From 6cf8d80ac0a9869d97373d6813441eabebce8980 Mon Sep 17 00:00:00 2001 From: herbelin Date: Thu, 23 Feb 2006 13:58:10 +0000 Subject: Nettoyage de l'archive doc et restructuration avant intégration à l'archive principale de Coq et publication des sources (HH) git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@8606 85f007b7-540e-0410-9357-904b9bb8a0f7 --- doc/refman/AddRefMan-pre.tex | 58 + doc/refman/Cases.tex | 698 +++++++++ doc/refman/Coercion.tex | 541 +++++++ doc/refman/Extraction.tex | 664 +++++++++ doc/refman/Helm.tex | 317 ++++ doc/refman/Natural.tex | 425 ++++++ doc/refman/Omega.tex | 226 +++ doc/refman/Polynom.tex | 504 +++++++ doc/refman/RefMan-add.tex | 54 + doc/refman/RefMan-cas.tex | 692 +++++++++ doc/refman/RefMan-cic.tex | 1480 +++++++++++++++++++ doc/refman/RefMan-coi.tex | 406 +++++ doc/refman/RefMan-com.tex | 280 ++++ doc/refman/RefMan-ext.tex | 1173 +++++++++++++++ doc/refman/RefMan-gal.tex | 1451 ++++++++++++++++++ doc/refman/RefMan-ide.tex | 327 +++++ doc/refman/RefMan-ind.tex | 498 +++++++ doc/refman/RefMan-int.tex | 147 ++ doc/refman/RefMan-lib.tex | 1102 ++++++++++++++ doc/refman/RefMan-ltac.tex | 1057 +++++++++++++ doc/refman/RefMan-mod.tex | 396 +++++ doc/refman/RefMan-modr.tex | 586 ++++++++ doc/refman/RefMan-oth.tex | 773 ++++++++++ doc/refman/RefMan-pre.tex | 519 +++++++ doc/refman/RefMan-pro.tex | 389 +++++ doc/refman/RefMan-syn.tex | 1016 +++++++++++++ doc/refman/RefMan-tac.tex | 3096 +++++++++++++++++++++++++++++++++++++++ doc/refman/RefMan-tacex.tex | 1208 +++++++++++++++ doc/refman/RefMan-tus.tex | 2015 +++++++++++++++++++++++++ doc/refman/RefMan-uti.tex | 276 ++++ doc/refman/Reference-Manual.tex | 124 ++ doc/refman/Setoid.tex | 158 ++ doc/refman/biblio.bib | 1144 +++++++++++++++ doc/refman/coqdoc.tex | 476 ++++++ doc/refman/coqide-queries.png | Bin 0 -> 27316 bytes doc/refman/coqide.png | Bin 0 -> 20953 bytes doc/refman/cover.html | 36 + doc/refman/headers.tex | 102 ++ doc/refman/hevea.sty | 78 + doc/refman/index.html | 29 + 40 files changed, 24521 insertions(+) create mode 100644 doc/refman/AddRefMan-pre.tex create mode 100644 doc/refman/Cases.tex create mode 100644 doc/refman/Coercion.tex create mode 100755 doc/refman/Extraction.tex create mode 100644 doc/refman/Helm.tex create mode 100755 doc/refman/Natural.tex create mode 100755 doc/refman/Omega.tex create mode 100644 doc/refman/Polynom.tex create mode 100755 doc/refman/RefMan-add.tex create mode 100755 doc/refman/RefMan-cas.tex create mode 100755 doc/refman/RefMan-cic.tex create mode 100755 doc/refman/RefMan-coi.tex create mode 100755 doc/refman/RefMan-com.tex create mode 100644 doc/refman/RefMan-ext.tex create mode 100644 doc/refman/RefMan-gal.tex create mode 100644 doc/refman/RefMan-ide.tex create mode 100755 doc/refman/RefMan-ind.tex create mode 100755 doc/refman/RefMan-int.tex create mode 100755 doc/refman/RefMan-lib.tex create mode 100644 doc/refman/RefMan-ltac.tex create mode 100644 doc/refman/RefMan-mod.tex create mode 100644 doc/refman/RefMan-modr.tex create mode 100644 doc/refman/RefMan-oth.tex create mode 100755 doc/refman/RefMan-pre.tex create mode 100755 doc/refman/RefMan-pro.tex create mode 100755 doc/refman/RefMan-syn.tex create mode 100644 doc/refman/RefMan-tac.tex create mode 100644 doc/refman/RefMan-tacex.tex create mode 100755 doc/refman/RefMan-tus.tex create mode 100755 doc/refman/RefMan-uti.tex create mode 100644 doc/refman/Reference-Manual.tex create mode 100644 doc/refman/Setoid.tex create mode 100755 doc/refman/biblio.bib create mode 100644 doc/refman/coqdoc.tex create mode 100644 doc/refman/coqide-queries.png create mode 100644 doc/refman/coqide.png create mode 100644 doc/refman/cover.html create mode 100644 doc/refman/headers.tex create mode 100644 doc/refman/hevea.sty create mode 100644 doc/refman/index.html (limited to 'doc/refman') diff --git a/doc/refman/AddRefMan-pre.tex b/doc/refman/AddRefMan-pre.tex new file mode 100644 index 0000000000..5312b8fc2f --- /dev/null +++ b/doc/refman/AddRefMan-pre.tex @@ -0,0 +1,58 @@ +%\coverpage{Addendum to the Reference Manual}{\ } +%\addcontentsline{toc}{part}{Additional documentation} +\setheaders{Presentation of the Addendum} +\chapter*{Presentation of the Addendum} + +Here you will find several pieces of additional documentation for the +\Coq\ Reference Manual. Each of this chapters is concentrated on a +particular topic, that should interest only a fraction of the \Coq\ +users: that's the reason why they are apart from the Reference +Manual. + +\begin{description} + +\item[Extended pattern-matching] This chapter details the use of + generalized pattern-matching. It is contributed by Cristina Cornes + and Hugo Herbelin. + +\item[Implicit coercions] This chapter details the use of the coercion + mechanism. It is contributed by Amokrane Saïbi. + +%\item[Proof of imperative programs] This chapter explains how to +% prove properties of annotated programs with imperative features. +% It is contributed by Jean-Christophe Filliâtre + +\item[Program extraction] This chapter explains how to extract in practice ML + files from $\FW$ terms. It is contributed by Jean-Christophe + Filliâtre and Pierre Letouzey. + +%\item[Natural] This chapter is due to Yann Coscoy. It is the user +% manual of the tools he wrote for printing proofs in natural +% language. At this time, French and English languages are supported. + +\item[omega] \texttt{omega}, written by Pierre Crégut, solves a whole + class of arithmetic problems. + +%\item[Program] The \texttt{Program} technology intends to inverse the +% extraction mechanism. It allows the developments of certified +% programs in \Coq. This chapter is due to Catherine Parent. {\bf This +% feature is not available in {\Coq} version 7.} + +\item[The {\tt ring} tactic] This is a tactic to do AC rewriting. This + chapter explains how to use it and how it works. + The chapter is contributed by Patrick Loiseleur. + +\item[The {\tt Setoid\_replace} tactic] This is a + tactic to do rewriting on types equipped with specific (only partially + substitutive) equality. The chapter is contributed by Clément Renard. + + +\end{description} + +\atableofcontents + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Cases.tex b/doc/refman/Cases.tex new file mode 100644 index 0000000000..95411afae9 --- /dev/null +++ b/doc/refman/Cases.tex @@ -0,0 +1,698 @@ +\achapter{Extended pattern-matching}\defaultheaders +\aauthor{Cristina Cornes} + +\label{Mult-match-full} +\ttindex{Cases} +\index{ML-like patterns} + +This section describes the full form of pattern-matching in {\Coq} terms. + +\asection{Patterns}\label{implementation} The full syntax of {\tt +match} is presented in figures~\ref{term-syntax} +and~\ref{term-syntax-aux}. Identifiers in patterns are either +constructor names or variables. Any identifier that is not the +constructor of an inductive or coinductive type is considered to be a +variable. A variable name cannot occur more than once in a given +pattern. It is recommended to start variable names by a lowercase +letter. + +If a pattern has the form $(c~\vec{x})$ where $c$ is a constructor +symbol and $\vec{x}$ is a linear vector of variables, it is called +{\em simple}: it is the kind of pattern recognized by the basic +version of {\tt match}. If a pattern is +not simple we call it {\em nested}. + +A variable pattern matches any value, and the identifier is bound to +that value. The pattern ``\texttt{\_}'' (called ``don't care'' or +``wildcard'' symbol) also matches any value, but does not bind anything. It +may occur an arbitrary number of times in a pattern. Alias patterns +written \texttt{(}{\sl pattern} \texttt{as} {\sl identifier}\texttt{)} are +also accepted. This pattern matches the same values as {\sl pattern} +does and {\sl identifier} is bound to the matched value. A list of +patterns separated with commas +is also considered as a pattern and is called {\em multiple +pattern}. + +Since extended {\tt match} expressions are compiled into the primitive +ones, the expressiveness of the theory remains the same. Once the +stage of parsing has finished only simple patterns remain. An easy way +to see the result of the expansion is by printing the term with +\texttt{Print} if the term is a constant, or using the command +\texttt{Check}. + +The extended \texttt{match} still accepts an optional {\em elimination +predicate} given after the keyword \texttt{return}. Given a pattern +matching expression, if all the right hand sides of \texttt{=>} ({\em +rhs} in short) have the same type, then this type can be sometimes +synthesized, and so we can omit the \texttt{return} part. Otherwise +the predicate after \texttt{return} has to be provided, like for the basic +\texttt{match}. + +Let us illustrate through examples the different aspects of extended +pattern matching. Consider for example the function that computes the +maximum of two natural numbers. We can write it in primitive syntax +by: + +\begin{coq_example} +Fixpoint max (n m:nat) {struct m} : nat := + match n with + | O => m + | S n' => match m with + | O => S n' + | S m' => S (max n' m') + end + end. +\end{coq_example} + +Using multiple patterns in the definition allows to write: + +\begin{coq_example} +Reset max. +Fixpoint max (n m:nat) {struct m} : nat := + match n, m with + | O, _ => m + | S n', O => S n' + | S n', S m' => S (max n' m') + end. +\end{coq_example} + +which will be compiled into the previous form. + +The pattern-matching compilation strategy examines patterns from left +to right. A \texttt{match} expression is generated {\bf only} when +there is at least one constructor in the column of patterns. E.g. the +following example does not build a \texttt{match} expression. + +\begin{coq_example} +Check (fun x:nat => match x return nat with + | y => y + end). +\end{coq_example} + +We can also use ``\texttt{as} patterns'' to associate a name to a +sub-pattern: + +\begin{coq_example} +Reset max. +Fixpoint max (n m:nat) {struct n} : nat := + match n, m with + | O, _ => m + | S n' as p, O => p + | S n', S m' => S (max n' m') + end. +\end{coq_example} + +Here is now an example of nested patterns: + +\begin{coq_example} +Fixpoint even (n:nat) : bool := + match n with + | O => true + | S O => false + | S (S n') => even n' + end. +\end{coq_example} + +This is compiled into: + +\begin{coq_example} +Print even. +\end{coq_example} + +In the previous examples patterns do not conflict with, but +sometimes it is comfortable to write patterns that admit a non +trivial superposition. Consider +the boolean function \texttt{lef} that given two natural numbers +yields \texttt{true} if the first one is less or equal than the second +one and \texttt{false} otherwise. We can write it as follows: + +\begin{coq_example} +Fixpoint lef (n m:nat) {struct m} : bool := + match n, m with + | O, x => true + | x, O => false + | S n, S m => lef n m + end. +\end{coq_example} + +Note that the first and the second multiple pattern superpose because +the couple of values \texttt{O O} matches both. Thus, what is the result +of the function on those values? To eliminate ambiguity we use the +{\em textual priority rule}: we consider patterns ordered from top to +bottom, then a value is matched by the pattern at the $ith$ row if and +only if it is not matched by some pattern of a previous row. Thus in the +example, +\texttt{O O} is matched by the first pattern, and so \texttt{(lef O O)} +yields \texttt{true}. + +Another way to write this function is: + +\begin{coq_example} +Reset lef. +Fixpoint lef (n m:nat) {struct m} : bool := + match n, m with + | O, x => true + | S n, S m => lef n m + | _, _ => false + end. +\end{coq_example} + + +Here the last pattern superposes with the first two. Because +of the priority rule, the last pattern +will be used only for values that do not match neither the first nor +the second one. + +Terms with useless patterns are not accepted by the +system. Here is an example: +% Test failure +\begin{coq_eval} +Set Printing Depth 50. + (********** The following is not correct and should produce **********) + (**************** Error: This clause is redundant ********************) +\end{coq_eval} +\begin{coq_example} +Check (fun x:nat => + match x with + | O => true + | S _ => false + | x => true + end). +\end{coq_example} + +\asection{About patterns of parametric types} +When matching objects of a parametric type, constructors in patterns +{\em do not expect} the parameter arguments. Their value is deduced +during expansion. + +Consider for example the polymorphic lists: + +\begin{coq_example} +Inductive List (A:Set) : Set := + | nil : List A + | cons : A -> List A -> List A. +\end{coq_example} + +We can check the function {\em tail}: + +\begin{coq_example} +Check + (fun l:List nat => + match l with + | nil => nil nat + | cons _ l' => l' + end). +\end{coq_example} + + +When we use parameters in patterns there is an error message: +% Test failure +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is not correct and should produce **********) +(******** Error: The constructor cons expects 2 arguments ************) +\end{coq_eval} +\begin{coq_example} +Check + (fun l:List nat => + match l with + | nil A => nil nat + | cons A _ l' => l' + end). +\end{coq_example} + + + +\asection{Matching objects of dependent types} +The previous examples illustrate pattern matching on objects of +non-dependent types, but we can also +use the expansion strategy to destructure objects of dependent type. +Consider the type \texttt{listn} of lists of a certain length: + +\begin{coq_example} +Inductive listn : nat -> Set := + | niln : listn 0 + | consn : forall n:nat, nat -> listn n -> listn (S n). +\end{coq_example} + +\asubsection{Understanding dependencies in patterns} +We can define the function \texttt{length} over \texttt{listn} by: + +\begin{coq_example} +Definition length (n:nat) (l:listn n) := n. +\end{coq_example} + +Just for illustrating pattern matching, +we can define it by case analysis: + +\begin{coq_example} +Reset length. +Definition length (n:nat) (l:listn n) := + match l with + | niln => 0 + | consn n _ _ => S n + end. +\end{coq_example} + +We can understand the meaning of this definition using the +same notions of usual pattern matching. + +% +% Constraining of dependencies is not longer valid in V7 +% +\iffalse +Now suppose we split the second pattern of \texttt{length} into two +cases so to give an +alternative definition using nested patterns: +\begin{coq_example} +Definition length1 (n:nat) (l:listn n) := + match l with + | niln => 0 + | consn n _ niln => S n + | consn n _ (consn _ _ _) => S n + end. +\end{coq_example} + +It is obvious that \texttt{length1} is another version of +\texttt{length}. We can also give the following definition: +\begin{coq_example} +Definition length2 (n:nat) (l:listn n) := + match l with + | niln => 0 + | consn n _ niln => 1 + | consn n _ (consn m _ _) => S (S m) + end. +\end{coq_example} + +If we forget that \texttt{listn} is a dependent type and we read these +definitions using the usual semantics of pattern matching, we can conclude +that \texttt{length1} +and \texttt{length2} are different functions. +In fact, they are equivalent +because the pattern \texttt{niln} implies that \texttt{n} can only match +the value $0$ and analogously the pattern \texttt{consn} determines that \texttt{n} can +only match values of the form $(S~v)$ where $v$ is the value matched by +\texttt{m}. + +The converse is also true. If +we destructure the length value with the pattern \texttt{O} then the list +value should be $niln$. +Thus, the following term \texttt{length3} corresponds to the function +\texttt{length} but this time defined by case analysis on the dependencies instead of on the list: + +\begin{coq_example} +Definition length3 (n:nat) (l:listn n) := + match l with + | niln => 0 + | consn O _ _ => 1 + | consn (S n) _ _ => S (S n) + end. +\end{coq_example} + +When we have nested patterns of dependent types, the semantics of +pattern matching becomes a little more difficult because +the set of values that are matched by a sub-pattern may be conditioned by the +values matched by another sub-pattern. Dependent nested patterns are +somehow constrained patterns. +In the examples, the expansion of +\texttt{length1} and \texttt{length2} yields exactly the same term + but the +expansion of \texttt{length3} is completely different. \texttt{length1} and +\texttt{length2} are expanded into two nested case analysis on +\texttt{listn} while \texttt{length3} is expanded into a case analysis on +\texttt{listn} containing a case analysis on natural numbers inside. + + +In practice the user can think about the patterns as independent and +it is the expansion algorithm that cares to relate them. \\ +\fi +% +% +% + +\asubsection{When the elimination predicate must be provided} +The examples given so far do not need an explicit elimination predicate + because all the rhs have the same type and the +strategy succeeds to synthesize it. +Unfortunately when dealing with dependent patterns it often happens +that we need to write cases where the type of the rhs are +different instances of the elimination predicate. +The function \texttt{concat} for \texttt{listn} +is an example where the branches have different type +and we need to provide the elimination predicate: + +\begin{coq_example} +Fixpoint concat (n:nat) (l:listn n) (m:nat) (l':listn m) {struct l} : + listn (n + m) := + match l in listn n return listn (n + m) with + | niln => l' + | consn n' a y => consn (n' + m) a (concat n' y m l') + end. +\end{coq_example} +The elimination predicate is {\tt fun (n:nat) (l:listn n) => listn~(n+m)}. +In general if $m$ has type $(I~q_1\ldots q_r~t_1\ldots t_s)$ where +$q_1\ldots q_r$ are parameters, the elimination predicate should be of +the form~: +{\tt fun $y_1$\ldots $y_s$ $x$:($I$~$q_1$\ldots $q_r$~$y_1$\ldots + $y_s$) => P}. + +In the concrete syntax, it should be written~: +\[ \kw{match}~m~\kw{as}~x~\kw{in}~(I~\_\ldots \_~y_1\ldots y_s)~\kw{return}~Q~\kw{with}~\ldots~\kw{end}\] + +The variables which appear in the \kw{in} and \kw{as} clause are new +and bounded in the property $Q$ in the \kw{return} clause. The +parameters of the inductive definitions should not be mentioned and +are replaced by \kw{\_}. + +Recall that a list of patterns is also a pattern. So, when +we destructure several terms at the same time and the branches have +different type we need to provide +the elimination predicate for this multiple pattern. +It is done using the same scheme, each term may be associated to an +\kw{as} and \kw{in} clause in order to introduce a dependent product. + +For example, an equivalent definition for \texttt{concat} (even though the matching on the second term is trivial) would have +been: + +\begin{coq_example} +Reset concat. +Fixpoint concat (n:nat) (l:listn n) (m:nat) (l':listn m) {struct l} : + listn (n + m) := + match l in listn n, l' return listn (n + m) with + | niln, x => x + | consn n' a y, x => consn (n' + m) a (concat n' y m x) + end. +\end{coq_example} + +% Notice that this time, the predicate \texttt{[n,\_:nat](listn (plus n +% m))} is binary because we +% destructure both \texttt{l} and \texttt{l'} whose types have arity one. +% In general, if we destructure the terms $e_1\ldots e_n$ +% the predicate will be of arity $m$ where $m$ is the sum of the +% number of dependencies of the type of $e_1, e_2,\ldots e_n$ +% (the $\lambda$-abstractions +% should correspond from left to right to each dependent argument of the +% type of $e_1\ldots e_n$). +When the arity of the predicate (i.e. number of abstractions) is not +correct Coq raises an error message. For example: + +% Test failure +\begin{coq_eval} +Reset concat. +Set Printing Depth 50. +(********** The following is not correct and should produce ***********) +(** Error: the term l' has type listn m while it is expected to have **) +(** type listn (?31 + ?32) **) +\end{coq_eval} +\begin{coq_example} +Fixpoint concat + (n:nat) (l:listn n) (m:nat) + (l':listn m) {struct l} : listn (n + m) := + match l, l' with + | niln, x => x + | consn n' a y, x => consn (n' + m) a (concat n' y m x) + end. +\end{coq_example} + +\asection{Using pattern matching to write proofs} +In all the previous examples the elimination predicate does not depend +on the object(s) matched. But it may depend and the typical case +is when we write a proof by induction or a function that yields an +object of dependent type. An example of proof using \texttt{match} in +given in section \ref{refine-example} + +For example, we can write +the function \texttt{buildlist} that given a natural number +$n$ builds a list of length $n$ containing zeros as follows: + +\begin{coq_example} +Fixpoint buildlist (n:nat) : listn n := + match n return listn n with + | O => niln + | S n => consn n 0 (buildlist n) + end. +\end{coq_example} + +We can also use multiple patterns. +Consider the following definition of the predicate less-equal +\texttt{Le}: + +\begin{coq_example} +Inductive LE : nat -> nat -> Prop := + | LEO : forall n:nat, LE 0 n + | LES : forall n m:nat, LE n m -> LE (S n) (S m). +\end{coq_example} + +We can use multiple patterns to write the proof of the lemma + \texttt{(n,m:nat) (LE n m)}\verb=\/=\texttt{(LE m n)}: + +\begin{coq_example} +Fixpoint dec (n m:nat) {struct n} : LE n m \/ LE m n := + match n, m return LE n m \/ LE m n with + | O, x => or_introl (LE x 0) (LEO x) + | x, O => or_intror (LE x 0) (LEO x) + | S n as n', S m as m' => + match dec n m with + | or_introl h => or_introl (LE m' n') (LES n m h) + | or_intror h => or_intror (LE n' m') (LES m n h) + end + end. +\end{coq_example} +In the example of \texttt{dec}, +the first \texttt{match} is dependent while +the second is not. + +% In general, consider the terms $e_1\ldots e_n$, +% where the type of $e_i$ is an instance of a family type +% $\lb (\vec{d_i}:\vec{D_i}) \mto T_i$ ($1\leq i +% \leq n$). Then, in expression \texttt{match} $e_1,\ldots, +% e_n$ \texttt{of} \ldots \texttt{end}, the +% elimination predicate ${\cal P}$ should be of the form: +% $[\vec{d_1}:\vec{D_1}][x_1:T_1]\ldots [\vec{d_n}:\vec{D_n}][x_n:T_n]Q.$ + +The user can also use \texttt{match} in combination with the tactic +\texttt{refine} (see section \ref{refine}) to build incomplete proofs +beginning with a \texttt{match} construction. + +\asection{Pattern-matching on inductive objects involving local +definitions} + +If local definitions occur in the type of a constructor, then there +are two ways to match on this constructor. Either the local +definitions are skipped and matching is done only on the true arguments +of the constructors, or the bindings for local definitions can also +be caught in the matching. + +Example. + +\begin{coq_eval} +Reset Initial. +Require Import Arith. +\end{coq_eval} + +\begin{coq_example*} +Inductive list : nat -> Set := + | nil : list 0 + | cons : forall n:nat, let m := (2 * n) in list m -> list (S (S m)). +\end{coq_example*} + +In the next example, the local definition is not caught. + +\begin{coq_example} +Fixpoint length n (l:list n) {struct l} : nat := + match l with + | nil => 0 + | cons n l0 => S (length (2 * n) l0) + end. +\end{coq_example} + +But in this example, it is. + +\begin{coq_example} +Fixpoint length' n (l:list n) {struct l} : nat := + match l with + | nil => 0 + | cons _ m l0 => S (length' m l0) + end. +\end{coq_example} + +\Rem for a given matching clause, either none of the local +definitions or all of them can be caught. + +\asection{Pattern-matching and coercions} + +If a mismatch occurs between the expected type of a pattern and its +actual type, a coercion made from constructors is sought. If such a +coercion can be found, it is automatically inserted around the +pattern. + +Example: + +\begin{coq_example} +Inductive I : Set := + | C1 : nat -> I + | C2 : I -> I. +Coercion C1 : nat >-> I. +Check (fun x => match x with + | C2 O => 0 + | _ => 0 + end). +\end{coq_example} + + +\asection{When does the expansion strategy fail ?}\label{limitations} +The strategy works very like in ML languages when treating +patterns of non-dependent type. +But there are new cases of failure that are due to the presence of +dependencies. + +The error messages of the current implementation may be sometimes +confusing. When the tactic fails because patterns are somehow +incorrect then error messages refer to the initial expression. But the +strategy may succeed to build an expression whose sub-expressions are +well typed when the whole expression is not. In this situation the +message makes reference to the expanded expression. We encourage +users, when they have patterns with the same outer constructor in +different equations, to name the variable patterns in the same +positions with the same name. +E.g. to write {\small\texttt{(cons n O x) => e1}} +and {\small\texttt{(cons n \_ x) => e2}} instead of +{\small\texttt{(cons n O x) => e1}} and +{\small\texttt{(cons n' \_ x') => e2}}. +This helps to maintain certain name correspondence between the +generated expression and the original. + +Here is a summary of the error messages corresponding to each situation: + +\begin{ErrMsgs} +\item \sverb{The constructor } {\sl + ident} \sverb{expects } {\sl num} \sverb{arguments} + + \sverb{The variable } {\sl ident} \sverb{is bound several times + in pattern } {\sl term} + + \sverb{Found a constructor of inductive type} {\term} + \sverb{while a constructor of} {\term} \sverb{is expected} + + Patterns are incorrect (because constructors are not applied to + the correct number of the arguments, because they are not linear or + they are wrongly typed) + +\item \errindex{Non exhaustive pattern-matching} + +the pattern matching is not exhaustive + +\item \sverb{The elimination predicate } {\sl term} \sverb{should be + of arity } {\sl num} \sverb{(for non dependent case) or } {\sl + num} \sverb{(for dependent case)} + +The elimination predicate provided to \texttt{match} has not the + expected arity + + +%\item the whole expression is wrongly typed + +% CADUC ? +% , or the synthesis of +% implicit arguments fails (for example to find the elimination +% predicate or to resolve implicit arguments in the rhs). + +% There are {\em nested patterns of dependent type}, the elimination +% predicate corresponds to non-dependent case and has the form +% $[x_1:T_1]...[x_n:T_n]T$ and {\bf some} $x_i$ occurs {\bf free} in +% $T$. Then, the strategy may fail to find out a correct elimination +% predicate during some step of compilation. In this situation we +% recommend the user to rewrite the nested dependent patterns into +% several \texttt{match} with {\em simple patterns}. + +\item {\tt Unable to infer a match predicate\\ + Either there is a type incompatiblity or the problem involves\\ + dependencies} + + There is a type mismatch between the different branches + + Then the user should provide an elimination predicate. + +% Obsolete ? +% \item because of nested patterns, it may happen that even though all +% the rhs have the same type, the strategy needs dependent elimination +% and so an elimination predicate must be provided. The system warns +% about this situation, trying to compile anyway with the +% non-dependent strategy. The risen message is: + +% \begin{itemize} +% \item {\tt Warning: This pattern matching may need dependent +% elimination to be compiled. I will try, but if fails try again +% giving dependent elimination predicate.} +% \end{itemize} + + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +% % LA PROPAGATION DES CONTRAINTES ARRIERE N'EST PAS FAITE DANS LA V7 +% TODO +% \item there are {\em nested patterns of dependent type} and the +% strategy builds a term that is well typed but recursive calls in fix +% point are reported as illegal: +% \begin{itemize} +% \item {\tt Error: Recursive call applied to an illegal term ...} +% \end{itemize} + +% This is because the strategy generates a term that is correct w.r.t. +% the initial term but which does not pass the guard condition. In +% this situation we recommend the user to transform the nested dependent +% patterns into {\em several \texttt{match} of simple patterns}. Let us +% explain this with an example. Consider the following definition of a +% function that yields the last element of a list and \texttt{O} if it is +% empty: + +% \begin{coq_example} +% Fixpoint last [n:nat; l:(listn n)] : nat := +% match l of +% (consn _ a niln) => a +% | (consn m _ x) => (last m x) | niln => O +% end. +% \end{coq_example} + +% It fails because of the priority between patterns, we know that this +% definition is equivalent to the following more explicit one (which +% fails too): + +% \begin{coq_example*} +% Fixpoint last [n:nat; l:(listn n)] : nat := +% match l of +% (consn _ a niln) => a +% | (consn n _ (consn m b x)) => (last n (consn m b x)) +% | niln => O +% end. +% \end{coq_example*} + +% Note that the recursive call {\tt (last n (consn m b x))} is not +% guarded. When treating with patterns of dependent types the strategy +% interprets the first definition of \texttt{last} as the second +% one\footnote{In languages of the ML family the first definition would +% be translated into a term where the variable \texttt{x} is shared in +% the expression. When patterns are of non-dependent types, Coq +% compiles as in ML languages using sharing. When patterns are of +% dependent types the compilation reconstructs the term as in the +% second definition of \texttt{last} so to ensure the result of +% expansion is well typed.}. Thus it generates a term where the +% recursive call is rejected by the guard condition. + +% You can get rid of this problem by writing the definition with +% \emph{simple patterns}: + +% \begin{coq_example} +% Fixpoint last [n:nat; l:(listn n)] : nat := +% <[_:nat]nat>match l of +% (consn m a x) => Cases x of niln => a | _ => (last m x) end +% | niln => O +% end. +% \end{coq_example} + +\end{ErrMsgs} + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Coercion.tex b/doc/refman/Coercion.tex new file mode 100644 index 0000000000..5445224b02 --- /dev/null +++ b/doc/refman/Coercion.tex @@ -0,0 +1,541 @@ +\achapter{Implicit Coercions} +\aauthor{Amokrane Saïbi} + +\label{Coercions-full} +\index{Coercions!presentation} + +\asection{General Presentation} + +This section describes the inheritance mechanism of {\Coq}. In {\Coq} with +inheritance, we are not interested in adding any expressive power to +our theory, but only convenience. Given a term, possibly not typable, +we are interested in the problem of determining if it can be well +typed modulo insertion of appropriate coercions. We allow to write: + +\begin{itemize} +\item $f~a$ where $f:forall~ x:A, B$ and $a:A'$ when $A'$ can + be seen in some sense as a subtype of $A$. +\item $x:A$ when $A$ is not a type, but can be seen in + a certain sense as a type: set, group, category etc. +\item $f~a$ when $f$ is not a function, but can be seen in a certain sense + as a function: bijection, functor, any structure morphism etc. +\end{itemize} + +\asection{Classes} +\index{Coercions!classes} + A class with $n$ parameters is any defined name with a type +$forall~ (x_1:A_1)..(x_n:A_n), s$ where $s$ is a sort. Thus a class with +parameters is considered as a single class and not as a family of +classes. An object of a class $C$ is any term of type $C~t_1 +.. t_n$. In addition to these user-classes, we have two abstract +classes: + +\begin{itemize} +\item {\tt Sortclass}, the class of sorts; + its objects are the terms whose type is a sort. +\item {\tt Funclass}, the class of functions; + its objects are all the terms with a functional + type, i.e. of form $forall~ x:A, B$. +\end{itemize} + +Formally, the syntax of a classes is defined on Figure~\ref{fig:classes}. +\begin{figure} +\begin{centerframe} +\begin{tabular}{lcl} +{\class} & ::= & {\qualid} \\ + & $|$ & {\tt Sortclass} \\ + & $|$ & {\tt Funclass} +\end{tabular} +\end{centerframe} +\caption{Syntax of classes} +\label{fig:classes} +\end{figure} + +\asection{Coercions} +\index{Coercions!Funclass} +\index{Coercions!Sortclass} + A name $f$ can be declared as a coercion between a source user-class +$C$ with $n$ parameters and a target class $D$ if one of these +conditions holds: + +\newcommand{\oftype}{\!:\!} + +\begin{itemize} +\item $D$ is a user-class, then the type of $f$ must have the form + $forall~ (x_1 \oftype A_1)..(x_n \oftype A_n)(y\oftype C~x_1..x_n), D~u_1..u_m$ where $m$ + is the number of parameters of $D$. +\item $D$ is {\tt Funclass}, then the type of $f$ must have the form + $forall~ (x_1\oftype A_1)..(x_n\oftype A_n)(y\oftype C~x_1..x_n)(x:A), B$. +\item $D$ is {\tt Sortclass}, then the type of $f$ must have the form + $forall~ (x_1\oftype A_1)..(x_n\oftype A_n)(y\oftype C~x_1..x_n), s$ with $s$ a sort. +\end{itemize} + +We then write $f:C \mbox{\texttt{>->}} D$. The restriction on the type +of coercions is called {\em the uniform inheritance condition}. +Remark that the abstract classes {\tt Funclass} and {\tt Sortclass} +cannot be source classes. + +To coerce an object $t:C~t_1..t_n$ of $C$ towards $D$, we have to +apply the coercion $f$ to it; the obtained term $f~t_1..t_n~t$ is +then an object of $D$. + +\asection{Identity Coercions} +\index{Coercions!identity} + + Identity coercions are special cases of coercions used to go around +the uniform inheritance condition. Let $C$ and $D$ be two classes +with respectively $n$ and $m$ parameters and +$f:forall~(x_1:T_1)..(x_k:T_k)(y:C~u_1..u_n), D~v_1..v_m$ a function which +does not verify the uniform inheritance condition. To declare $f$ as +coercion, one has first to declare a subclass $C'$ of $C$: + +$$C' := fun~ (x_1:T_1)..(x_k:T_k) => C~u_1..u_n$$ + +\noindent We then define an {\em identity coercion} between $C'$ and $C$: +\begin{eqnarray*} +Id\_C'\_C & := & fun~ (x_1:T_1)..(x_k:T_k)(y:C'~x_1..x_k) => (y:C~u_1..u_n)\\ +\end{eqnarray*} + +We can now declare $f$ as coercion from $C'$ to $D$, since we can +``cast'' its type as +$forall~ (x_1:T_1)..(x_k:T_k)(y:C'~x_1..x_k),D~v_1..v_m$.\\ The identity +coercions have a special status: to coerce an object $t:C'~t_1..t_k$ +of $C'$ towards $C$, we does not have to insert explicitly $Id\_C'\_C$ +since $Id\_C'\_C~t_1..t_k~t$ is convertible with $t$. However we +``rewrite'' the type of $t$ to become an object of $C$; in this case, +it becomes $C~u_1^*..u_k^*$ where each $u_i^*$ is the result of the +substitution in $u_i$ of the variables $x_j$ by $t_j$. + + +\asection{Inheritance Graph} +\index{Coercions!inheritance graph} +Coercions form an inheritance graph with classes as nodes. We call +{\em coercion path} an ordered list of coercions between two nodes of +the graph. A class $C$ is said to be a subclass of $D$ if there is a +coercion path in the graph from $C$ to $D$; we also say that $C$ +inherits from $D$. Our mechanism supports multiple inheritance since a +class may inherit from several classes, contrary to simple inheritance +where a class inherits from at most one class. However there must be +at most one path between two classes. If this is not the case, only +the {\em oldest} one is valid and the others are ignored. So the order +of declaration of coercions is important. + +We extend notations for coercions to coercion paths. For instance +$[f_1;..;f_k]:C \mbox{\texttt{>->}} D$ is the coercion path composed +by the coercions $f_1..f_k$. The application of a coercion path to a +term consists of the successive application of its coercions. + +\asection{Declaration of Coercions} + +%%%%% "Class" is useless, since classes are implicitely defined via coercions. + +% \asubsection{\tt Class {\qualid}.}\comindex{Class} +% Declares {\qualid} as a new class. + +% \begin{ErrMsgs} +% \item {\qualid} \errindex{not declared} +% \item {\qualid} \errindex{is already a class} +% \item \errindex{Type of {\qualid} does not end with a sort} +% \end{ErrMsgs} + +% \begin{Variant} +% \item {\tt Class Local {\qualid}.} \\ +% Declares the construction denoted by {\qualid} as a new local class to +% the current section. +% \end{Variant} + +% END "Class" is useless + +\asubsection{\tt Coercion {\qualid} : {\class$_1$} >-> {\class$_2$}.} +\comindex{Coercion} + +Declares the construction denoted by {\qualid} as a coercion between +{\class$_1$} and {\class$_2$}. + +% Useless information +% The classes {\class$_1$} and {\class$_2$} are first declared if necessary. + +\begin{ErrMsgs} +\item {\qualid} \errindex{not declared} +\item {\qualid} \errindex{is already a coercion} +\item \errindex{Funclass cannot be a source class} +\item \errindex{Sortclass cannot be a source class} +\item {\qualid} \errindex{is not a function} +\item \errindex{Cannot find the source class of {\qualid}} +\item \errindex{Cannot recognize {\class$_1$} as a source class of {\qualid}} +\item {\qualid} \errindex{does not respect the inheritance uniform condition} +\item \errindex{Found target class {\class} instead of {\class$_2$}} + +\end{ErrMsgs} + +When the coercion {\qualid} is added to the inheritance graph, non +valid coercion paths are ignored; they are signaled by a warning. +\\[0.3cm] +\noindent {\bf Warning :} +\begin{enumerate} +\item \begin{tabbing} +{\tt Ambiguous paths: }\= $[f_1^1;..;f_{n_1}^1] : C_1\mbox{\tt >->}D_1$\\ + \> ... \\ + \>$[f_1^m;..;f_{n_m}^m] : C_m\mbox{\tt >->}D_m$ + \end{tabbing} +\end{enumerate} + +\begin{Variants} +\item {\tt Coercion Local {\qualid} : {\class$_1$} >-> {\class$_2$}.} +\comindex{Coercion Local}\\ + Declares the construction denoted by {\qualid} as a coercion local to + the current section. + +\item {\tt Coercion {\ident} := {\term}}\comindex{Coercion}\\ + This defines {\ident} just like \texttt{Definition {\ident} := + {\term}}, and then declares {\ident} as a coercion between it + source and its target. + +\item {\tt Coercion {\ident} := {\term} : {\type}}\\ + This defines {\ident} just like + \texttt{Definition {\ident} : {\type} := {\term}}, and then + declares {\ident} as a coercion between it source and its target. + +\item {\tt Coercion Local {\ident} := {\term}}\comindex{Coercion Local}\\ + This defines {\ident} just like \texttt{Local {\ident} := + {\term}}, and then declares {\ident} as a coercion between it + source and its target. + +\item Assumptions can be declared as coercions at declaration +time. This extends the grammar of declarations from Figure +\ref{sentences-syntax} as follows: +\comindex{Variable \mbox{\rm (and coercions)}} +\comindex{Axiom \mbox{\rm (and coercions)}} +\comindex{Parameter \mbox{\rm (and coercions)}} +\comindex{Hypothesis \mbox{\rm (and coercions)}} + +\begin{tabular}{lcl} +%% Declarations +{\declaration} & ::= & {\declarationkeyword} {\assums} {\tt .} \\ +&&\\ +{\assums} & ::= & {\simpleassums} \\ + & $|$ & \nelist{{\tt (} \simpleassums {\tt )}}{} \\ +&&\\ +{\simpleassums} & ::= & \nelist{\ident}{} {\tt :}\zeroone{{\tt >}} {\term}\\ +\end{tabular} + +If the extra {\tt >} is present before the type of some assumptions, these +assumptions are declared as coercions. + +\item Constructors of inductive types can be declared as coercions at +definition time of the inductive type. This extends and modifies the +grammar of inductive types from Figure \ref{sentences-syntax} as follows: +\comindex{Inductive \mbox{\rm (and coercions)}} +\comindex{CoInductive \mbox{\rm (and coercions)}} + +\begin{center} +\begin{tabular}{lcl} +%% Inductives +{\inductive} & ::= & + {\tt Inductive} \nelist{\inductivebody}{with} {\tt .} \\ + & $|$ & {\tt CoInductive} \nelist{\inductivebody}{with} {\tt .} \\ + & & \\ +{\inductivebody} & ::= & + {\ident} \sequence{\binderlet}{} {\tt :} {\term} {\tt :=} \\ + && ~~~\zeroone{\zeroone{\tt |} \nelist{\constructor}{|}} \\ + & & \\ +{\constructor} & ::= & {\ident} \sequence{\binderlet}{} \zeroone{{\tt :}\zeroone{\tt >} {\term}} \\ +\end{tabular} +\end{center} + +Especially, if the extra {\tt >} is present in a constructor +declaration, this constructor is declared as a coercion. +\end{Variants} + +\asubsection{\tt Identity Coercion {\ident}:{\class$_1$} >-> {\class$_2$}.} +\comindex{Identity Coercion} + +We check that {\class$_1$} is a constant with a value of the form +$fun~ (x_1:T_1)..(x_n:T_n) => (\mbox{\class}_2~t_1..t_m)$ where $m$ is the +number of parameters of \class$_2$. Then we define an identity +function with the type +$forall~ (x_1:T_1)..(x_n:T_n)(y:\mbox{\class}_1~x_1..x_n), +{\mbox{\class}_2}~t_1..t_m$, and we declare it as an identity +coercion between {\class$_1$} and {\class$_2$}. + +\begin{ErrMsgs} +\item {\class$_1$} \errindex{must be a transparent constant} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Identity Coercion Local {\ident}:{\ident$_1$} >-> {\ident$_2$}.} \\ +Idem but locally to the current section. + +\item {\tt SubClass {\ident} := {\type}.} \\ +\comindex{SubClass} + If {\type} is a class +{\ident'} applied to some arguments then {\ident} is defined and an +identity coercion of name {\tt Id\_{\ident}\_{\ident'}} is +declared. Otherwise said, this is an abbreviation for + +{\tt Definition {\ident} := {\type}.} + + followed by + +{\tt Identity Coercion Id\_{\ident}\_{\ident'}:{\ident} >-> {\ident'}}. + +\item {\tt Local SubClass {\ident} := {\type}.} \\ +Same as before but locally to the current section. + +\end{Variants} + +\asection{Displaying Available Coercions} + +\asubsection{\tt Print Classes.} +\comindex{Print Classes} +Print the list of declared classes in the current context. + +\asubsection{\tt Print Coercions.} +\comindex{Print Coercions} +Print the list of declared coercions in the current context. + +\asubsection{\tt Print Graph.} +\comindex{Print Graph} +Print the list of valid coercion paths in the current context. + +\asubsection{\tt Print Coercion Paths {\class$_1$} {\class$_2$}.} +\comindex{Print Coercion Paths} +Print the list of valid coercion paths from {\class$_1$} to {\class$_2$}. + +\asection{Activating the Printing of Coercions} + +\asubsection{\tt Set Printing Coercions.} +\comindex{Set Printing Coercions} +\comindex{Unset Printing Coercions} + +This command forces all the coercions to be printed. +Conversely, to skip the printing of coercions, use + {\tt Unset Printing Coercions}. +By default, coercions are not printed. + +\asubsection{\tt Set Printing Coercion {\qualid}.} +\comindex{Set Printing Coercion} +\comindex{Unset Printing Coercion} + +This command forces coercion denoted by {\qualid} to be printed. +To skip the printing of coercion {\qualid}, use + {\tt Unset Printing Coercion {\qualid}}. +By default, a coercion is never printed. + +\asection{Classes as Records} +\label{Coercions-and-records} +\index{Coercions!and records} +We allow the definition of {\em Structures with Inheritance} (or +classes as records) by extending the existing {\tt Record} macro +(see section~\ref{Record}). Its new syntax is: + +\begin{center} +\begin{tabular}{l} +{\tt Record \zeroone{>}~{\ident} {\binderlet} : {\sort} := \zeroone{\ident$_0$} \verb+{+} \\ +~~~~\begin{tabular}{l} + {\tt \ident$_1$ $[$:$|$:>$]$ \term$_1$ ;} \\ + ... \\ + {\tt \ident$_n$ $[$:$|$:>$]$ \term$_n$ \verb+}+. } + \end{tabular} +\end{tabular} +\end{center} +The identifier {\ident} is the name of the defined record and {\sort} +is its type. The identifier {\ident$_0$} is the name of its +constructor. The identifiers {\ident$_1$}, .., {\ident$_n$} are the +names of its fields and {\term$_1$}, .., {\term$_n$} their respective +types. The alternative {\tt $[$:$|$:>$]$} is ``{\tt :}'' or ``{\tt +:>}''. If {\tt {\ident$_i$}:>{\term$_i$}}, then {\ident$_i$} is +automatically declared as coercion from {\ident} to the class of +{\term$_i$}. Remark that {\ident$_i$} always verifies the uniform +inheritance condition. If the optional ``{\tt >}'' before {\ident} is +present, then {\ident$_0$} (or the default name {\tt Build\_{\ident}} +if {\ident$_0$} is omitted) is automatically declared as a coercion +from the class of {\term$_n$} to {\ident} (this may fail if the +uniform inheritance condition is not satisfied). + +\Rem The keyword {\tt Structure}\comindex{Structure} is a synonym of {\tt +Record}. + +\asection{Coercions and Sections} +\index{Coercions!and sections} + The inheritance mechanism is compatible with the section +mechanism. The global classes and coercions defined inside a section +are redefined after its closing, using their new value and new +type. The classes and coercions which are local to the section are +simply forgotten. +Coercions with a local source class or a local target class, and +coercions which do not verify the uniform inheritance condition any longer +are also forgotten. + +\asection{Examples} + + There are three situations: + +\begin{itemize} +\item $f~a$ is ill-typed where $f:forall~x:A,B$ and $a:A'$. If there is a + coercion path between $A'$ and $A$, $f~a$ is transformed into + $f~a'$ where $a'$ is the result of the application of this + coercion path to $a$. + +We first give an example of coercion between atomic inductive types + +%\begin{\small} +\begin{coq_example} +Definition bool_in_nat (b:bool) := if b then 0 else 1. +Coercion bool_in_nat : bool >-> nat. +Check (0 = true). +Set Printing Coercions. +Check (0 = true). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + +\Warning ``\verb|Check true=O.|'' fails. This is ``normal'' behaviour of +coercions. To validate \verb|true=O|, the coercion is searched from +\verb=nat= to \verb=bool=. There is none. + +We give an example of coercion between classes with parameters. + +%\begin{\small} +\begin{coq_example} +Parameters + (C : nat -> Set) (D : nat -> bool -> Set) (E : bool -> Set). +Parameter f : forall n:nat, C n -> D (S n) true. +Coercion f : C >-> D. +Parameter g : forall (n:nat) (b:bool), D n b -> E b. +Coercion g : D >-> E. +Parameter c : C 0. +Parameter T : E true -> nat. +Check (T c). +Set Printing Coercions. +Check (T c). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + +We give now an example using identity coercions. + +%\begin{small} +\begin{coq_example} +Definition D' (b:bool) := D 1 b. +Identity Coercion IdD'D : D' >-> D. +Print IdD'D. +Parameter d' : D' true. +Check (T d'). +Set Printing Coercions. +Check (T d'). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + + + In the case of functional arguments, we use the monotonic rule of +sub-typing. Approximatively, to coerce $t:forall~x:A, B$ towards +$forall~x:A',B'$, one have to coerce $A'$ towards $A$ and $B$ towards +$B'$. An example is given below: + +%\begin{small} +\begin{coq_example} +Parameters (A B : Set) (h : A -> B). +Coercion h : A >-> B. +Parameter U : (A -> E true) -> nat. +Parameter t : B -> C 0. +Check (U t). +Set Printing Coercions. +Check (U t). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + + Remark the changes in the result following the modification of the +previous example. + +%\begin{small} +\begin{coq_example} +Parameter U' : (C 0 -> B) -> nat. +Parameter t' : E true -> A. +Check (U' t'). +Set Printing Coercions. +Check (U' t'). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + +\item An assumption $x:A$ when $A$ is not a type, is ill-typed. It is + replaced by $x:A'$ where $A'$ is the result of the application + to $A$ of the coercion path between the class of $A$ and {\tt + Sortclass} if it exists. This case occurs in the abstraction + $fun~ x:A => t$, universal quantification $forall~x:A, B$, + global variables and parameters of (co-)inductive definitions + and functions. In $forall~x:A, B$, such a coercion path may be + applied to $B$ also if necessary. + +%\begin{small} +\begin{coq_example} +Parameter Graph : Type. +Parameter Node : Graph -> Type. +Coercion Node : Graph >-> Sortclass. +Parameter G : Graph. +Parameter Arrows : G -> G -> Type. +Check Arrows. +Parameter fg : G -> G. +Check fg. +Set Printing Coercions. +Check fg. +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + +\item $f~a$ is ill-typed because $f:A$ is not a function. The term + $f$ is replaced by the term obtained by applying to $f$ the + coercion path between $A$ and {\tt Funclass} if it exists. + +%\begin{small} +\begin{coq_example} +Parameter bij : Set -> Set -> Set. +Parameter ap : forall A B:Set, bij A B -> A -> B. +Coercion ap : bij >-> Funclass. +Parameter b : bij nat nat. +Check (b 0). +Set Printing Coercions. +Check (b 0). +\end{coq_example} +%\end{small} + +\begin{coq_eval} +Unset Printing Coercions. +\end{coq_eval} + +Let us see the resulting graph of this session. + +%\begin{small} +\begin{coq_example} +Print Graph. +\end{coq_example} +%\end{small} + +\end{itemize} + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Extraction.tex b/doc/refman/Extraction.tex new file mode 100755 index 0000000000..a68969c38a --- /dev/null +++ b/doc/refman/Extraction.tex @@ -0,0 +1,664 @@ +\achapter{Extraction of programs in Objective Caml and Haskell} +\label{Extraction} +\aauthor{Jean-Christophe Filliâtre and Pierre Letouzey} +\index{Extraction} + +\begin{flushleft} + \em The status of extraction is experimental. +\end{flushleft} +We present here the \Coq\ extraction commands, used to build certified +and relatively efficient functional programs, extracting them from the +proofs of their specifications. The functional languages available as +output are currently \ocaml{}, \textsc{Haskell} and \textsc{Scheme}. +In the following, ``ML'' will be used (abusively) to refer to any of +the three. + +\paragraph{Differences with old versions.} +The current extraction mechanism is new for version 7.0 of {\Coq}. +In particular, the \FW\ toplevel used as an intermediate step between +\Coq\ and ML has been withdrawn. It is also not possible +any more to import ML objects in this \FW\ toplevel. +The current mechanism also differs from +the one in previous versions of \Coq: there is no more +an explicit toplevel for the language (formerly called \textsc{Fml}). + +\asection{Generating ML code} +\comindex{Extraction} +\comindex{Recursive Extraction} +\comindex{Extraction Module} +\comindex{Recursive Extraction Module} + +The next two commands are meant to be used for rapid preview of +extraction. They both display extracted term(s) inside \Coq. + +\begin{description} +\item {\tt Extraction \qualid.} ~\par + Extracts one constant or module in the \Coq\ toplevel. + +\item {\tt Recursive Extraction \qualid$_1$ \dots\ \qualid$_n$.} ~\par + Recursive extraction of all the globals (or modules) \qualid$_1$ \dots\ + \qualid$_n$ and all their dependencies in the \Coq\ toplevel. +\end{description} + +%% TODO error messages + +All the following commands produce real ML files. User can choose to produce +one monolithic file or one file per \Coq\ library. + +\begin{description} +\item {\tt Extraction "{\em file}"} + \qualid$_1$ \dots\ \qualid$_n$. ~\par + Recursive extraction of all the globals (or modules) \qualid$_1$ \dots\ + \qualid$_n$ and all their dependencies in one monolithic file {\em file}. + Global and local identifiers are renamed according to the choosen ML + language to fullfill its syntactic conventions, keeping original + names as much as possible. + +\item {\tt Extraction Library} \ident. ~\par + Extraction of the whole \Coq\ library {\tt\ident.v} to an ML module + {\tt\ident.ml}. In case of name clash, identifiers are here renamed + using prefixes \verb!coq_! or \verb!Coq_! to ensure a + session-independent renaming. + +\item {\tt Recursive Extraction Library} \ident. ~\par + Extraction of the \Coq\ library {\tt\ident.v} and all other modules + {\tt\ident.v} depends on. +\end{description} + +The list of globals \qualid$_i$ does not need to be +exhaustive: it is automatically completed into a complete and minimal +environment. + +\asection{Extraction options} + +\asubsection{Setting the target language} +\comindex{Extraction Language} + +The ability to fix target language is the first and more important +of the extraction options. Default is Ocaml. Besides Haskell and +Scheme, another language called Toplevel is provided. It is a pseudo-Ocaml, +with no renaming on global names: so names are printed as in \Coq. +This third language is available only at the \Coq\ Toplevel. +\begin{description} +\item {\tt Extraction Language Ocaml}. +\item {\tt Extraction Language Haskell}. +\item {\tt Extraction Language Scheme}. +\item {\tt Extraction Language Toplevel}. +\end{description} + +\asubsection{Inlining and optimizations} + +Since Objective Caml is a strict language, the extracted +code has to be optimized in order to be efficient (for instance, when +using induction principles we do not want to compute all the recursive +calls but only the needed ones). So the extraction mechanism provides +an automatic optimization routine that will be +called each time the user want to generate Ocaml programs. Essentially, +it performs constants inlining and reductions. Therefore some +constants may not appear in resulting monolithic Ocaml program (a warning is +printed for each such constant). In the case of modular extraction, +even if some inlining is done, the inlined constant are nevertheless +printed, to ensure session-independent programs. + +Concerning Haskell, such optimizations are less useful because of +lazyness. We still make some optimizations, for example in order to +produce more readable code. + +All these optimizations are controled by the following \Coq\ options: + +\begin{description} + +\item \comindex{Set Extraction Optimize} +{\tt Set Extraction Optimize.} + +\item \comindex{Unset Extraction Optimize} +{\tt Unset Extraction Optimize.} + +Default is Set. This control all optimizations made on the ML terms +(mostly reduction of dummy beta/iota redexes, but also simplications on +Cases, etc). Put this option to Unset if you want a ML term as close as +possible to the Coq term. + +\item \comindex{Set Extraction AutoInline} +{\tt Set Extraction AutoInline.} + +\item \comindex{Unset Extraction AutoInline} +{\tt Unset Extraction AutoInline.} + +Default is Set, so by default, the extraction mechanism feels free to +inline the bodies of some defined constants, according to some heuristics +like size of bodies, useness of some arguments, etc. Those heuristics are +not always perfect, you may want to disable this feature, do it by Unset. + +\item \comindex{Extraction Inline} +{\tt Extraction Inline} \qualid$_1$ \dots\ \qualid$_n$. + +\item \comindex{Extraction NoInline} +{\tt Extraction NoInline} \qualid$_1$ \dots\ \qualid$_n$. + +In addition to the automatic inline feature, you can now tell precisely to +inline some more constants by the {\tt Extraction Inline} command. Conversely, +you can forbid the automatic inlining of some specific constants by +the {\tt Extraction NoInline} command. +Those two commands enable a precise control of what is inlined and what is not. + +\item \comindex{Print Extraction Inline} +{\tt Print Extraction Inline}. + +Prints the current state of the table recording the custom inlinings +declared by the two previous commands. + +\item \comindex{Reset Extraction Inline} +{\tt Reset Extraction Inline}. + +Puts the table recording the custom inlinings back to empty. + +\end{description} + + +\paragraph{Inlining and printing of a constant declaration.} + +A user can explicitely asks a constant to be extracted by two means: +\begin{itemize} +\item by mentioning it on the extraction command line +\item by extracting the whole \Coq\ module of this constant. +\end{itemize} +In both cases, the declaration of this constant will be present in the +produced file. +But this same constant may or may not be inlined in the following +terms, depending on the automatic/custom inlining mechanism. + + +For the constants non-explicitely required but needed for dependancy +reasons, there are two cases: +\begin{itemize} +\item If an inlining decision is taken, wether automatically or not, +all occurences of this constant are replaced by its extracted body, and +this constant is not declared in the generated file. +\item If no inlining decision is taken, the constant is normally + declared in the produced file. +\end{itemize} + +\asubsection{Realizing axioms}\label{extraction:axioms} + +Extraction will fail if it encounters an informative +axiom not realized (see section \ref{extraction:axioms}). +A warning will be issued if it encounters an logical axiom, to remind +user that inconsistant logical axioms may lead to incorrect or +non-terminating extracted terms. + +It is possible to assume some axioms while developing a proof. Since +these axioms can be any kind of proposition or object or type, they may +perfectly well have some computational content. But a program must be +a closed term, and of course the system cannot guess the program which +realizes an axiom. Therefore, it is possible to tell the system +what ML term corresponds to a given axiom. + +\comindex{Extract Constant} +\begin{description} +\item{\tt Extract Constant \qualid\ => \str.} ~\par + Give an ML extraction for the given constant. + The \str\ may be an identifier or a quoted string. +\item{\tt Extract Inlined Constant \qualid\ => \str.} ~\par + Same as the previous one, except that the given ML terms will + be inlined everywhere instead of being declared via a let. +\end{description} + +Note that the {\tt Extract Inlined Constant} command is sugar +for an {\tt Extract Constant} followed by a {\tt Extraction Inline}. +Hence a {\tt Reset Extraction Inline} will have an effect on the +realized and inlined xaxiom. + +Of course, it is the responsability of the user to ensure that the ML +terms given to realize the axioms do have the expected types. In +fact, the strings containing realizing code are just copied in the +extracted files. The extraction recognize whether the realized axiom +should become a ML type constant or a ML object declaration. + +\Example +\begin{coq_example} +Axiom X:Set. +Axiom x:X. +Extract Constant X => "int". +Extract Constant x => "0". +\end{coq_example} + +Notice that in the case of type scheme axiom (i.e. whose type is an +arity, that is a sequence of product finished by a sort), then some type +variables has to be given. The syntax is then: + +\begin{description} +\item{\tt Extract Constant \qualid\ \str$_1$ \ldots \str$_n$ => \str.} ~\par +\end{description} + +The number of type variable given is checked by the system. + +\Example +\begin{coq_example} +Axiom Y : Set -> Set -> Set. +Extract Constant Y "'a" "'b" => " 'a*'b ". +\end{coq_example} + +Realizing an axiom via {\tt Extract Constant} is only useful in the +case of an informative axiom (of sort Type or Set). A logical axiom +have no computational content and hence will not appears in extracted +terms. But a warning is nonetheless issued if extraction encounters a +logical axiom. This warning reminds user that inconsistant logical +axioms may lead to incorrect or non-terminating extracted terms. + +If an informative axiom has not been realized before an extraction, a +warning is also issued and the definition of the axiom is filled with +an exception labelled {\tt AXIOM TO BE REALIZED}. The user must then +search these exceptions inside the extracted file and replace them by +real code. + +\comindex{Extract Inductive} + +The system also provides a mechanism to specify ML terms for inductive +types and constructors. For instance, the user may want to use the ML +native boolean type instead of \Coq\ one. The syntax is the following: + +\begin{description} +\item{\tt Extract Inductive \qualid\ => \str\ [ \str\ \dots \str\ ].} ~\par + Give an ML extraction for the given inductive type. You must specify + extractions for the type itself (first \str) and all its + constructors (between square brackets). The ML extraction must be an + ML recursive datatype. +\end{description} + +\Example +Typical examples are the following: +\begin{coq_example} +Extract Inductive unit => "unit" [ "()" ]. +Extract Inductive bool => "bool" [ "true" "false" ]. +Extract Inductive sumbool => "bool" [ "true" "false" ]. +\end{coq_example} + + +\asection{Differences between \Coq\ and ML type systems} + + +Due to differences between \Coq\ and ML type systems, +some extracted programs are not directly typable in ML. +We now solve this problem (at least in Ocaml) by adding +when needed some unsafe casting {\tt Obj.magic}, which give +a generic type {\tt 'a} to any term. + +For example, Here are two kinds of problem that can occur: + +\begin{itemize} + \item If some part of the program is {\em very} polymorphic, there + may be no ML type for it. In that case the extraction to ML works + all right but the generated code may be refused by the ML + type-checker. A very well known example is the {\em distr-pair} + function: +\begin{verbatim} +Definition dp := + fun (A B:Set)(x:A)(y:B)(f:forall C:Set, C->C) => (f A x, f B y). +\end{verbatim} + +In Ocaml, for instance, the direct extracted term would be: + +\begin{verbatim} +let dp x y f = Pair((f () x),(f () y)) +\end{verbatim} + +and would have type: +\begin{verbatim} +dp : 'a -> 'a -> (unit -> 'a -> 'b) -> ('b,'b) prod +\end{verbatim} + +which is not its original type, but a restriction. + +We now produce the following correct version: +\begin{verbatim} +let dp x y f = Pair ((Obj.magic f () x), (Obj.magic f () y)) +\end{verbatim} + + \item Some definitions of \Coq\ may have no counterpart in ML. This + happens when there is a quantification over types inside the type + of a constructor; for example: +\begin{verbatim} +Inductive anything : Set := dummy : forall A:Set, A -> anything. +\end{verbatim} + +which corresponds to the definition of an ML dynamic type. +In Ocaml, we must cast any argument of the constructor dummy. + +\end{itemize} + +Even with those unsafe castings, you should never get error like +``segmentation fault''. In fact even if your program may seem +ill-typed to the Ocaml type-checker, it can't go wrong: it comes +from a Coq well-typed terms, so for example inductives will always +have the correct number of arguments, etc. + +More details about the correctness of the extracted programs can be +found in \cite{Let02}. + +We have to say, though, that in most ``realistic'' programs, these +problems do not occur. For example all the programs of Coq library are +accepted by Caml type-checker without any {\tt Obj.magic} (see examples below). + + + +\asection{Some examples} + +We present here two examples of extractions, taken from the +\Coq\ Standard Library. We choose \ocaml\ as target language, +but all can be done in the other dialects with slight modifications. +We then indicate where to find other examples and tests of Extraction. + +\asubsection{A detailed example: Euclidean division} + +The file {\tt Euclid} contains the proof of Euclidean division +(theorem {\tt eucl\_dev}). The natural numbers defined in the example +files are unary integers defined by two constructors $O$ and $S$: +\begin{coq_example*} +Inductive nat : Set := + | O : nat + | S : nat -> nat. +\end{coq_example*} + +This module contains a theorem {\tt eucl\_dev}, and its extracted term +is of type +\begin{verbatim} +forall b:nat, b > 0 -> forall a:nat, diveucl a b +\end{verbatim} +where {\tt diveucl} is a type for the pair of the quotient and the modulo. +We can now extract this program to \ocaml: + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Require Import Euclid. +Extraction Inline Wf_nat.gt_wf_rec Wf_nat.lt_wf_rec. +Recursive Extraction eucl_dev. +\end{coq_example} + +The inlining of {\tt gt\_wf\_rec} and {\tt lt\_wf\_rec} is not +mandatory. It only enhances readability of extracted code. +You can then copy-paste the output to a file {\tt euclid.ml} or let +\Coq\ do it for you with the following command: + +\begin{coq_example} +Extraction "euclid" eucl_dev. +\end{coq_example} + +Let us play the resulting program: + +\begin{verbatim} +# #use "euclid.ml";; +type sumbool = Left | Right +type nat = O | S of nat +type diveucl = Divex of nat * nat +val minus : nat -> nat -> nat = +val le_lt_dec : nat -> nat -> sumbool = +val le_gt_dec : nat -> nat -> sumbool = +val eucl_dev : nat -> nat -> diveucl = +# eucl_dev (S (S O)) (S (S (S (S (S O)))));; +- : diveucl = Divex (S (S O), S O) +\end{verbatim} +It is easier to test on \ocaml\ integers: +\begin{verbatim} +# let rec i2n = function 0 -> O | n -> S (i2n (n-1));; +val i2n : int -> nat = +# let rec n2i = function O -> 0 | S p -> 1+(n2i p);; +val n2i : nat -> int = +# let div a b = + let Divex (q,r) = eucl_dev (i2n b) (i2n a) in (n2i q, n2i r);; +div : int -> int -> int * int = +# div 173 15;; +- : int * int = 11, 8 +\end{verbatim} + +\asubsection{Another detailed example: Heapsort} + +The file {\tt Heap.v} +contains the proof of an efficient list sorting algorithm described by +Bjerner. Is is an adaptation of the well-known {\em heapsort} +algorithm to functional languages. The main function is {\tt +treesort}, whose type is shown below: + + +\begin{coq_eval} +Reset Initial. +Require Import Relation_Definitions. +Require Import List. +Require Import Sorting. +Require Import Permutation. +\end{coq_eval} +\begin{coq_example} +Require Import Heap. +Check treesort. +\end{coq_example} + +Let's now extract this function: + +\begin{coq_example} +Extraction Inline sort_rec is_heap_rec. +Extraction NoInline list_to_heap. +Extraction "heapsort" treesort. +\end{coq_example} + +One more time, the {\tt Extraction Inline} and {\tt NoInline} +directives are cosmetic. Without it, everything goes right, +but the output is less readable. +Here is the produced file {\tt heapsort.ml}: + +\begin{verbatim} +type nat = + | O + | S of nat + +type 'a sig2 = + 'a + (* singleton inductive, whose constructor was exist2 *) + +type sumbool = + | Left + | Right + +type 'a list = + | Nil + | Cons of 'a * 'a list + +type 'a multiset = + 'a -> nat + (* singleton inductive, whose constructor was Bag *) + +type 'a merge_lem = + 'a list + (* singleton inductive, whose constructor was merge_exist *) + +(** val merge : ('a1 -> 'a1 -> sumbool) -> ('a1 -> 'a1 -> sumbool) -> + 'a1 list -> 'a1 list -> 'a1 merge_lem **) + +let rec merge leA_dec eqA_dec l1 l2 = + match l1 with + | Nil -> l2 + | Cons (a, l) -> + let rec f = function + | Nil -> Cons (a, l) + | Cons (a0, l3) -> + (match leA_dec a a0 with + | Left -> Cons (a, + (merge leA_dec eqA_dec l (Cons (a0, l3)))) + | Right -> Cons (a0, (f l3))) + in f l2 + +type 'a tree = + | Tree_Leaf + | Tree_Node of 'a * 'a tree * 'a tree + +type 'a insert_spec = + 'a tree + (* singleton inductive, whose constructor was insert_exist *) + +(** val insert : ('a1 -> 'a1 -> sumbool) -> ('a1 -> 'a1 -> sumbool) -> + 'a1 tree -> 'a1 -> 'a1 insert_spec **) + +let rec insert leA_dec eqA_dec t a = + match t with + | Tree_Leaf -> Tree_Node (a, Tree_Leaf, Tree_Leaf) + | Tree_Node (a0, t0, t1) -> + let h3 = fun x -> insert leA_dec eqA_dec t0 x in + (match leA_dec a0 a with + | Left -> Tree_Node (a0, t1, (h3 a)) + | Right -> Tree_Node (a, t1, (h3 a0))) + +type 'a build_heap = + 'a tree + (* singleton inductive, whose constructor was heap_exist *) + +(** val list_to_heap : ('a1 -> 'a1 -> sumbool) -> ('a1 -> 'a1 -> + sumbool) -> 'a1 list -> 'a1 build_heap **) + +let rec list_to_heap leA_dec eqA_dec = function + | Nil -> Tree_Leaf + | Cons (a, l0) -> + insert leA_dec eqA_dec (list_to_heap leA_dec eqA_dec l0) a + +type 'a flat_spec = + 'a list + (* singleton inductive, whose constructor was flat_exist *) + +(** val heap_to_list : ('a1 -> 'a1 -> sumbool) -> ('a1 -> 'a1 -> + sumbool) -> 'a1 tree -> 'a1 flat_spec **) + +let rec heap_to_list leA_dec eqA_dec = function + | Tree_Leaf -> Nil + | Tree_Node (a, t0, t1) -> Cons (a, + (merge leA_dec eqA_dec (heap_to_list leA_dec eqA_dec t0) + (heap_to_list leA_dec eqA_dec t1))) + +(** val treesort : ('a1 -> 'a1 -> sumbool) -> ('a1 -> 'a1 -> sumbool) + -> 'a1 list -> 'a1 list sig2 **) + +let treesort leA_dec eqA_dec l = + heap_to_list leA_dec eqA_dec (list_to_heap leA_dec eqA_dec l) + +\end{verbatim} + +Let's test it: +% Format.set_margin 72;; +\begin{verbatim} +# #use "heapsort.ml";; +type sumbool = Left | Right +type nat = O | S of nat +type 'a tree = Tree_Leaf | Tree_Node of 'a * 'a tree * 'a tree +type 'a list = Nil | Cons of 'a * 'a list +val merge : + ('a -> 'a -> sumbool) -> 'b -> 'a list -> 'a list -> 'a list = +val heap_to_list : + ('a -> 'a -> sumbool) -> 'b -> 'a tree -> 'a list = +val insert : + ('a -> 'a -> sumbool) -> 'b -> 'a tree -> 'a -> 'a tree = +val list_to_heap : + ('a -> 'a -> sumbool) -> 'b -> 'a list -> 'a tree = +val treesort : + ('a -> 'a -> sumbool) -> 'b -> 'a list -> 'a list = +\end{verbatim} + +One can remark that the argument of {\tt treesort} corresponding to +{\tt eqAdec} is never used in the informative part of the terms, +only in the logical parts. So the extracted {\tt treesort} never use +it, hence this {\tt 'b} argument. We will use {\tt ()} for this +argument. Only remains the {\tt leAdec} +argument (of type {\tt 'a -> 'a -> sumbool}) to really provide. + +\begin{verbatim} +# let leAdec x y = if x <= y then Left else Right;; +val leAdec : 'a -> 'a -> sumbool = +# let rec listn = function 0 -> Nil + | n -> Cons(Random.int 10000,listn (n-1));; +val listn : int -> int list = +# treesort leAdec () (listn 9);; +- : int list = Cons (160, Cons (883, Cons (1874, Cons (3275, Cons + (5392, Cons (7320, Cons (8512, Cons (9632, Cons (9876, Nil))))))))) +\end{verbatim} + +Some tests on longer lists (10000 elements) show that the program is +quite efficient for Caml code. + + +\asubsection{The Standard Library} + +As a test, we propose an automatic extraction of the +Standard Library of \Coq. In particular, we will find back the +two previous examples, {\tt Euclid} and {\tt Heapsort}. +Go to directory {\tt contrib/extraction/test} of the sources of \Coq, +and run commands: +\begin{verbatim} +make tree; make +\end{verbatim} +This will extract all Standard Library files and compile them. +It is done via many {\tt Extraction Module}, with some customization +(see subdirectory {\tt custom}). + +%The result of this extraction of the Standard Library can be browsed +%at URL +%\begin{flushleft} +%\url{http://www.lri.fr/~letouzey/extraction}. +%\end{flushleft} + +%Reals theory is normally not extracted, since it is an axiomatic +%development. We propose nonetheless a dummy realization of those +%axioms, to test, run: \\ +% +%\mbox{\tt make reals}\\ + +This test works also with Haskell. In the same directory, run: +\begin{verbatim} +make tree; make -f Makefile.haskell +\end{verbatim} +The haskell compiler currently used is {\tt hbc}. Any other should +also work, just adapt the {\tt Makefile.haskell}. In particular {\tt + ghc} is known to work. + +\asubsection{Extraction's horror museum} + +Some pathological examples of extraction are grouped in the file +\begin{verbatim} +contrib/extraction/test_extraction.v +\end{verbatim} +of the sources of \Coq. + +\asubsection{Users' Contributions} + + Several of the \Coq\ Users' Contributions use extraction to produce + certified programs. In particular the following ones have an automatic + extraction test (just run {\tt make} in those directories): + + \begin{itemize} + \item Bordeaux/Additions + \item Bordeaux/EXCEPTIONS + \item Bordeaux/SearchTrees + \item Dyade/BDDS + \item Lannion + \item Lyon/CIRCUITS + \item Lyon/FIRING-SQUAD + \item Marseille/CIRCUITS + \item Muenchen/Higman + \item Nancy/FOUnify + \item Rocq/ARITH/Chinese + \item Rocq/COC + \item Rocq/GRAPHS + \item Rocq/HIGMAN + \item Sophia-Antipolis/Stalmarck + \item Suresnes/BDD + \end{itemize} + + Lannion, Rocq/HIGMAN and Lyon/CIRCUITS are a bit particular. They are + the only examples of developments where {\tt Obj.magic} are needed. + This is probably due to an heavy use of impredicativity. + After compilation those two examples run nonetheless, + thanks to the correction of the extraction~\cite{Let02}. + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Helm.tex b/doc/refman/Helm.tex new file mode 100644 index 0000000000..262c3315d2 --- /dev/null +++ b/doc/refman/Helm.tex @@ -0,0 +1,317 @@ +\label{Helm} +\index{XML exportation} +\index{Proof rendering} + +This section describes the exportation of {\Coq} theories to XML that +has been contributed by Claudio Sacerdoti Coen. Currently, the main +applications are the rendering and searching tool +developed within the HELM\footnote{Hypertextual Electronic Library of +Mathematics} and MoWGLI\footnote{Mathematics on the Web, Get it by +Logic and Interfaces} projects mainly at the University of Bologna and +partly at INRIA-Sophia Antipolis. + +\asubsection{Practical use of the XML exportation tool} + +The basic way to export the logical content of a file into XML format +is to use {\tt coqc} with option {\tt -xml}. +When the {\tt -xml} flag is set, every definition or declaration is +immediately exported to XML once concluded. +The system environment variable {\tt COQ\_XML\_LIBRARY\_ROOT} must be +previously set to a directory in which the logical structure of the +exported objects is reflected. + + For {\tt Makefile} files generated by \verb+coq_makefile+ (see section + \ref{Makefile}), it is sufficient to compile the files using + \begin{quotation} + \verb+make COQ_XML=-xml+ + \end{quotation} + (or, equivalently, setting the environment variable \verb+COQ_XML+) + + To export a development to XML, the suggested procedure is then: + + \begin{enumerate} + \item add to your own contribution a valid \verb+Make+ file and use + \verb+coq_makefile+ to generate the \verb+Makefile+ from the \verb+Make+ + file. + + \Warning Since logical names are used to structure the XML + hierarchy, always add to the \verb+Make+ file at least one \verb+"-R"+ + option to map physical file names to logical module paths. + \item set the \verb+COQ_XML_LIBRARY_ROOT+ environment variable to + the directory where the XML file hierarchy must be physically + rooted. + \item compile your contribution with \verb+"make COQ_XML=-xml"+ + \end{enumerate} + +\Rem In case the system variable {\tt COQ\_XML\_LIBRARY\_ROOT} is not set, +the output is done on the standard output. Also, the files are +compressed using {\tt gzip} after creation. This is to save disk space +since the XML format is very verbose. + +\asubsection{Reflection of the logical structure into the file system} + +For each {\Coq} logical object, several independent files associated +to this object are created. The structure of the long name of the +object is reflected in the directory structure of the file system. +E.g. an object of long name {\tt +{\ident$_1$}.{\ldots}.{\ident$_n$}.{\ident}} is exported to files in the +subdirectory {{\ident$_1$}/{\ldots}/{\ident$_n$}} of the directory +bound to the environment variable {\tt COQ\_XML\_LIBRARY\_ROOT}. + +\asubsection{What is exported?} + +The XML exportation tool exports the logical content of {\Coq} +theories. This covers global definitions (including lemmas, theorems, +...), global assumptions (parameters and axioms), local assumptions or +definitions, and inductive definitions. + +Vernacular files are exported to {\tt .theory.xml} files. +%Variables, +%definitions, theorems, axioms and proofs are exported to individual +%files whose suffixes range from {\tt .var.xml}, {\tt .con.xml}, {\tt +%.con.body.xml}, {\tt .con.types.xml} to {\tt .con.proof_tree.xml}. +Comments are pre-processed with {\sf coqdoc} (see section +\ref{coqdoc}). Especially, they have to be enclosed within {\tt (**} +and {\tt *)} to be exported. + +For each inductive definition of name +{\ident$_1$}.{\ldots}.{\ident$_n$}.{\ident}, a file named {\tt +{\ident}.ind.xml} is created in the subdirectory {\tt +{\ident$_1$}/{\ldots}/{\ident$_n$}} of the xml library root +directory. It contains the arities and constructors of the type. For mutual inductive definitions, the file is named after the +name of the first inductive type of the block. + +For each global definition of base name {\tt +{\ident$_1$}.{\ldots}.{\ident$_n$}.{\ident}}, files named +{\tt {\ident}.con.body.xml} and {\tt {\ident}.con.xml} are created in the +subdirectory {\tt {\ident$_1$}/{\ldots}/{\ident$_n$}}. They +respectively contain the body and the type of the definition. + +For each global assumption of base name {\tt +{\ident$_1$}.{\ident$_2$}.{\ldots}.{\ident$_n$}.{\ident}}, a file +named {\tt {\ident}.con.xml} is created in the subdirectory {\tt +{\ident$_1$}/{\ldots}/{\ident$_n$}}. It contains the type of the +global assumption. + +For each local assumption or definition of base name {\ident} located +in sections {\ident$'_1$}, {\ldots}, {\ident$'_p$} of the module {\tt +{\ident$_1$}.{\ident$_2$}.{\ldots}.{\ident$_n$}.{\ident}}, a file +named {\tt {\ident}.var.xml} is created in the subdirectory {\tt +{\ident$_1$}/{\ldots}/{\ident$_n$}/{\ident$'_1$}/\ldots/{\ident$'_p$}}. +It contains its type and, if a definition, its body. + +In order to do proof-rendering (for example in natural language), some +redundant typing information is required, i.e. the type of at least +some of the subterms of the bodies and types of the CIC objects. These +types are called inner types and are exported to files of suffix {\tt +.types.xml} by the exportation tool. + + +% Deactivated +%% \asubsection{Proof trees} + +%% For each definition or theorem that has been built with tactics, an +%% extra file of suffix {\tt proof\_tree.xml} is created. It contains the +%% proof scripts and is used for rendering the proof. + +\asubsection{Inner types} +\label{inner-types} + +The type of a subterm of a construction is called an {\em inner type} +if it respects the following conditions. + +\begin{enumerate} + \item Its sort is \verb+Prop+\footnote{or {\tt CProp} which is the + "sort"-like definition used in C-CoRN (see + \url{http://vacuumcleaner.cs.kun.nl/c-corn}) to type + computationally relevant predicative propositions.}. + \item It is not a type cast nor an atomic term (variable, constructor or constant). + \item If it's root is an abstraction, then the root's parent node is + not an abstraction, i.e. only the type of the outer abstraction of + a block of nested abstractions is printed. +\end{enumerate} + +The rationale for the 3$^{rd}$ condition is that the type of the inner +abstractions could be easily computed starting from the type of the +outer ones; moreover, the types of the inner abstractions requires a +lot of disk/memory space: removing the 3$^{rd}$ condition leads to XML +file that are two times as big as the ones exported applying the 3$^{rd}$ +condition. + +\asubsection{Interactive exportation commands} + +There are also commands to be used interactively in {\tt coqtop}. + +\asubsubsection{\tt Print XML {\qualid}} +\comindex{Print XML} + +If the variable {\tt COQ\_XML\_LIBRARY\_ROOT} is set, this command creates +files containing the logical content in XML format of {\qualid}. If +the variable is not set, the result is displayed on the standard +output. + +\begin{Variants} +\item {\tt Print XML File {\str} {\qualid}}\\ +This writes the logical content of {\qualid} in XML format to files +whose prefix is {\str}. +\end{Variants} + +\asubsubsection{{\tt Show XML Proof}} +\comindex{Show XML Proof} + +If the variable {\tt COQ\_XML\_LIBRARY\_ROOT} is set, this command creates +files containing the current proof in progress in XML format. It +writes also an XML file made of inner types. If the variable is not +set, the result is displayed on the standard output. + +\begin{Variants} +\item {\tt Show XML File {\str} Proof}\\ This writes the +logical content of {\qualid} in XML format to files whose prefix is +{\str}. +\end{Variants} + +\asubsection{Applications: rendering, searching and publishing} + +The HELM team at the University of Bologna has developed tools +exploiting the XML exportation of {\Coq} libraries. This covers +rendering, searching and publishing tools. + +All these tools require a running http server and, if possible, a +MathML compliant browser. The procedure to install the suite of tools +ultimately allowing rendering and searching can be found on the HELM +web site \url{http://helm.cs.unibo.it/library.html}. + +It may be easier though to upload your developments on the HELM http +server and to re-use the infrastructure running on it. This requires +publishing your development. To this aim, follow the instructions on +\url{http://mowgli.cs.unibo.it}. + +Notice that the HELM server already hosts a copy of the standard +library of {\Coq} and of the {\Coq} user contributions. + +\asubsection{Technical informations} + +\asubsubsection{CIC with Explicit Named Substitutions} + +The exported files are XML encoding of the lambda-terms used by the +\Coq\ system. The implementative details of the \Coq\ system are hidden as much +as possible, so that the XML DTD is a straightforward encoding of the +Calculus of (Co)Inductive Constructions. + +Nevertheless, there is a feature of the \Coq\ system that can not be +hidden in a completely satisfactory way: discharging (see Sect.\ref{Section}). +In \Coq\ it is possible +to open a section, declare variables and use them in the rest of the section +as if they were axiom declarations. Once the section is closed, every definition and theorem in the section is discharged by abstracting it over the section +variables. Variable declarations as well as section declarations are entirely +dropped. Since we are interested in an XML encoding of definitions and +theorems as close as possible to those directly provided the user, we +do not want to export discharged forms. Exporting non-discharged theorem +and definitions together with theorems that rely on the discharged forms +obliges the tools that work on the XML encoding to implement discharging to +achieve logical consistency. Moreover, the rendering of the files can be +misleading, since hyperlinks can be shown between occurrences of the discharge +form of a definition and the non-discharged definition, that are different +objects. + +To overcome the previous limitations, Claudio Sacerdoti Coen developed in his +PhD. thesis an extension of CIC, called Calculus of (Co)Inductive Constructions +with Explicit Named Substitutions, that is a slight extension of CIC where +discharging is not necessary. The DTD of the exported XML files describes +constants, inductive types and variables of the Calculus of (Co)Inductive +Constructions with Explicit Named Substitutions. The conversion to the new +calculus is performed during the exportation phase. + +The following example shows a very small \Coq\ development together with its +version in CIC with Explicit Named Substitutions. + +\begin{verbatim} +# CIC version: # +Section S. + Variable A : Prop. + + Definition impl := A -> A. + + Theorem t : impl. (* uses the undischarged form of impl *) + Proof. + exact (fun (a:A) => a). + Qed. + +End S. + +Theorem t' : (impl False). (* uses the discharged form of impl *) + Proof. + exact (t False). (* uses the discharged form of t *) + Qed. +\end{verbatim} + +\begin{verbatim} +# Corresponding CIC with Explicit Named Substitutions version: # +Section S. + Variable A : Prop. + + Definition impl(A) := A -> A. (* theorems and definitions are + explicitly abstracted over the + variables. The name is sufficient to + completely describe the abstraction *) + + Theorem t(A) : impl. (* impl where A is not instantiated *) + Proof. + exact (fun (a:A) => a). + Qed. + +End S. + +Theorem t'() : impl{False/A}. (* impl where A is instantiated with False + Notice that t' does not depend on A *) + Proof. + exact t{False/A}. (* t where A is instantiated with False *) + Qed. +\end{verbatim} + +Further details on the typing and reduction rules of the calculus can be +found in Claudio Sacerdoti Coen PhD. dissertation, where the consistency +of the calculus is also proved. + +\asubsubsection{The CIC with Explicit Named Substitutions XML DTD} + +A copy of the DTD can be found in the file ``\verb+cic.dtd+'' in the +\verb+contrib/xml+ source directory of \Coq. +The following is a very brief overview of the elements described in the DTD. + +\begin{description} + \item[]\texttt{} + is the root element of the files that correspond to constant types. + \item[]\texttt{} + is the root element of the files that correspond to constant bodies. + It is used only for closed definitions and theorems (i.e. when no + metavariable occurs in the body or type of the constant) + \item[]\texttt{} + is the root element of the file that correspond to the body of a constant + that depends on metavariables (e.g. unfinished proofs) + \item[]\texttt{} + is the root element of the files that correspond to variables + \item[]\texttt{} + is the root element of the files that correspond to blocks + of mutually defined inductive definitions +\end{description} + +The elements + \verb++, \verb++, \verb++, \verb++, \verb++, + \verb++, \verb++, \verb++, \verb++, \verb++, \verb++, \verb++, \verb++, \verb++, + \verb++ and \verb++ are used to encode the constructors of CIC. + The \verb+sort+ or \verb+type+ attribute of the element, if present, is + respectively the sort or the type of the term, that is a sort because of the + typing rules of CIC. + +The element \verb++ correspond to the application of an explicit +named substitution to its first argument, that is a reference to a definition +or declaration in the environment. + +All the other elements are just syntactic sugar. + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Natural.tex b/doc/refman/Natural.tex new file mode 100755 index 0000000000..69dfab87c6 --- /dev/null +++ b/doc/refman/Natural.tex @@ -0,0 +1,425 @@ +\achapter{\texttt{Natural} : proofs in natural language} +\aauthor{Yann Coscoy} + +\asection{Introduction} + +\Natural~ is a package allowing the writing of proofs in natural +language. For instance, the proof in \Coq~of the induction principle on pairs +of natural numbers looks like this: + +\begin{coq_example*} +Require Natural. +\end{coq_example*} +\begin{coq_example} +Print nat_double_ind. +\end{coq_example} + +Piping it through the \Natural~pretty-printer gives: + +\comindex{Print Natural} +\begin{coq_example} +Print Natural nat_double_ind. +\end{coq_example} + +\asection{Activating \Natural} + +To enable the printing of proofs in natural language, you should +type under \texttt{coqtop} or \texttt{coqtop -full} the command + +\begin{coq_example*} +Require Natural. +\end{coq_example*} + +By default, proofs are transcripted in english. If you wish to print them +in French, set the French option by + +\comindex{Set Natural} +\begin{coq_example*} +Set Natural French. +\end{coq_example*} + +If you want to go back to English, type in + +\begin{coq_example*} +Set Natural English. +\end{coq_example*} + +Currently, only \verb=French= and \verb=English= are available. + +You may see for example the natural transcription of the proof of +the induction principle on pairs of natural numbers: + +\begin{coq_example*} +Print Natural nat_double_ind. +\end{coq_example*} + +You may also show in natural language the current proof in progress: + +\comindex{Show Natural} +\begin{coq_example} +Goal (n:nat)(le O n). +Induction n. +Show Natural Proof. +\end{coq_example} + +\subsection*{Restrictions} + +For \Natural, a proof is an object of type a proposition (i.e. an +object of type something of type {\tt Prop}). Only proofs are written +in natural language when typing {\tt Print Natural \ident}. All other +objects (the objects of type something which is of type {\tt Set} or +{\tt Type}) are written as usual $\lambda$-terms. + +\asection{Customizing \Natural} + +The transcription of proofs in natural language is mainly a paraphrase of +the formal proofs, but some specific hints in the transcription +can be given. +Three kinds of customization are available. + +\asubsection{Implicit proof steps} + +\subsubsection*{Implicit lemmas} + +Applying a given lemma or theorem \verb=lem1= of statement, say $A +\Rightarrow B$, to an hypothesis, say $H$ (assuming $A$) produces the +following kind of output translation: + +\begin{verbatim} +... +Using lem1 with H we get B. +... +\end{verbatim} + +But sometimes, you may prefer not to see the explicit invocation to +the lemma. You may prefer to see: + +\begin{verbatim} +... +With H we have A. +... +\end{verbatim} + +This is possible by declaring the lemma as implicit. You should type: + +\comindex{Add Natural} +\begin{coq_example*} +Add Natural Implicit lem1. +\end{coq_example*} + +By default, the lemmas \verb=proj1=, \verb=proj2=, \verb=sym_equal= +and \verb=sym_eqT= are declared implicit. To remove a lemma or a theorem +previously declared as implicit, say \verb=lem1=, use the command + +\comindex{Remove Natural} +\begin{coq_example*} +Remove Natural Implicit lem1. +\end{coq_example*} + +To test if the lemma or theorem \verb=lem1= is, or is not, +declared as implicit, type + +\comindex{Test Natural} +\begin{coq_example*} +Test Natural Implicit lem1. +\end{coq_example*} + +\subsubsection*{Implicit proof constructors} + +Let \verb=constr1= be a proof constructor of a given inductive +proposition (or predicate) +\verb=Q= (of type \verb=Prop=). Assume \verb=constr1= proves +\verb=(x:A)(P x)->(Q x)=. Then, applying \verb=constr1= to an hypothesis, +say \verb=H= (assuming \verb=(P a)=) produces the following kind of output: + +\begin{verbatim} +... +By the definition of Q, with H we have (Q a). +... +\end{verbatim} + +But sometimes, you may prefer not to see the explicit invocation to +this constructor. You may prefer to see: + +\begin{verbatim} +... +With H we have (Q a). +... +\end{verbatim} + +This is possible by declaring the constructor as implicit. You should +type, as before: + +\comindex{Add Natural Implicit} +\begin{coq_example*} +Add Natural Implicit constr1. +\end{coq_example*} + +By default, the proposition (or predicate) constructors + +\verb=conj=, \verb=or_introl=, \verb=or_intror=, \verb=ex_intro=, +\verb=exT_intro=, \verb=refl_equal=, \verb=refl_eqT= and \verb=exist= + +\noindent are declared implicit. Note that declaring implicit the +constructor of a datatype (i.e. an inductive type of type \verb=Set=) +has no effect. + +As above, you can remove or test a constant declared implicit. + +\subsubsection*{Implicit inductive constants} + +Let \verb=Ind= be an inductive type (either a proposition (or a +predicate) -- on \verb=Prop= --, or a datatype -- on \verb=Set=). +Suppose the proof proceeds by induction on an hypothesis \verb=h= +proving \verb=Ind= (or more generally \verb=(Ind A1 ... An)=). The +following kind of output is produced: + +\begin{verbatim} +... +With H, we will prove A by induction on the definition of Ind. +Case 1. ... +Case 2. ... +... +\end{verbatim} + +But sometimes, you may prefer not to see the explicit invocation to +\verb=Ind=. You may prefer to see: + +\begin{verbatim} +... +We will prove A by induction on H. +Case 1. ... +Case 2. ... +... +\end{verbatim} + +This is possible by declaring the inductive type as implicit. You should +type, as before: + +\comindex{Add Natural Implicit} +\begin{coq_example*} +Add Natural Implicit Ind. +\end{coq_example*} + +This kind of parameterization works for any inductively defined +proposition (or predicate) or datatype. Especially, it works whatever +the definition is recursive or purely by cases. + +By default, the data type \verb=nat= and the inductive connectives +\verb=and=, \verb=or=, \verb=sig=, \verb=False=, \verb=eq=, +\verb=eqT=, \verb=ex= and \verb=exT= are declared implicit. + +As above, you can remove or test a constant declared implicit. Use +{\tt Remove Natural Contractible $id$} or {\tt Test Natural +Contractible $id$}. + +\asubsection{Contractible proof steps} + +\subsubsection*{Contractible lemmas or constructors} + +Some lemmas, theorems or proof constructors of inductive predicates are +often applied in a row and you obtain an output of this kind: + +\begin{verbatim} +... +Using T with H1 and H2 we get P. + * By H3 we have Q. + Using T with theses results we get R. +... +\end{verbatim} + +where \verb=T=, \verb=H1=, \verb=H2= and \verb=H3= prove statements +of the form \verb=(X,Y:Prop)X->Y->(L X Y)=, \verb=A=, \verb=B= and \verb=C= +respectively (and thus \verb=R= is \verb=(L (L A B) C)=). + +You may obtain a condensed output of the form + +\begin{verbatim} +... +Using T with H1, H2, and H3 we get R. +... +\end{verbatim} + +by declaring \verb=T= as contractible: + +\comindex{Add Natural Contractible} +\begin{coq_example*} +Add Natural Contractible T. +\end{coq_example*} + +By default, the lemmas \verb=proj1=, \verb=proj2= and the proof +constructors \verb=conj=, \verb=or_introl=, \verb=or_intror= are +declared contractible. As for implicit notions, you can remove or +test a lemma or constructor declared contractible. + +\subsubsection*{Contractible induction steps} + +Let \verb=Ind= be an inductive type. When the proof proceeds by +induction in a row, you may obtain an output of this kind: + +\begin{verbatim} +... +We have (Ind A (Ind B C)). +We use definition of Ind in a study in two cases. +Case 1: We have A. +Case 2: We have (Ind B C). + We use definition of Ind in a study of two cases. + Case 2.1: We have B. + Case 2.2: We have C. +... +\end{verbatim} + +You may prefer to see + +\begin{verbatim} +... +We have (Ind A (Ind B C)). +We use definition of Ind in a study in three cases. +Case 1: We have A. +Case 2: We have B. +Case 3: We have C. +... +\end{verbatim} + +This is possible by declaring \verb=Ind= as contractible: + +\begin{coq_example*} +Add Natural Contractible T. +\end{coq_example*} + +By default, only \verb=or= is declared as a contractible inductive +constant. +As for implicit notions, you can remove or test an inductive notion declared +contractible. + +\asubsection{Transparent definitions} + +``Normal'' definitions are all constructions except proofs and proof constructors. + +\subsubsection*{Transparent non inductive normal definitions} + +When using the definition of a non inductive constant, say \verb=D=, the +following kind of output is produced: + +\begin{verbatim} +... +We have proved C which is equivalent to D. +... +\end{verbatim} + +But you may prefer to hide that D comes from the definition of C as +follows: + +\begin{verbatim} +... +We have prove D. +... +\end{verbatim} + +This is possible by declaring \verb=C= as transparent: + +\comindex{Add Natural Transparent} +\begin{coq_example*} +Add Natural Transparent D. +\end{coq_example*} + +By default, only \verb=not= (normally written \verb=~=) is declared as +a non inductive transparent definition. +As for implicit and contractible definitions, you can remove or test a +non inductive definition declared transparent. +Use \texttt{Remove Natural Transparent} \ident or +\texttt{Test Natural Transparent} \ident. + +\subsubsection*{Transparent inductive definitions} + +Let \verb=Ind= be an inductive proposition (more generally: a +predicate \verb=(Ind x1 ... xn)=). Suppose the definition of +\verb=Ind= is non recursive and built with just +one constructor proving something like \verb=A -> B -> Ind=. +When coming back to the definition of \verb=Ind= the +following kind of output is produced: + +\begin{verbatim} +... +Assume Ind (H). + We use H with definition of Ind. + We have A and B. + ... +\end{verbatim} + +When \verb=H= is not used a second time in the proof, you may prefer +to hide that \verb=A= and \verb=B= comes from the definition of +\verb=Ind=. You may prefer to get directly: + +\begin{verbatim} +... +Assume A and B. +... +\end{verbatim} + +This is possible by declaring \verb=Ind= as transparent: + +\begin{coq_example*} +Add Natural Transparent Ind. +\end{coq_example*} + +By default, \verb=and=, \verb=or=, \verb=ex=, \verb=exT=, \verb=sig= +are declared as inductive transparent constants. As for implicit and +contractible constants, you can remove or test an inductive +constant declared transparent. + +As for implicit and contractible constants, you can remove or test an +inductive constant declared transparent. + +\asubsection{Extending the maximal depth of nested text} + +The depth of nested text is limited. To know the current depth, do: + +\comindex{Set Natural Depth} +\begin{coq_example} +Set Natural Depth. +\end{coq_example} + +To change the maximal depth of nested text (for instance to 125) do: + +\begin{coq_example} +Set Natural Depth 125. +\end{coq_example} + +\asubsection{Restoring the default parameterization} + +The command \verb=Set Natural Default= sets back the parameterization tables of +\Natural~ to their default values, as listed in the above sections. +Moreover, the language is set back to English and the max depth of +nested text is set back to its initial value. + +\asubsection{Printing the current parameterization} + +The commands {\tt Print Natural Implicit}, {\tt Print Natural +Contractible} and {\tt Print \\ Natural Transparent} print the list of +constructions declared {\tt Implicit}, {\tt Contractible}, +{\tt Transparent} respectively. + +\asubsection{Interferences with \texttt{Reset}} + +The customization of \texttt{Natural} is dependent of the \texttt{Reset} +command. If you reset the environment back to a point preceding an +\verb=Add Natural ...= command, the effect of the command will be +erased. Similarly, a reset back to a point before a +\verb=Remove Natural ... = command invalidates the removal. + +\asection{Error messages} + +An error occurs when trying to \verb=Print=, to \verb=Add=, to +\verb=Test=, or to \verb=remove= an undefined ident. Similarly, an +error occurs when trying to set a language unknown from \Natural. +Errors may also occur when trying to parameterize the printing of +proofs: some parameterization are effectively forbidden. +Note that to \verb=Remove= an ident absent from a table or to +\verb=Add= to a table an already present ident does not lead to an +error. + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Omega.tex b/doc/refman/Omega.tex new file mode 100755 index 0000000000..bbf17f6304 --- /dev/null +++ b/doc/refman/Omega.tex @@ -0,0 +1,226 @@ +\achapter{Omega: a solver of quantifier-free problems in +Presburger Arithmetic} +\aauthor{Pierre Crégut} +\label{OmegaChapter} + +\asection{Description of {\tt omega}} +\tacindex{omega} +\label{description} + +{\tt omega} solves a goal in Presburger arithmetic, i.e. a universally +quantified formula made of equations and inequations. Equations may +be specified either on the type \verb=nat= of natural numbers or on +the type \verb=Z= of binary-encoded integer numbers. Formulas on +\verb=nat= are automatically injected into \verb=Z=. The procedure +may use any hypothesis of the current proof session to solve the goal. + +Multiplication is handled by {\tt omega} but only goals where at +least one of the two multiplicands of products is a constant are +solvable. This is the restriction meaned by ``Presburger arithmetic''. + +If the tactic cannot solve the goal, it fails with an error message. +In any case, the computation eventually stops. + +\asubsection{Arithmetical goals recognized by {\tt omega}} + +{\tt omega} applied only to quantifier-free formulas built from the +connectors + +\begin{quote} +\verb=/\, \/, ~, ->= +\end{quote} + +on atomic formulas. Atomic formulas are built from the predicates + +\begin{quote} +\verb!=, le, lt, gt, ge! +\end{quote} + + on \verb=nat= or from the predicates + +\begin{quote} +\verb!=, <, <=, >, >=! +\end{quote} + + on \verb=Z=. In expressions of type \verb=nat=, {\tt omega} recognizes + +\begin{quote} +\verb!plus, minus, mult, pred, S, O! +\end{quote} + +and in expressions of type \verb=Z=, {\tt omega} recognizes + +\begin{quote} +\verb!+, -, *, Zsucc!, and constants. +\end{quote} + +All expressions of type \verb=nat= or \verb=Z= not built on these +operators are considered abstractly as if they +were arbitrary variables of type \verb=nat= or \verb=Z=. + +\asubsection{Messages from {\tt omega}} +\label{errors} + +When {\tt omega} does not solve the goal, one of the following errors +is generated: + +\begin{ErrMsgs} + +\item \errindex{omega can't solve this system} + + This may happen if your goal is not quantifier-free (if it is + universally quantified, try {\tt intros} first; if it contains + existentials quantifiers too, {\tt omega} is not strong enough to solve your + goal). This may happen also if your goal contains arithmetical + operators unknown from {\tt omega}. Finally, your goal may be really + wrong! + +\item \errindex{omega: Not a quantifier-free goal} + + If your goal is universally quantified, you should first apply {\tt + intro} as many time as needed. + +\item \errindex{omega: Unrecognized predicate or connective: {\sl ident}} + +\item \errindex{omega: Unrecognized atomic proposition: {\sl prop}} + +\item \errindex{omega: Can't solve a goal with proposition variables} + +\item \errindex{omega: Unrecognized proposition} + +\item \errindex{omega: Can't solve a goal with non-linear products} + +\item \errindex{omega: Can't solve a goal with equality on {\sl type}} + +\end{ErrMsgs} + +%% Ce code est débranché pour l'instant +%% +% \asubsection{Control over the output} +% There are some flags that can be set to get more information on the procedure + +% \begin{itemize} +% \item \verb=Time= to get the time used by the procedure +% \item \verb=System= to visualize the normalized systems. +% \item \verb=Action= to visualize the actions performed by the OMEGA +% procedure (see \ref{technical}). +% \end{itemize} + +% \comindex{Set omega Time} +% \comindex{UnSet omega Time} +% \comindex{Switch omega Time} +% \comindex{Set omega System} +% \comindex{UnSet omega System} +% \comindex{Switch omega System} +% \comindex{Set omega Action} +% \comindex{UnSet omega Action} +% \comindex{Switch omega Action} + +% Use {\tt Set omega {\rm\sl flag}} to set the flag +% {\rm\sl flag}. Use {\tt Unset omega {\rm\sl flag}} to unset it and +% {\tt Switch omega {\rm\sl flag}} to toggle it. + +\section{Using {\tt omega}} + +The {\tt omega} tactic does not belong to the core system. It should be +loaded by +\begin{coq_example*} +Require Import Omega. +Open Scope Z_scope. +\end{coq_example*} + +\example{} + +\begin{coq_example} +Goal forall m n:Z, 1 + 2 * m <> 2 * n. +intros; omega. +\end{coq_example} +\begin{coq_eval} +Abort. +\end{coq_eval} + +\example{} + +\begin{coq_example} +Goal forall z:Z, z > 0 -> 2 * z + 1 > z. +intro; omega. +\end{coq_example} + +% Other examples can be found in \verb+$COQLIB/theories/DEMOS/OMEGA+. + +\asection{Technical data} +\label{technical} + +\asubsection{Overview of the tactic} +\begin{itemize} + +\item The goal is negated twice and the first negation is introduced as an + hypothesis. +\item Hypothesis are decomposed in simple equations or inequations. Multiple + goals may result from this phase. +\item Equations and inequations over \verb=nat= are translated over + \verb=Z=, multiple goals may result from the translation of + substraction. +\item Equations and inequations are normalized. +\item Goals are solved by the {\it OMEGA} decision procedure. +\item The script of the solution is replayed. + +\end{itemize} + +\asubsection{Overview of the {\it OMEGA} decision procedure} + +The {\it OMEGA} decision procedure involved in the {\tt omega} tactic uses +a small subset of the decision procedure presented in + +\begin{quote} + "The Omega Test: a fast and practical integer programming +algorithm for dependence analysis", William Pugh, Communication of the +ACM , 1992, p 102-114. +\end{quote} + +Here is an overview, look at the original paper for more information. + +\begin{itemize} + +\item Equations and inequations are normalized by division by the GCD of their + coefficients. +\item Equations are eliminated, using the Banerjee test to get a coefficient + equal to one. +\item Note that each inequation defines a half space in the space of real value + of the variables. + \item Inequations are solved by projecting on the hyperspace + defined by cancelling one of the variable. They are partitioned + according to the sign of the coefficient of the eliminated + variable. Pairs of inequations from different classes define a + new edge in the projection. + \item Redundant inequations are eliminated or merged in new + equations that can be eliminated by the Banerjee test. +\item The last two steps are iterated until a contradiction is reached + (success) or there is no more variable to eliminate (failure). + +\end{itemize} + +It may happen that there is a real solution and no integer one. The last +steps of the Omega procedure (dark shadow) are not implemented, so the +decision procedure is only partial. + +\asection{Bugs} + +\begin{itemize} +\item The simplification procedure is very dumb and this results in + many redundant cases to explore. + +\item Much too slow. + +\item Certainly other bugs! You can report them to + +\begin{quote} + \url{Pierre.Cregut@cnet.francetelecom.fr} +\end{quote} + +\end{itemize} + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/Polynom.tex b/doc/refman/Polynom.tex new file mode 100644 index 0000000000..70889c9d45 --- /dev/null +++ b/doc/refman/Polynom.tex @@ -0,0 +1,504 @@ +\achapter{The \texttt{ring} tactic} +\aauthor{Patrick Loiseleur and Samuel Boutin} +\label{ring} +\tacindex{ring} + +This chapter presents the \texttt{ring} tactic. + +\asection{What does this tactic?} + +\texttt{ring} does associative-commutative rewriting in ring and semi-ring +structures. Assume you have two binary functions $\oplus$ and $\otimes$ +that are associative and commutative, with $\oplus$ distributive on +$\otimes$, and two constants 0 and 1 that are unities for $\oplus$ and +$\otimes$. A \textit{polynomial} is an expression built on variables $V_0, V_1, +\dots$ and constants by application of $\oplus$ and $\otimes$. + +Let an {\it ordered product} be a product of variables $V_{i_1} \otimes +\ldots \otimes V_{i_n}$ verifying $i_1 \le i_2 \le \dots \le i_n$. Let a +\textit{monomial} be the product of a constant (possibly equal to 1, in +which case we omit it) and an ordered product. We can order the +monomials by the lexicographic order on products of variables. Let a +\textit{canonical sum} be an ordered sum of monomials that are all +different, i.e. each monomial in the sum is strictly less than the +following monomial according to the lexicographic order. It is an easy +theorem to show that every polynomial is equivalent (modulo the ring +properties) to exactly one canonical sum. This canonical sum is called +the \textit{normal form} of the polynomial. So what does \texttt{ring}? It +normalizes polynomials over any ring or semi-ring structure. The basic +use of \texttt{ring} is to simplify ring expressions, so that the user +does not have to deal manually with the theorems of associativity and +commutativity. + +\begin{Examples} +\item In the ring of integers, the normal form of +$x (3 + yx + 25(1 - z)) + zx$ is $28x + (-24)xz + xxy$. +\item For the classical propositional calculus (or the boolean rings) + the normal form is what logicians call \textit{disjunctive normal + form}: every formula is equivalent to a disjunction of + conjunctions of atoms. (Here $\oplus$ is $\vee$, $\otimes$ is + $\wedge$, variables are atoms and the only constants are T and F) +\end{Examples} + +\asection{The variables map} + +It is frequent to have an expression built with + and + $\times$, but rarely on variables only. +Let us associate a number to each subterm of a ring +expression in the \gallina\ language. For example in the ring +\texttt{nat}, consider the expression: + +\begin{quotation} +\begin{verbatim} +(plus (mult (plus (f (5)) x) x) + (mult (if b then (4) else (f (3))) (2))) +\end{verbatim} +\end{quotation} + +\noindent As a ring expression, is has 3 subterms. Give each subterm a +number in an arbitrary order: + +\begin{tabular}{ccl} +0 & $\mapsto$ & \verb|if b then (4) else (f (3))| \\ +1 & $\mapsto$ & \verb|(f (5))| \\ +2 & $\mapsto$ & \verb|x| \\ +\end{tabular} + +\noindent Then normalize the ``abstract'' polynomial + +$$((V_1 \otimes V_2) \oplus V_2) \oplus (V_0 \otimes 2) $$ + +\noindent In our example the normal form is: + +$$(2 \otimes V_0) \oplus (V_1 \otimes V_2) \oplus (V_2 \otimes V_2)$$ + +\noindent Then substitute the variables by their values in the variables map to +get the concrete normal polynomial: + +\begin{quotation} +\begin{verbatim} +(plus (mult (2) (if b then (4) else (f (3)))) + (plus (mult (f (5)) x) (mult x x))) +\end{verbatim} +\end{quotation} + +\asection{Is it automatic?} + +Yes, building the variables map and doing the substitution after +normalizing is automatically done by the tactic. So you can just forget +this paragraph and use the tactic according to your intuition. + +\asection{Concrete usage in \Coq} + +Under a session launched by \texttt{coqtop} or \texttt{coqtop -full}, +load the \texttt{ring} files with the command: + +\begin{quotation} +\begin{verbatim} +Require Ring. +\end{verbatim} +\end{quotation} + +It does not work under \texttt{coqtop -opt} because the compiled ML +objects used by the tactic are not linked in this binary image, and +dynamic loading of native code is not possible in \ocaml. + +In order to use \texttt{ring} on naturals, load \texttt{ArithRing} +instead; for binary integers, load the module \texttt{ZArithRing}. + +Then, to normalize the terms $term_1$, \dots, $term_n$ in +the current subgoal, use the tactic: + +\begin{quotation} +\texttt{ring} $term_1$ \dots{} $term_n$ +\end{quotation} +\tacindex{ring} + +Then the tactic guesses the type of given terms, the ring theory to +use, the variables map, and replace each term with its normal form. +The variables map is common to all terms + +\Warning \texttt{ring $term_1$; ring $term_2$} is not equivalent to +\texttt{ring $term_1$ $term_2$}. In the latter case the variables map +is shared between the two terms, and common subterm $t$ of $term_1$ +and $term_2$ will have the same associated variable number. + +\begin{ErrMsgs} +\item \errindex{All terms must have the same type} +\item \errindex{Don't know what to do with this goal} +\item \errindex{No Declared Ring Theory for \term.} + + \texttt{Use Add [Semi] Ring to declare it} + + That happens when all terms have the same type \term, but there is + no declared ring theory for this set. See below. +\end{ErrMsgs} + +\begin{Variants} +\item \texttt{ring} + + That works if the current goal is an equality between two + polynomials. It will normalize both sides of the + equality, solve it if the normal forms are equal and in other cases + try to simplify the equality using \texttt{congr\_eqT} and \texttt{refl\_equal} + to reduce $x + y = x + z$ to $y = z$ and $x * z = x * y$ to $y = z$. + + \ErrMsg\errindex{This goal is not an equality} + +\end{Variants} + +\asection{Add a ring structure} + +It can be done in the \Coq toplevel (No ML file to edit and to link +with \Coq). First, \texttt{ring} can handle two kinds of structure: +rings and semi-rings. Semi-rings are like rings without an opposite to +addition. Their precise specification (in \gallina) can be found in +the file + +\begin{quotation} +\begin{verbatim} +contrib/ring/Ring_theory.v +\end{verbatim} +\end{quotation} + +The typical example of ring is \texttt{Z}, the typical +example of semi-ring is \texttt{nat}. + +The specification of a +ring is divided in two parts: first the record of constants +($\oplus$, $\otimes$, 1, 0, $\ominus$) and then the theorems +(associativity, commutativity, etc.). + +\begin{small} +\begin{flushleft} +\begin{verbatim} +Section Theory_of_semi_rings. + +Variable A : Type. +Variable Aplus : A -> A -> A. +Variable Amult : A -> A -> A. +Variable Aone : A. +Variable Azero : A. +(* There is also a "weakly decidable" equality on A. That means + that if (A_eq x y)=true then x=y but x=y can arise when + (A_eq x y)=false. On an abstract ring the function [x,y:A]false + is a good choice. The proof of A_eq_prop is in this case easy. *) +Variable Aeq : A -> A -> bool. + +Record Semi_Ring_Theory : Prop := +{ SR_plus_sym : (n,m:A)[| n + m == m + n |]; + SR_plus_assoc : (n,m,p:A)[| n + (m + p) == (n + m) + p |]; + + SR_mult_sym : (n,m:A)[| n*m == m*n |]; + SR_mult_assoc : (n,m,p:A)[| n*(m*p) == (n*m)*p |]; + SR_plus_zero_left :(n:A)[| 0 + n == n|]; + SR_mult_one_left : (n:A)[| 1*n == n |]; + SR_mult_zero_left : (n:A)[| 0*n == 0 |]; + SR_distr_left : (n,m,p:A) [| (n + m)*p == n*p + m*p |]; + SR_plus_reg_left : (n,m,p:A)[| n + m == n + p |] -> m==p; + SR_eq_prop : (x,y:A) (Is_true (Aeq x y)) -> x==y +}. +\end{verbatim} +\end{flushleft} +\end{small} + +\begin{small} +\begin{flushleft} +\begin{verbatim} +Section Theory_of_rings. + +Variable A : Type. + +Variable Aplus : A -> A -> A. +Variable Amult : A -> A -> A. +Variable Aone : A. +Variable Azero : A. +Variable Aopp : A -> A. +Variable Aeq : A -> A -> bool. + + +Record Ring_Theory : Prop := +{ Th_plus_sym : (n,m:A)[| n + m == m + n |]; + Th_plus_assoc : (n,m,p:A)[| n + (m + p) == (n + m) + p |]; + Th_mult_sym : (n,m:A)[| n*m == m*n |]; + Th_mult_assoc : (n,m,p:A)[| n*(m*p) == (n*m)*p |]; + Th_plus_zero_left :(n:A)[| 0 + n == n|]; + Th_mult_one_left : (n:A)[| 1*n == n |]; + Th_opp_def : (n:A) [| n + (-n) == 0 |]; + Th_distr_left : (n,m,p:A) [| (n + m)*p == n*p + m*p |]; + Th_eq_prop : (x,y:A) (Is_true (Aeq x y)) -> x==y +}. +\end{verbatim} +\end{flushleft} +\end{small} + +To define a ring structure on A, you must provide an addition, a +multiplication, an opposite function and two unities 0 and 1. + +You must then prove all theorems that make +(A,Aplus,Amult,Aone,Azero,Aeq) +a ring structure, and pack them with the \verb|Build_Ring_Theory| +constructor. + +Finally to register a ring the syntax is: + +\comindex{Add Ring} +\begin{quotation} + \texttt{Add Ring} \textit{A Aplus Amult Aone Azero Ainv Aeq T} + \texttt{[} \textit{c1 \dots cn} \texttt{].} +\end{quotation} + +\noindent where \textit{A} is a term of type \texttt{Set}, +\textit{Aplus} is a term of type \texttt{A->A->A}, +\textit{Amult} is a term of type \texttt{A->A->A}, +\textit{Aone} is a term of type \texttt{A}, +\textit{Azero} is a term of type \texttt{A}, +\textit{Ainv} is a term of type \texttt{A->A}, +\textit{Aeq} is a term of type \texttt{A->bool}, +\textit{T} is a term of type +\texttt{(Ring\_Theory }\textit{A Aplus Amult Aone Azero Ainv + Aeq}\texttt{)}. +The arguments \textit{c1 \dots cn}, +are the names of constructors which define closed terms: a +subterm will be considered as a constant if it is either one of the +terms \textit{c1 \dots cn} or the application of one of these terms to +closed terms. For \texttt{nat}, the given constructors are \texttt{S} +and \texttt{O}, and the closed terms are \texttt{O}, \texttt{(S O)}, +\texttt{(S (S O))}, \ldots + +\begin{Variants} +\item \texttt{Add Semi Ring} \textit{A Aplus Amult Aone Azero Aeq T} + \texttt{[} \textit{c1 \dots\ cn} \texttt{].}\comindex{Add Semi + Ring} + + There are two differences with the \texttt{Add Ring} command: there + is no inverse function and the term $T$ must be of type + \texttt{(Semi\_Ring\_Theory }\textit{A Aplus Amult Aone Azero + Aeq}\texttt{)}. + +\item \texttt{Add Abstract Ring} \textit{A Aplus Amult Aone Azero Ainv + Aeq T}\texttt{.}\comindex{Add Abstract Ring} + + This command should be used for when the operations of rings are not + computable; for example the real numbers of + \texttt{theories/REALS/}. Here $0+1$ is not beta-reduced to $1$ but + you still may want to \textit{rewrite} it to $1$ using the ring + axioms. The argument \texttt{Aeq} is not used; a good choice for + that function is \verb+[x:A]false+. + +\item \texttt{Add Abstract Semi Ring} \textit{A Aplus Amult Aone Azero + Aeq T}\texttt{.}\comindex{Add Abstract Semi Ring} + +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{Not a valid (semi)ring theory}. + + That happens when the typing condition does not hold. +\end{ErrMsgs} + +Currently, the hypothesis is made than no more than one ring structure +may be declared for a given type in \texttt{Set} or \texttt{Type}. +This allows automatic detection of the theory used to achieve the +normalization. On popular demand, we can change that and allow several +ring structures on the same set. + +The table of ring theories is compatible with the \Coq\ +sectioning mechanism. If you declare a ring inside a section, the +declaration will be thrown away when closing the section. +And when you load a compiled file, all the \texttt{Add Ring} +commands of this file that are not inside a section will be loaded. + +The typical example of ring is \texttt{Z}, and the typical example of +semi-ring is \texttt{nat}. Another ring structure is defined on the +booleans. + +\Warning Only the ring of booleans is loaded by default with the +\texttt{Ring} module. To load the ring structure for \texttt{nat}, +load the module \texttt{ArithRing}, and for \texttt{Z}, +load the module \texttt{ZArithRing}. + + +\asection{How does it work?} + +The code of \texttt{ring} is a good example of tactic written using +\textit{reflection} (or \textit{internalization}, it is synonymous). +What is reflection? Basically, it is writing \Coq{} tactics in \Coq, +rather than in \ocaml. From the philosophical point of view, it is +using the ability of the Calculus of Constructions to speak and reason +about itself. For the \texttt{ring} tactic we used \Coq\ as a +programming language and also as a proof environment to build a tactic +and to prove it correctness. + +The interested reader is strongly advised to have a look at the file +\texttt{Ring\_normalize.v}. Here a type for polynomials is defined: + +\begin{small} +\begin{flushleft} +\begin{verbatim} +Inductive Type polynomial := + Pvar : idx -> polynomial +| Pconst : A -> polynomial +| Pplus : polynomial -> polynomial -> polynomial +| Pmult : polynomial -> polynomial -> polynomial +| Popp : polynomial -> polynomial. +\end{verbatim} +\end{flushleft} +\end{small} + +There is also a type to represent variables maps, and an +interpretation function, that maps a variables map and a polynomial to an +element of the concrete ring: + +\begin{small} +\begin{flushleft} +\begin{verbatim} +Definition polynomial_simplify := [...] +Definition interp : (varmap A) -> (polynomial A) -> A := [...] +\end{verbatim} +\end{flushleft} +\end{small} + +A function to normalize polynomials is defined, and the big theorem is +its correctness w.r.t interpretation, that is: + +\begin{small} +\begin{flushleft} +\begin{verbatim} +Theorem polynomial_simplify_correct : forall (v:(varmap A))(p:polynomial) + (interp v (polynomial_simplify p)) + =(interp v p). +\end{verbatim} +\end{flushleft} +\end{small} + +(The actual code is slightly more complex: for efficiency, +there is a special datatype to represent normalized polynomials, +i.e. ``canonical sums''. But the idea is still the same). + +So now, what is the scheme for a normalization proof? Let \texttt{p} +be the polynomial expression that the user wants to normalize. First a +little piece of ML code guesses the type of \texttt{p}, the ring +theory \texttt{T} to use, an abstract polynomial \texttt{ap} and a +variables map \texttt{v} such that \texttt{p} is +$\beta\delta\iota$-equivalent to \verb|(interp v ap)|. Then we +replace it by \verb|(interp v (polynomial_simplify ap))|, using the +main correctness theorem and we reduce it to a concrete expression +\texttt{p'}, which is the concrete normal form of +\texttt{p}. This is summarized in this diagram: +\begin{center} +\begin{tabular}{rcl} +\texttt{p} & $\rightarrow_{\beta\delta\iota}$ + & \texttt{(interp v ap)} \\ + & & $=_{\mathrm{(by\ the\ main\ correctness\ theorem)}}$ \\ +\texttt{p'} + & $\leftarrow_{\beta\delta\iota}$ + & \texttt{(interp v (polynomial\_simplify ap))} +\end{tabular} +\end{center} +The user do not see the right part of the diagram. +From outside, the tactic behaves like a +$\beta\delta\iota$ simplification extended with AC rewriting rules. +Basically, the proof is only the application of the main +correctness theorem to well-chosen arguments. + +\asection{History of \texttt{ring}} + +First Samuel Boutin designed the tactic \texttt{ACDSimpl}. +This tactic did lot of rewriting. But the proofs +terms generated by rewriting were too big for \Coq's type-checker. +Let us see why: + +\begin{coq_eval} +Require Import ZArith. +Open Scope Z_scope. +\end{coq_eval} +\begin{coq_example} +Goal forall x y z:Z, x + 3 + y + y * z = x + 3 + y + z * y. +\end{coq_example} +\begin{coq_example*} +intros; rewrite (Zmult_comm y z); reflexivity. +Save toto. +\end{coq_example*} +\begin{coq_example} +Print toto. +\end{coq_example} + +At each step of rewriting, the whole context is duplicated in the proof +term. Then, a tactic that does hundreds of rewriting generates huge proof +terms. Since \texttt{ACDSimpl} was too slow, Samuel Boutin rewrote it +using reflection (see his article in TACS'97 \cite{Bou97}). Later, the +stuff was rewritten by Patrick +Loiseleur: the new tactic does not any more require \texttt{ACDSimpl} +to compile and it makes use of $\beta\delta\iota$-reduction +not only to replace the rewriting steps, but also to achieve the +interleaving of computation and +reasoning (see \ref{DiscussReflection}). He also wrote a +few ML code for the \texttt{Add Ring} command, that allow to register +new rings dynamically. + +Proofs terms generated by \texttt{ring} are quite small, they are +linear in the number of $\oplus$ and $\otimes$ operations in the +normalized terms. Type-checking those terms requires some time because it +makes a large use of the conversion rule, but +memory requirements are much smaller. + +\asection{Discussion} +\label{DiscussReflection} + +Efficiency is not the only motivation to use reflection +here. \texttt{ring} also deals with constants, it rewrites for example the +expression $34 + 2*x -x + 12$ to the expected result $x + 46$. For the +tactic \texttt{ACDSimpl}, the only constants were 0 and 1. So the +expression $34 + 2*(x - 1) + 12$ is interpreted as +$V_0 \oplus V_1 \otimes (V_2 \ominus 1) \oplus V_3$, +with the variables mapping +$\{V_0 \mt 34; V_1 \mt 2; V_2 \mt x; V_3 \mt 12 \}$. Then it is +rewritten to $34 - x + 2*x + 12$, very far from the expected +result. Here rewriting is not sufficient: you have to do some kind of +reduction (some kind of \textit{computation}) to achieve the +normalization. + +The tactic \texttt{ring} is not only faster than a classical one: +using reflection, we get for free integration of computation and +reasoning that would be very complex to implement in the classic fashion. + +Is it the ultimate way to write tactics? +The answer is: yes and no. The \texttt{ring} tactic +uses intensively the conversion +rule of \CIC, that is replaces proof by computation the most as it is +possible. It can be useful in all situations where a classical tactic +generates huge proof terms. Symbolic Processing and Tautologies are +in that case. But there are also tactics like \texttt{Auto} or +\texttt{Linear}: that do many complex computations, using side-effects +and backtracking, and generate + a small proof term. Clearly, it would be a non-sense to +replace them by tactics using reflection. + +Another argument against the reflection is that \Coq, as a +programming language, has many nice features, like dependent types, +but is very far from the +speed and the expressive power of \ocaml. Wait a minute! With \Coq\ +it is possible to extract ML code from \CIC\ terms, right? So, why not +to link the extracted code with \Coq\ to inherit the benefits of the +reflection and the speed of ML tactics? That is called \textit{total + reflection}, and is still an active research subject. With these +technologies it will become possible to bootstrap the type-checker of +\CIC, but there is still some work to achieve that goal. + +Another brilliant idea from Benjamin Werner: reflection could be used +to couple a external tool (a rewriting program or a model checker) +with \Coq. We define (in \Coq) a type of terms, a type of +\emph{traces}, and prove a correction theorem that states that +\emph{replaying traces} is safe w.r.t some interpretation. Then we let +the external tool do every computation (using side-effects, +backtracking, exception, or others features that are not available in +pure lambda calculus) to produce the trace: now we replay the trace in +Coq{}, and apply the correction lemma. So internalization seems to be +the best way to import \dots{} external proofs! + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-add.tex b/doc/refman/RefMan-add.tex new file mode 100755 index 0000000000..d04d1468fe --- /dev/null +++ b/doc/refman/RefMan-add.tex @@ -0,0 +1,54 @@ +\chapter{List of additional documentation}\label{Addoc} + +\section{Tutorials}\label{Tutorial} +A companion volume to this reference manual, the \Coq\ Tutorial, is +aimed at gently introducing new users to developing proofs in \Coq\ +without assuming prior knowledge of type theory. In a second step, the +user can read also the tutorial on recursive types (document {\tt +RecTutorial.ps}). + +\section{The \Coq\ standard library}\label{Addoc-library} +A brief description of the \Coq\ standard library is given in the additional +document {\tt Library.dvi}. + +\section{Installation and un-installation procedures}\label{Addoc-install} +A \verb!INSTALL! file in the distribution explains how to install +\Coq. + +\section{{\tt Extraction} of programs}\label{Addoc-extract} +{\tt Extraction} is a package offering some special facilities to +extract ML program files. It is described in the separate document +{\tt Extraction.dvi} +\index{Extraction of programs} + +\section{Proof printing in {\tt Natural} language}\label{Addoc-natural} +{\tt Natural} is a tool to print proofs in natural language. +It is described in the separate document {\tt Natural.dvi}. +\index{Natural@{\tt Print Natural}} +\index{Printing in natural language} + +\section{The {\tt Omega} decision tactic}\label{Addoc-omega} +{\bf Omega} is a tactic to automatically solve arithmetical goals in +Presburger arithmetic (i.e. arithmetic without multiplication). +It is described in the separate document {\tt Omega.dvi}. +\index{Omega@{\tt Omega}} + +\section{Simplification on rings}\label{Addoc-polynom} +A documentation of the package {\tt polynom} (simplification on rings) +can be found in the document {\tt Polynom.dvi} +\index{Polynom@{\tt Polynom}} +\index{Simplification on rings} + +%\section{Anomalies}\label{Addoc-anomalies} +%The separate document {\tt Anomalies.*} gives a list of known +%anomalies and bugs of the system. Before communicating us an +%anomalous behavior, please check first whether it has been already +%reported in this document. + +% $Id$ + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-cas.tex b/doc/refman/RefMan-cas.tex new file mode 100755 index 0000000000..c79c14e9bb --- /dev/null +++ b/doc/refman/RefMan-cas.tex @@ -0,0 +1,692 @@ +%\documentstyle[11pt,../tools/coq-tex/coq,fullpage]{article} + +%\pagestyle{plain} + +%\begin{document} +%\nocite{Augustsson85,wadler87,HuetLevy79,MaSi94,maranget94,Laville91,saidi94,dowek93,Leroy90,puel-suarez90} + +%\input{title} +%\input{macros} +%\coverpage{The Macro \verb+Cases+}{Cristina Cornes} +%\pagestyle{plain} +\chapter{The Macro {\tt Cases}}\label{Cases}\index{Cases@{\tt Cases}} + +\marginparwidth 0pt \oddsidemargin 0pt \evensidemargin 0pt \marginparsep 0pt +\topmargin 0pt \textwidth 6.5in \textheight 8.5in + + +\verb+Cases+ is an extension to the concrete syntax of Coq that allows +to write case expressions using patterns in a syntax close to that of ML languages. +This construction is just a macro that is +expanded during parsing into a sequence of the primitive construction + \verb+Case+. +The current implementation contains two strategies, one for compiling +non-dependent case and another one for dependent case. +\section{Patterns}\label{implementation} +A pattern is a term that indicates the {\em shape} of a value, i.e. a +term where the variables can be seen as holes. When a value is +matched against a pattern (this is called {\em pattern matching}) +the pattern behaves as a filter, and associates a sub-term of the value +to each hole (i.e. to each variable pattern). + + +The syntax of patterns is presented in figure \ref{grammar}\footnote{ +Notation: \{$P$\}$^*$ denotes zero or more repetitions of $P$ and + \{$P$\}$^+$ denotes one or more repetitions of $P$. {\sl command} is the +non-terminal corresponding to terms in Coq.}. +Patterns are built up from constructors and variables. Any identifier +that is not a constructor of an inductive or coinductive type is +considered to be +a variable. Identifiers in patterns should be linear except for +the ``don't care'' pattern denoted by ``\verb+_+''. +We can use patterns to build more complex patterns. +We call {\em simple pattern} a variable or a pattern of the form +$(c~\vec{x})$ where $c$ is a constructor symbol and $\vec{x}$ is a +linear vector of variables. If a pattern is +not simple we call it {\em nested}. + + +A variable pattern matches any value, and the +identifier is bound to that value. The pattern ``\verb+_+'' also matches +any value, but it is not binding. Alias patterns written \verb+(+{\sl pattern} \verb+as+ {\sl +identifier}\verb+)+ are also accepted. This pattern matches the same values as +{\sl pattern} does and +{\sl identifier} is bound to the matched value. +A list of patterns is also considered as a pattern and is called {\em +multiple pattern}. + +\begin{figure}[t] +\begin{center} +\begin{sl} +\begin{tabular}{|l|}\hline \\ +\begin{tabular}{rcl}%\hline && \\ +simple\_pattern & := & pattern \verb+as+ identifier \\ + &$|$ & pattern \verb+,+ pattern \\ + &$|$ & pattern pattern\_list \\ && \\ + +pattern & := & identifier $|$ \verb+(+ simple\_pattern \verb+)+ \\ &&\\ + + +equation & := & \{pattern\}$^+$ ~\verb+=>+ ~term \\ && \\ + +ne\_eqn\_list & := & \verb+|+$^{opt}$ equation~ \{\verb+|+ equation\}$^*$ \\ &&\\ + +eqn\_list & := & \{~equation~ \{\verb+|+ equation\}$^*$~\}$^*$\\ &&\\ + + +term & := & +\verb+Cases+ \{term \}$^+$ \verb+of+ ne\_eqn\_list \verb+end+ \\ +&$|$ & \verb+<+term\verb+>+ \verb+Cases+ \{ term \}$^+$ +\verb+of+ eqn\_list \verb+end+ \\&& %\\ \hline +\end{tabular} \\ \hline +\end{tabular} +\end{sl} \end{center} +\caption{Macro Cases syntax.} +\label{grammar} +\end{figure} + + +Pattern matching improves readability. Compare for example the term +of the function {\em is\_zero} of natural +numbers written with patterns and the one written in primitive +concrete syntax (note that the first bar \verb+|+ is optional)~: + +\begin{center} +\begin{small} +\begin{tabular}{l} +\verb+[n:nat] Cases n of | O => true | _ => false end+,\\ +\verb+[n:nat] Case n of true [_:nat]false end+. +\end{tabular} +\end{small} +\end{center} + +In Coq pattern matching is compiled into the primitive constructions, +thus the expressiveness of the theory remains the same. Once the stage +of parsing has finished patterns disappear. An easy way to see the +result of the expansion is by printing the term with \texttt{Print} if +the term is a constant, or +using the command \texttt{Check} that displays +the term with its type : + +\begin{coq_example} +Check (fun n:nat => match n with + | O => true + | _ => false + end). +\end{coq_example} + + +\verb+Cases+ accepts optionally an infix term enclosed between +brackets \verb+<>+ that we +call the {\em elimination predicate}. +This term is the same argument as the one expected by the primitive +\verb+Case+. Given a pattern matching +expression, if all the right hand sides of \verb+=>+ ({\em rhs} in +short) have the same type, then this term +can be sometimes synthesized, and so we can omit the \verb+<>+. +Otherwise we have to +provide the predicate between \verb+<>+ as for the primitive \verb+Case+. + +Let us illustrate through examples the different aspects of pattern matching. +Consider for example the function that computes the maximum of two +natural numbers. We can write it in primitive syntax by: +\begin{coq_example} +Fixpoint max (n m:nat) {struct m} : nat := + match n with + | O => + (* O *) m + (* S n' *) + | S n' => + match m with + | O => + (* O *) S n' + (* S m' *) + | S m' => S (max n' m') + end + end. +\end{coq_example} + +Using patterns in the definitions gives: + +\begin{coq_example} +Reset max. +Fixpoint max (n m:nat) {struct m} : nat := + match n with + | O => m + | S n' => match m with + | O => S n' + | S m' => S (max n' m') + end + end. +\end{coq_example} + +Another way to write this definition is to use a multiple pattern to + match \verb+n+ and \verb+m+: + +\begin{coq_example} +Reset max. +Fixpoint max (n m:nat) {struct m} : nat := + match n, m with + | O, _ => m + | S n', O => S n' + | S n', S m' => S (max n' m') + end. +\end{coq_example} + + +The strategy examines patterns +from left to right. A case expression is generated {\bf only} when there is at least one constructor in the column of patterns. +For example, +\begin{coq_example} +Check (fun x:nat => match x return nat with + | y => y + end). +\end{coq_example} + + + +We can also use ``\verb+as+ patterns'' to associate a name to a +sub-pattern: + +\begin{coq_example} +Reset max. +Fixpoint max (n m:nat) {struct n} : nat := + match n, m with + | O, _ => m + | S n' as N, O => N + | S n', S m' => S (max n' m') + end. +\end{coq_example} + + +In the previous examples patterns do not conflict with, but +sometimes it is comfortable to write patterns that admits a non +trivial superposition. Consider +the boolean function $lef$ that given two natural numbers +yields \verb+true+ if the first one is less or equal than the second +one and \verb+false+ otherwise. We can write it as follows: + +\begin{coq_example} +Fixpoint lef (n m:nat) {struct m} : bool := + match n, m with + | O, x => true + | x, O => false + | S n, S m => lef n m + end. +\end{coq_example} + +Note that the first and the second multiple pattern superpose because the couple of +values \verb+O O+ matches both. Thus, what is the result of the +function on those values? +To eliminate ambiguity we use the {\em textual priority rule}: we +consider patterns ordered from top to bottom, then a value is matched +by the pattern at the $ith$ row if and only if is not matched by some +pattern of a previous row. Thus in the example, +\verb+O O+ is matched by the first pattern, and so \verb+(lef O O)+ +yields \verb+true+. + +Another way to write this function is: + +\begin{coq_example} +Reset lef. +Fixpoint lef (n m:nat) {struct m} : bool := + match n, m with + | O, x => true + | S n, S m => lef n m + | _, _ => false + end. +\end{coq_example} + + +Here the last pattern superposes with the first two. Because +of the priority rule, the last pattern +will be used only for values that do not match neither the first nor +the second one. + +Terms with useless patterns are accepted by the +system. For example, +\begin{coq_example} +Check + (fun x:nat => match x with + | O => true + | S _ => false + | x => true + end). +\end{coq_example} + +is accepted even though the last pattern is never used. +Beware, the +current implementation rises no warning message when there are unused +patterns in a term. + + + + +\subsection{About patterns of parametric types} +When matching objects of a parametric type, constructors in patterns +{\em do not expect} the parameter arguments. Their value is deduced +during expansion. + +Consider for example the polymorphic lists: + +\begin{coq_example} +Inductive List (A:Set) : Set := + | nil : List A + | cons : A -> List A -> List A. +\end{coq_example} + +We can check the function {\em tail}: + +\begin{coq_example} +Check + (fun l:List nat => + match l with + | nil => nil nat + | cons _ l' => l' + end). +\end{coq_example} + + +When we use parameters in patterns there is an error message: +\begin{coq_example} +Check + (fun l:List nat => + match l with + | nil nat => nil nat + | cons nat _ l' => l' + end). +\end{coq_example} + + + +\subsection{Matching objects of dependent types} +The previous examples illustrate pattern matching on objects of +non-dependent types, but we can also +use the macro to destructure objects of dependent type. +Consider the type \verb+listn+ of lists of a certain length: + +\begin{coq_example} +Inductive listn : nat -> Set := + | niln : listn 0%N + | consn : forall n:nat, nat -> listn n -> listn (S n). +\end{coq_example} + +\subsubsection{Understanding dependencies in patterns} +We can define the function \verb+length+ over \verb+listn+ by : + +\begin{coq_example} +Definition length (n:nat) (l:listn n) := n. +\end{coq_example} + +Just for illustrating pattern matching, +we can define it by case analysis: +\begin{coq_example} +Reset length. +Definition length (n:nat) (l:listn n) := + match l with + | niln => 0%N + | consn n _ _ => S n + end. +\end{coq_example} + +We can understand the meaning of this definition using the +same notions of usual pattern matching. + +Now suppose we split the second pattern of \verb+length+ into two +cases so to give an +alternative definition using nested patterns: +\begin{coq_example} +Definition length1 (n:nat) (l:listn n) := + match l with + | niln => 0%N + | consn n _ niln => S n + | consn n _ (consn _ _ _) => S n + end. +\end{coq_example} + +It is obvious that \verb+length1+ is another version of +\verb+length+. We can also give the following definition: +\begin{coq_example} +Definition length2 (n:nat) (l:listn n) := + match l with + | niln => 0%N + | consn n _ niln => 1%N + | consn n _ (consn m _ _) => S (S m) + end. +\end{coq_example} + +If we forget that \verb+listn+ is a dependent type and we read these +definitions using the usual semantics of pattern matching, we can conclude +that \verb+length1+ +and \verb+length2+ are different functions. +In fact, they are equivalent +because the pattern \verb+niln+ implies that \verb+n+ can only match +the value $0$ and analogously the pattern \verb+consn+ determines that \verb+n+ can +only match values of the form $(S~v)$ where $v$ is the value matched by +\verb+m+. + + +The converse is also true. If +we destructure the length value with the pattern \verb+O+ then the list +value should be $niln$. +Thus, the following term \verb+length3+ corresponds to the function +\verb+length+ but this time defined by case analysis on the dependencies instead of on the list: + +\begin{coq_example} +Definition length3 (n:nat) (l:listn n) := + match l with + | niln => 0%N + | consn O _ _ => 1%N + | consn (S n) _ _ => S (S n) + end. +\end{coq_example} + +When we have nested patterns of dependent types, the semantics of +pattern matching becomes a little more difficult because +the set of values that are matched by a sub-pattern may be conditioned by the +values matched by another sub-pattern. Dependent nested patterns are +somehow constrained patterns. +In the examples, the expansion of +\verb+length1+ and \verb+length2+ yields exactly the same term + but the +expansion of \verb+length3+ is completely different. \verb+length1+ and +\verb+length2+ are expanded into two nested case analysis on +\verb+listn+ while \verb+length3+ is expanded into a case analysis on +\verb+listn+ containing a case analysis on natural numbers inside. + + +In practice the user can think about the patterns as independent and +it is the expansion algorithm that cares to relate them. \\ + + +\subsubsection{When the elimination predicate must be provided} +The examples given so far do not need an explicit elimination predicate +between \verb+<>+ because all the rhs have the same type and the +strategy succeeds to synthesize it. +Unfortunately when dealing with dependent patterns it often happens +that we need to write cases where the type of the rhs are +different instances of the elimination predicate. +The function \verb+concat+ for \verb+listn+ +is an example where the branches have different type +and we need to provide the elimination predicate: + +\begin{coq_example} +Fixpoint concat (n:nat) (l:listn n) (m:nat) (l':listn m) {struct l} : + listn (n + m) := + match l in listn n return listn (n + m) with + | niln => l' + | consn n' a y => consn (n' + m) a (concat n' y m l') + end. +\end{coq_example} + +Recall that a list of patterns is also a pattern. So, when +we destructure several terms at the same time and the branches have +different type we need to provide +the elimination predicate for this multiple pattern. + +For example, an equivalent definition for \verb+concat+ (even though with a useless extra pattern) would have +been: + +\begin{coq_example} +Reset concat. +Fixpoint concat (n:nat) (l:listn n) (m:nat) (l':listn m) {struct l} : + listn (n + m) := + match l in listn n, l' return listn (n + m) with + | niln, x => x + | consn n' a y, x => consn (n' + m) a (concat n' y m x) + end. +\end{coq_example} + +Note that this time, the predicate \verb+[n,_:nat](listn (plus n m))+ is binary because we +destructure both \verb+l+ and \verb+l'+ whose types have arity one. +In general, if we destructure the terms $e_1\ldots e_n$ +the predicate will be of arity $m$ where $m$ is the sum of the +number of dependencies of the type of $e_1, e_2,\ldots e_n$ (the $\lambda$-abstractions +should correspond from left to right to each dependent argument of the +type of $e_1\ldots e_n$). +When the arity of the predicate (i.e. number of abstractions) is not +correct Coq rises an error message. For example: + +\begin{coq_example} +Fixpoint concat (n:nat) (l:listn n) (m:nat) (l':listn m) {struct l} : + listn (n + m) := + match l, l' with + | niln, x => x + | consn n' a y, x => consn (n' + m) a (concat n' y m x) + end. +\end{coq_example} + + +\subsection{Using pattern matching to write proofs} +In all the previous examples the elimination predicate does not depend on the object(s) matched. +The typical case where this is not possible is when we write a proof by +induction or a function that yields an object of dependent type. + +For example, we can write +the function \verb+buildlist+ that given a natural number +$n$ builds a list length $n$ containing zeros as follows: + +\begin{coq_example} +Fixpoint buildlist (n:nat) : listn n := + match n return listn n with + | O => niln + | S n => consn n 0 (buildlist n) + end. +\end{coq_example} + +We can also use multiple patterns whenever the elimination predicate has +the correct arity. + +Consider the following definition of the predicate less-equal +\verb+Le+: + +\begin{coq_example} +Inductive Le : nat -> nat -> Prop := + | LeO : forall n:nat, Le 0%N n + | LeS : forall n m:nat, Le n m -> Le (S n) (S m). +\end{coq_example} + +We can use multiple patterns to write the proof of the lemma + \verb+(n,m:nat) (Le n m)\/(Le m n)+: + +\begin{coq_example} +Fixpoint dec (n m:nat) {struct n} : Le n m \/ Le m n := + match n, m return Le n m \/ Le m n with + | O, x => or_introl (Le x 0) (LeO x) + | x, O => or_intror (Le x 0) (LeO x) + | S n as N, S m as M => + match dec n m with + | or_introl h => or_introl (Le M N) (LeS n m h) + | or_intror h => or_intror (Le N M) (LeS m n h) + end + end. +\end{coq_example} +In the example of \verb+dec+ the elimination predicate is binary +because we destructure two arguments of \verb+nat+ that is a +non-dependent type. Note the first \verb+Cases+ is dependent while the +second is not. + +In general, consider the terms $e_1\ldots e_n$, +where the type of $e_i$ is an instance of a family type +$[\vec{d_i}:\vec{D_i}]T_i$ ($1\leq i +\leq n$). Then to write \verb+<+${\cal P}$\verb+>Cases+ $e_1\ldots +e_n$ \verb+of+ \ldots \verb+end+, the +elimination predicate ${\cal P}$ should be of the form: +$[\vec{d_1}:\vec{D_1}][x_1:T_1]\ldots [\vec{d_n}:\vec{D_n}][x_n:T_n]Q.$ + + + + +\section{Extending the syntax of pattern} +The primitive syntax for patterns considers only those patterns containing +symbols of constructors and variables. Nevertheless, we +may define our own syntax for constructors and may be interested in +using this syntax to write patterns. +Because not any term is a pattern, the fact of extending the terms +syntax does not imply the extension of pattern syntax. Thus, +the grammar of patterns should be explicitly extended whenever we +want to use a particular syntax for a constructor. +The grammar rules for the macro \verb+Cases+ (and thus for patterns) +are defined in the file \verb+Multcase.v+ in the directory +\verb+src/syntax+. To extend the grammar of patterns +we need to extend the non-terminals corresponding to patterns +(we refer the reader to chapter of grammar extensions). + + +We have already extended the pattern syntax so as to note +the constructor \verb+pair+ of cartesian product with "( , )" in patterns. +This allows for example, to write the first projection +of pairs as follows: +\begin{coq_example} +Definition fst (A B:Set) (H:A * B) := match H with + | pair x y => x + end. +\end{coq_example} +The grammar presented in figure \ref{grammar} actually +contains this extension. + +\section{When does the expansion strategy fail?}\label{limitations} +The strategy works very like in ML languages when treating +patterns of non-dependent type. +But there are new cases of failure that are due to the presence of +dependencies. + +The error messages of the current implementation may be +sometimes confusing. +When the tactic fails because patterns are somehow incorrect then +error messages refer to the initial expression. But the strategy +may succeed to build an expression whose sub-expressions are well typed but +the whole expression is not. In this situation the message makes +reference to the expanded expression. +We encourage users, when they have patterns with the same outer constructor in different equations, to name the variable patterns in the same positions with the same name. +E.g. to write {\small\verb+(cons n O x) => e1+} and {\small\verb+(cons n \_ x) => e2+} instead of +{\small\verb+(cons n O x) => e1+} and {\small\verb+(cons n' \_ x') => e2+}. This helps to maintain certain name correspondence between the generated expression and the original. + + +Here is a summary of the error messages corresponding to each situation: +\begin{itemize} +\item patterns are incorrect (because constructors are not +applied to the correct number of the arguments, because they are not linear or they are +wrongly typed) +\begin{itemize} +\item \sverb{In pattern } {\sl term} \sverb{the constructor } {\sl ident} +\sverb{expects } {\sl num} \sverb{arguments} + +\item \sverb{The variable } {\sl ident} \sverb{is bound several times in pattern } {\sl term} + +\item \sverb{Constructor pattern: } {\sl term} \sverb{cannot match values of type } {\sl term} +\end{itemize} + +\item the pattern matching is not exhaustive +\begin{itemize} +\item \sverb{This pattern-matching is not exhaustive} +\end{itemize} +\item the elimination predicate provided to \verb+Cases+ has not the expected arity + +\begin{itemize} +\item \sverb{The elimination predicate } {\sl term} \sverb{should be +of arity } {\sl num} \sverb{(for non dependent case) or } {\sl num} \sverb{(for dependent case)} +\end{itemize} + + \item the whole expression is wrongly typed, or the synthesis of implicit arguments fails (for example to find +the elimination predicate or to resolve implicit arguments in the rhs). + + +There are {\em nested patterns of dependent type}, the +elimination predicate corresponds to non-dependent case and has the form $[x_1:T_1]...[x_n:T_n]T$ +and {\bf some} $x_i$ occurs {\bf free} in +$T$. Then, the strategy may fail to find out a correct elimination +predicate during some step of compilation. +In this situation we recommend the user to rewrite the nested +dependent patterns into several \verb+Cases+ with {\em simple patterns}. + +In all these cases we have the following error message: + + \begin{itemize} + \item + {\tt Expansion strategy failed to build a well typed case expression. + There is a branch that mismatches the expected type. + The risen type error on the result of expansion was:} + \end{itemize} + +\item because of nested patterns, it may happen that even though all +the rhs have the same type, the strategy needs +dependent elimination and so an elimination predicate must be +provided. The system +warns about this situation, trying to compile anyway with the +non-dependent strategy. The risen message is: +\begin{itemize} +\item {\tt Warning: This pattern matching may need dependent elimination to be compiled. +I will try, but if fails try again giving dependent elimination predicate.} +\end{itemize} + +\item there are {\em nested patterns of dependent type} and the strategy +builds a term that is well typed but recursive +calls in fix point are reported as illegal: +\begin{itemize} +\item {\tt Error: Recursive call applied to an illegal term ...} +\end{itemize} + +This is because the strategy generates a term that is correct +w.r.t. to the initial term but which does not pass the guard condition. +In this situation we recommend the user to transform the nested dependent +patterns into {\em several \verb+Cases+ of simple patterns}. +Let us explain this with an example. +Consider the following defintion of a function that yields the last +element of a list and \verb+O+ if it is empty: + +\begin{coq_example} +Fixpoint last (n:nat) (l:listn n) {struct l} : nat := + match l with + | consn _ a niln => a + | consn m _ x => last m x + | niln => 0%N + end. +\end{coq_example} + +It fails because of the priority between patterns, we know that this +definition is equivalent to the following more explicit one (which +fails too): + +\begin{coq_example*} +Fixpoint last (n:nat) (l:listn n) {struct l} : nat := + match l with + | consn _ a niln => a + | consn n _ (consn m b x) => last n (consn m b x) + | niln => 0%N + end. +\end{coq_example*} + +Note that the recursive call \sverb{(last n (consn m b x)) } is not +guarded. When treating with patterns of dependent types the strategy +interprets the first definition of \texttt{last} as the second +onefootnote{In languages of the ML family +the first definition would be translated into a term where the +variable \texttt{x} is shared in the expression. When +patterns are of non-dependent types, Coq compiles as in ML languages +using sharing. When patterns are of dependent types the compilation +reconstructs the term as in the second definition of \texttt{last} so to +ensure the result of expansion is well typed.}. +Thus it generates a +term where the recursive call is rejected by the +guard condition. + +You can get rid of this problem by writing the definition with \emph{simple +patterns}: + +\begin{coq_example} +Fixpoint last (n:nat) (l:listn n) {struct l} : nat := + match l return nat with + | consn m a x => match x with + | niln => a + | _ => last m x + end + | niln => 0%N + end. +\end{coq_example} + + +\end{itemize} + +%\end{document} + diff --git a/doc/refman/RefMan-cic.tex b/doc/refman/RefMan-cic.tex new file mode 100755 index 0000000000..52745b7a95 --- /dev/null +++ b/doc/refman/RefMan-cic.tex @@ -0,0 +1,1480 @@ +\chapter{Calculus of Inductive Constructions} +\label{Cic} +\index{Cic@\textsc{CIC}} +\index{pCic@p\textsc{CIC}} +\index{Calculus of (Co)Inductive Constructions} + +The underlying formal language of {\Coq} is a {\em Calculus of + Constructions} with {\em Inductive Definitions}. It is presented in +this chapter. +For {\Coq} version V7, this Calculus was known as the +{\em Calculus of (Co)Inductive Constructions}\index{Calculus of + (Co)Inductive Constructions} (\iCIC\ in short). +The underlying calculus of {\Coq} version V8.0 and up is a weaker + calculus where the sort \Set{} satisfies predicative rules. +We call this calculus the +{\em Predicative Calculus of (Co)Inductive + Constructions}\index{Predicative Calculus of + (Co)Inductive Constructions} (\pCIC\ in short). +In section~\ref{impredicativity} we give the extra-rules for \iCIC. A + compiling option of \Coq{} allows to type-check theories in this + extended system. + +In \CIC\, all objects have a {\em type}. There are types for functions (or +programs), there are atomic types (especially datatypes)... but also +types for proofs and types for the types themselves. +Especially, any object handled in the formalism must belong to a +type. For instance, the statement {\it ``for all x, P''} is not +allowed in type theory; you must say instead: {\it ``for all x +belonging to T, P''}. The expression {\it ``x belonging to T''} is +written {\it ``x:T''}. One also says: {\it ``x has type T''}. +The terms of {\CIC} are detailed in section \ref{Terms}. + +In \CIC\, there is an internal reduction mechanism. In particular, it +allows to decide if two programs are {\em intentionally} equal (one +says {\em convertible}). Convertibility is presented in section +\ref{convertibility}. + +The remaining sections are concerned with the type-checking of terms. +The beginner can skip them. + +The reader seeking a background on the Calculus of Inductive +Constructions may read several papers. Giménez~\cite{Gim98} provides +an introduction to inductive and coinductive definitions in Coq. In +their book~\cite{CoqArt}, Bertot and Castéran give a precise +description of the \CIC{} based on numerous practical examples. +Barras~\cite{Bar99}, Werner~\cite{Wer94} and +Paulin-Mohring~\cite{Moh97} are the most recent theses dealing with +Inductive Definitions. Coquand-Huet~\cite{CoHu85a,CoHu85b,CoHu86} +introduces the Calculus of Constructions. Coquand-Paulin~\cite{CoPa89} +extended this calculus to inductive definitions. The {\CIC} is a +formulation of type theory including the possibility of inductive +constructions, Barendregt~\cite{Bar91} studies the modern form of type +theory. + +\section{The terms}\label{Terms} + +In most type theories, one usually makes a syntactic distinction +between types and terms. This is not the case for \CIC\ which defines +both types and terms in the same syntactical structure. This is +because the type-theory itself forces terms and types to be defined in +a mutual recursive way and also because similar constructions can be +applied to both terms and types and consequently can share the same +syntactic structure. + +Consider for instance the $\ra$ constructor and assume \nat\ is the +type of natural numbers. Then $\ra$ is used both to denote +$\nat\ra\nat$ which is the type of functions from \nat\ to \nat, and +to denote $\nat \ra \Prop$ which is the type of unary predicates over +the natural numbers. Consider abstraction which builds functions. It +serves to build ``ordinary'' functions as $\kw{fun}~x:\nat \Ra ({\tt mult} ~x~x)$ (assuming {\tt mult} is already defined) but may build also +predicates over the natural numbers. For instance $\kw{fun}~x:\nat \Ra +(x=x)$ will +represent a predicate $P$, informally written in mathematics +$P(x)\equiv x=x$. If $P$ has type $\nat \ra \Prop$, $(P~x)$ is a +proposition, furthermore $\kw{forall}~x:\nat,(P~x)$ will represent the type of +functions which associate to each natural number $n$ an object of +type $(P~n)$ and consequently represent proofs of the formula +``$\forall x.P(x)$''. + +\subsection{Sorts}\label{Sorts} +\index{Sorts} +Types are seen as terms of the language and then should belong to +another type. The type of a type is always a constant of the language +called a {\em sort}. + +The two basic sorts in the language of \CIC\ are \Set\ and \Prop. + +The sort \Prop\ intends to be the type of logical propositions. If +$M$ is a logical proposition then it denotes a class, namely the class +of terms representing proofs of $M$. An object $m$ belonging to $M$ +witnesses the fact that $M$ is true. An object of type \Prop\ is +called a {\em proposition}. + +The sort \Set\ intends to be the type of specifications. This includes +programs and the usual sets such as booleans, naturals, lists +etc. + +These sorts themselves can be manipulated as ordinary terms. +Consequently sorts also should be given a type. Because assuming +simply that \Set\ has type \Set\ leads to an inconsistent theory, we +have infinitely many sorts in the language of \CIC. These are, in +addition to \Set\ and \Prop\, a hierarchy of universes \Type$(i)$ +for any integer $i$. We call \Sort\ the set of sorts +which is defined by: +\[\Sort \equiv \{\Prop,\Set,\Type(i)| i \in \NN\} \] +\index{Type@{\Type}} +\index{Prop@{\Prop}} +\index{Set@{\Set}} +The sorts enjoy the following properties: {\Prop:\Type(0)}, {\Set:\Type(0)} and + {\Type$(i)$:\Type$(i+1)$}. + +The user will never mention explicitly the index $i$ when referring to +the universe \Type$(i)$. One only writes \Type. The +system itself generates for each instance of \Type\ a new +index for the universe and checks that the constraints between these +indexes can be solved. From the user point of view we consequently +have {\sf Type :Type}. + +We shall make precise in the typing rules the constraints between the +indexes. + +\subsection{Constants} +Besides the sorts, the language also contains constants denoting +objects in the environment. These constants may denote previously +defined objects but also objects related to inductive definitions +(either the type itself or one of its constructors or destructors). + +\medskip\noindent {\bf Remark. } In other presentations of \CIC, +the inductive objects are not seen as +external declarations but as first-class terms. Usually the +definitions are also completely ignored. This is a nice theoretical +point of view but not so practical. An inductive definition is +specified by a possibly huge set of declarations, clearly we want to +share this specification among the various inductive objects and not +to duplicate it. So the specification should exist somewhere and the +various objects should refer to it. We choose one more level of +indirection where the objects are just represented as constants and +the environment gives the information on the kind of object the +constant refers to. + +\medskip +Our inductive objects will be manipulated as constants declared in the +environment. This roughly corresponds to the way they are actually +implemented in the \Coq\ system. It is simple to map this presentation +in a theory where inductive objects are represented by terms. + +\subsection{Terms} + +Terms are built from variables, global names, constructors, +abstraction, application, local declarations bindings (``let-in'' +expressions) and product. + +From a syntactic point of view, types cannot be distinguished from terms, +except that they cannot start by an abstraction, and that if a term is +a sort or a product, it should be a type. + +More precisely the language of the {\em Calculus of Inductive + Constructions} is built from the following rules: + +\begin{enumerate} +\item the sorts {\sf Set, Prop, Type} are terms. +\item names for global constants of the environment are terms. +\item variables are terms. +\item if $x$ is a variable and $T$, $U$ are terms then $\forall~x:T,U$ + ($\kw{forall}~x:T,U$ in \Coq{} concrete syntax) is a term. If $x$ + occurs in $U$, $\forall~x:T,U$ reads as {\it ``for all x of type T, + U''}. As $U$ depends on $x$, one says that $\forall~x:T,U$ is a + {\em dependent product}. If $x$ doesn't occurs in $U$ then + $\forall~x:T,U$ reads as {\it ``if T then U''}. A non dependent + product can be written: $T \rightarrow U$. +\item if $x$ is a variable and $T$, $U$ are terms then $\lb~x:T \mto U$ + ($\kw{fun}~x:T\Ra U$ in \Coq{} concrete syntax) is a term. This is a + notation for the $\lambda$-abstraction of + $\lambda$-calculus\index{lambda-calculus@$\lambda$-calculus} + \cite{Bar81}. The term $\lb~x:T \mto U$ is a function which maps + elements of $T$ to $U$. +\item if $T$ and $U$ are terms then $(T\ U)$ is a term + ($T~U$ in \Coq{} concrete syntax). The term $(T\ + U)$ reads as {\it ``T applied to U''}. +\item if $x$ is a variable, and $T$, $U$ are terms then + $\kw{let}~x:=T~\kw{in}~U$ is a + term which denotes the term $U$ where the variable $x$ is locally + bound to $T$. This stands for the common ``let-in'' construction of + functional programs such as ML or Scheme. +\end{enumerate} + +\paragraph{Notations.} Application associates to the left such that +$(t~t_1\ldots t_n)$ represents $(\ldots (t~t_1)\ldots t_n)$. The +products and arrows associate to the right such that $\forall~x:A,B\ra C\ra +D$ represents $\forall~x:A,(B\ra (C\ra D))$. One uses sometimes +$\forall~x~y:A,B$ or +$\lb~x~y:A\mto B$ to denote the abstraction or product of several variables +of the same type. The equivalent formulation is $\forall~x:A, \forall y:A,B$ or +$\lb~x:A \mto \lb y:A \mto B$ + +\paragraph{Free variables.} +The notion of free variables is defined as usual. In the expressions +$\lb~x:T\mto U$ and $\forall x:T, U$ the occurrences of $x$ in $U$ +are bound. They are represented by de Bruijn indexes in the internal +structure of terms. + +\paragraph{Substitution.} \index{Substitution} +The notion of substituting a term $t$ to free occurrences of a +variable $x$ in a term $u$ is defined as usual. The resulting term +is written $\subst{u}{x}{t}$. + + +\section{Typed terms}\label{Typed-terms} + +As objects of type theory, terms are subjected to {\em type +discipline}. The well typing of a term depends on an environment which +consists in a global environment (see below) and a local context. + +\paragraph{Local context.} +A {\em local context} (or shortly context) is an ordered list of +declarations of variables. The declaration of some variable $x$ is +either an assumption, written $x:T$ ($T$ is a type) or a definition, +written $x:=t:T$. We use brackets to write contexts. A +typical example is $[x:T;y:=u:U;z:V]$. Notice that the variables +declared in a context must be distinct. If $\Gamma$ declares some $x$, +we write $x \in\Gamma$. By writing $(x:T)\in\Gamma$ we mean that +either $x:T$ is an assumption in $\Gamma$ or that there exists some $t$ such +that $x:=t:T$ is a definition in $\Gamma$. If $\Gamma$ defines some +$x:=t:T$, we also write $(x:=t:T)\in\Gamma$. Contexts must be +themselves {\em well formed}. For the rest of the chapter, the +notation $\Gamma::(y:T)$ (resp $\Gamma::(y:=t:T)$) denotes the context +$\Gamma$ enriched with the declaration $y:T$ (resp $y:=t:T$). The +notation $[]$ denotes the empty context. \index{Context} + +% Does not seem to be used further... +% Si dans l'explication WF(E)[Gamma] concernant les constantes +% definies ds un contexte + +We define the inclusion of two contexts $\Gamma$ and $\Delta$ (written +as $\Gamma \subset \Delta$) as the property, for all variable $x$, +type $T$ and term $t$, if $(x:T) \in \Gamma$ then $(x:T)\in \Delta$ +and if $(x:=t:T) \in \Gamma$ then $(x:=t:T)\in \Delta$. +%We write +% $|\Delta|$ for the length of the context $\Delta$, that is for the number +% of declarations (assumptions or definitions) in $\Delta$. + +A variable $x$ is said to be free in $\Gamma$ if $\Gamma$ contains a +declaration $y:T$ such that $x$ is free in $T$. + +\paragraph{Environment.}\index{Environment} +Because we are manipulating global declarations (constants and global +assumptions), we also need to consider a global environment $E$. + +An environment is an ordered list of declarations of global +names. Declarations are either assumptions or ``standard'' +definitions, that is abbreviations for well-formed terms +but also definitions of inductive objects. In the latter +case, an object in the environment will define one or more constants +(that is types and constructors, see section \ref{Cic-inductive-definitions}). + +An assumption will be represented in the environment as +\Assum{\Gamma}{c}{T} which means that $c$ is assumed of some type $T$ +well-defined in some context $\Gamma$. An (ordinary) definition will +be represented in the environment as \Def{\Gamma}{c}{t}{T} which means +that $c$ is a constant which is valid in some context $\Gamma$ whose +value is $t$ and type is $T$. + +The rules for inductive definitions (see section +\ref{Cic-inductive-definitions}) have to be considered as assumption +rules to which the following definitions apply: if the name $c$ is +declared in $E$, we write $c \in E$ and if $c:T$ or $c:=t:T$ is +declared in $E$, we write $(c : T) \in E$. + +\paragraph{Typing rules.}\label{Typing-rules}\index{Typing rules} +In the following, we assume $E$ is a valid environment wrt to +inductive definitions. We define simultaneously two +judgments. The first one \WTEG{t}{T} means the term $t$ is well-typed +and has type $T$ in the environment $E$ and context $\Gamma$. The +second judgment \WFE{\Gamma} means that the environment $E$ is +well-formed and the context $\Gamma$ is a valid context in this +environment. It also means a third property which makes sure that any +constant in $E$ was defined in an environment which is included in +$\Gamma$ +\footnote{This requirement could be relaxed if we instead introduced + an explicit mechanism for instantiating constants. At the external + level, the Coq engine works accordingly to this view that all the + definitions in the environment were built in a sub-context of the + current context.}. + +A term $t$ is well typed in an environment $E$ iff there exists a +context $\Gamma$ and a term $T$ such that the judgment \WTEG{t}{T} can +be derived from the following rules. +\begin{description} +\item[W-E] \inference{\WF{[]}{[]}} +\item[W-S] % Ce n'est pas vrai : x peut apparaitre plusieurs fois dans Gamma +\inference{\frac{\WTEG{T}{s}~~~~s\in \Sort~~~~x \not\in + \Gamma % \cup E + } + {\WFE{\Gamma::(x:T)}}~~~~~ + \frac{\WTEG{t}{T}~~~~x \not\in + \Gamma % \cup E + }{\WFE{\Gamma::(x:=t:T)}}} +\item[Def] \inference{\frac{\WTEG{t}{T}~~~c \notin E\cup \Gamma} + {\WF{E;\Def{\Gamma}{c}{t}{T}}{\Gamma}}} +\item[Ax] \index{Typing rules!Ax} +\inference{\frac{\WFE{\Gamma}}{\WTEG{\Prop}{\Type(p)}}~~~~~ +\frac{\WFE{\Gamma}}{\WTEG{\Set}{\Type(q)}}} +\inference{\frac{\WFE{\Gamma}~~~~i nat. +Inductive list (A:Set) : Set := + | nil : list A + | cons : A -> list A -> list A. +\end{coq_example*} +\begin{coq_example*} +Inductive Length (A:Set) : list A -> nat -> Prop := + | Lnil : Length A (nil A) O + | Lcons : + forall (a:A) (l:list A) (n:nat), + Length A l n -> Length A (cons A a l) (S n). +Inductive tree : Set := + node : forest -> tree +with forest : Set := + | emptyf : forest + | consf : tree -> forest -> forest. +\end{coq_example*} +The inductive declaration in \Coq\ is slightly different from the one +we described theoretically. The difference is that in the type of +constructors the inductive definition is explicitly applied to the +parameters variables. The \Coq\ type-checker verifies that all +parameters are applied in the correct manner in each recursive call. +In particular, the following definition will not be accepted because +there is an occurrence of \List\ which is not applied to the parameter +variable: +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is not correct and should produce **********) +(********* Error: The 1st argument of list' must be A in ... *********) +\end{coq_eval} +\begin{coq_example} +Inductive list' (A:Set) : Set := + | nil' : list' A + | cons' : A -> list' (A -> A) -> list' A. +\end{coq_example} + +\subsection{Types of inductive objects} +We have to give the type of constants in an environment $E$ which +contains an inductive declaration. + +\begin{description} +\item[Ind-Const] Assuming $\Gamma_P$ is $[p_1:P_1;\ldots;p_r:P_r]$, + $\Gamma_I$ is $[I_1:A_1;\ldots;I_k:A_k]$, and $\Gamma_C$ is + $[c_1:C_1;\ldots;c_n:C_n]$, + +\inference{\frac{\Ind{\Gamma}{\Gamma_P}{\Gamma_I}{\Gamma_C} \in E + ~~j=1\ldots k}{(I_j:\forall~p_1:P_1,\ldots\forall p_r:P_r,A_j) \in E}} + +\inference{\frac{\Ind{\Gamma}{\Gamma_P}{\Gamma_I}{\Gamma_C} \in E + ~~~~i=1.. n} + {(c_i:\forall~p_1:P_1,\ldots \forall p_r:P_r,\subst{C_i}{I_j}{(I_j~p_1\ldots + p_r)}_{j=1\ldots k})\in E}} +\end{description} + +\paragraph{Example.} +We have $(\List:\Set \ra \Set), (\cons:\forall~A:\Set,A\ra(\List~A)\ra +(\List~A))$, \\ +$(\Length:\forall~A:\Set, (\List~A)\ra\nat\ra\Prop)$, $\tree:\Set$ and $\forest:\Set$. + +From now on, we write $\ListA$ instead of $(\List~A)$ and $\LengthA$ +for $(\Length~A)$. + +%\paragraph{Parameters.} +%%The parameters introduce a distortion between the inside specification +%%of the inductive declaration where parameters are supposed to be +%%instantiated (this representation is appropriate for checking the +%%correctness or deriving the destructor principle) and the outside +%%typing rules where the inductive objects are seen as objects +%%abstracted with respect to the parameters. + +%In the definition of \List\ or \Length\, $A$ is a parameter because +%what is effectively inductively defined is $\ListA$ or $\LengthA$ for +%a given $A$ which is constant in the type of constructors. But when +%we define $(\LengthA~l~n)$, $l$ and $n$ are not parameters because the +%constructors manipulate different instances of this family. + +\subsection{Well-formed inductive definitions} +We cannot accept any inductive declaration because some of them lead +to inconsistent systems. We restrict ourselves to definitions which +satisfy a syntactic criterion of positivity. Before giving the formal +rules, we need a few definitions: + +\paragraph{Definitions}\index{Positivity}\label{Positivity} + +A type $T$ is an {\em arity of sort $s$}\index{Arity} if it converts +to the sort $s$ or to a product $\forall~x:T,U$ with $U$ an arity +of sort $s$. (For instance $A\ra \Set$ or $\forall~A:\Prop,A\ra +\Prop$ are arities of sort respectively \Set\ and \Prop). A {\em type + of constructor of $I$}\index{Type of constructor} is either a term +$(I~t_1\ldots ~t_n)$ or $\fa x:T,C$ with $C$ a {\em type of constructor + of $I$}. + +\smallskip + +The type of constructor $T$ will be said to {\em satisfy the positivity +condition} for a constant $X$ in the following cases: + +\begin{itemize} +\item $T=(X~t_1\ldots ~t_n)$ and $X$ does not occur free in +any $t_i$ +\item $T=\forall~x:U,V$ and $X$ occurs only strictly positively in $U$ and +the type $V$ satisfies the positivity condition for $X$ +\end{itemize} + +The constant $X$ {\em occurs strictly positively} in $T$ in the +following cases: + +\begin{itemize} +\item $X$ does not occur in $T$ +\item $T$ converts to $(X~t_1 \ldots ~t_n)$ and $X$ does not occur in + any of $t_i$ +\item $T$ converts to $\forall~x:U,V$ and $X$ does not occur in + type $U$ but occurs strictly positively in type $V$ +\item $T$ converts to $(I~a_1 \ldots ~a_m ~ t_1 \ldots ~t_p)$ where + $I$ is the name of an inductive declaration of the form + $\Ind{\Gamma}{p_1:P_1;\ldots;p_m:P_m}{I:A}{c_1:C_1;\ldots;c_n:C_n}$ + (in particular, it is not mutually defined and it has $m$ + parameters) and $X$ does not occur in any of the $t_i$, and the + types of constructor $C_i\{p_j/a_j\}_{j=1\ldots m}$ of $I$ satisfy + the nested positivity condition for $X$ +%\item more generally, when $T$ is not a type, $X$ occurs strictly +%positively in $T[x:U]u$ if $X$ does not occur in $U$ but occurs +%strictly positively in $u$ +\end{itemize} + +The type of constructor $T$ of $I$ {\em satisfies the nested +positivity condition} for a constant $X$ in the following +cases: + +\begin{itemize} +\item $T=(I~t_1\ldots ~t_n)$ and $X$ does not occur in +any $t_i$ +\item $T=\forall~x:U,V$ and $X$ occurs only strictly positively in $U$ and +the type $V$ satisfies the nested positivity condition for $X$ +\end{itemize} + +\paragraph{Example} + +$X$ occurs strictly positively in $A\ra X$ or $X*A$ or $({\tt list} +X)$ but not in $X \ra A$ or $(X \ra A)\ra A$ nor $({\tt neg}~A)$ +assuming the notion of product and lists were already defined and {\tt + neg} is an inductive definition with declaration \Ind{}{A:\Set}{{\tt + neg}:\Set}{{\tt neg}:(A\ra{\tt False}) \ra {\tt neg}}. Assuming +$X$ has arity ${\tt nat \ra Prop}$ and {\tt ex} is the inductively +defined existential quantifier, the occurrence of $X$ in ${\tt (ex~ + nat~ \lb~n:nat\mto (X~ n))}$ is also strictly positive. + +\paragraph{Correctness rules.} +We shall now describe the rules allowing the introduction of a new +inductive definition. + +\begin{description} +\item[W-Ind] Let $E$ be an environment and + $\Gamma,\Gamma_P,\Gamma_I,\Gamma_C$ are contexts such that + $\Gamma_I$ is $[I_1:A_1;\ldots;I_k:A_k]$ and $\Gamma_C$ is + $[c_1:C_1;\ldots;c_n:C_n]$. +\inference{ + \frac{ + (\WTE{\Gamma;\Gamma_P}{A_j}{s'_j})_{j=1\ldots k} + ~~ (\WTE{\Gamma;\Gamma_P;\Gamma_I}{C_i}{s_{p_i}})_{i=1\ldots n} +} + {\WF{E;\Ind{\Gamma}{\Gamma_P}{\Gamma_I}{\Gamma_C}}{\Gamma}}} +providing the following side conditions hold: +\begin{itemize} +\item $k>0$, $I_j$, $c_i$ are different names for $j=1\ldots k$ and $i=1\ldots n$, +\item for $j=1\ldots k$ we have $A_j$ is an arity of sort $s_j$ and $I_j + \notin \Gamma \cup E$, +\item for $i=1\ldots n$ we have $C_i$ is a type of constructor of + $I_{p_i}$ which satisfies the positivity condition for $I_1 \ldots I_k$ + and $c_i \notin \Gamma \cup E$. +\end{itemize} +\end{description} +One can remark that there is a constraint between the sort of the +arity of the inductive type and the sort of the type of its +constructors which will always be satisfied for the impredicative sort +(\Prop) but may fail to define inductive definition +on sort \Set{} and generate constraints between universes for +inductive definitions in types. + +\paragraph{Examples} +It is well known that existential quantifier can be encoded as an +inductive definition. +The following declaration introduces the second-order existential +quantifier $\exists X.P(X)$. +\begin{coq_example*} +Inductive exProp (P:Prop->Prop) : Prop + := exP_intro : forall X:Prop, P X -> exProp P. +\end{coq_example*} +The same definition on \Set{} is not allowed and fails~: +\begin{coq_eval} +(********** The following is not correct and should produce **********) +(*** Error: Large non-propositional inductive types must be in Type***) +\end{coq_eval} +\begin{coq_example} +Inductive exSet (P:Set->Prop) : Set + := exS_intro : forall X:Set, P X -> exSet P. +\end{coq_example} +It is possible to declare the same inductive definition in the +universe \Type. +The \texttt{exType} inductive definition has type $(\Type_i \ra\Prop)\ra +\Type_j$ with the constraint that the parameter \texttt{X} of \texttt{exT\_intro} has type $\Type_k$ with $kProp) : Type + := exT_intro : forall X:Type, P X -> exType P. +\end{coq_example*} +%We shall assume for the following definitions that, if necessary, we +%annotated the type of constructors such that we know if the argument +%is recursive or not. We shall write the type $(x:_R T)C$ if it is +%a recursive argument and $(x:_P T)C$ if the argument is not recursive. + +\subsection{Destructors} +The specification of inductive definitions with arities and +constructors is quite natural. But we still have to say how to use an +object in an inductive type. + +This problem is rather delicate. There are actually several different +ways to do that. Some of them are logically equivalent but not always +equivalent from the computational point of view or from the user point +of view. + +From the computational point of view, we want to be able to define a +function whose domain is an inductively defined type by using a +combination of case analysis over the possible constructors of the +object and recursion. + +Because we need to keep a consistent theory and also we prefer to keep +a strongly normalizing reduction, we cannot accept any sort of +recursion (even terminating). So the basic idea is to restrict +ourselves to primitive recursive functions and functionals. + +For instance, assuming a parameter $A:\Set$ exists in the context, we +want to build a function \length\ of type $\ListA\ra \nat$ which +computes the length of the list, so such that $(\length~\Nil) = \nO$ +and $(\length~(\cons~A~a~l)) = (\nS~(\length~l))$. We want these +equalities to be recognized implicitly and taken into account in the +conversion rule. + +From the logical point of view, we have built a type family by giving +a set of constructors. We want to capture the fact that we do not +have any other way to build an object in this type. So when trying to +prove a property $(P~m)$ for $m$ in an inductive definition it is +enough to enumerate all the cases where $m$ starts with a different +constructor. + +In case the inductive definition is effectively a recursive one, we +want to capture the extra property that we have built the smallest +fixed point of this recursive equation. This says that we are only +manipulating finite objects. This analysis provides induction +principles. + +For instance, in order to prove $\forall l:\ListA,(\LengthA~l~(\length~l))$ +it is enough to prove: + +\noindent $(\LengthA~\Nil~(\length~\Nil))$ and + +\smallskip +$\forall a:A, \forall l:\ListA, (\LengthA~l~(\length~l)) \ra +(\LengthA~(\cons~A~a~l)~(\length~(\cons~A~a~l)))$. +\smallskip + +\noindent which given the conversion equalities satisfied by \length\ is the +same as proving: +$(\LengthA~\Nil~\nO)$ and $\forall a:A, \forall l:\ListA, +(\LengthA~l~(\length~l)) \ra +(\LengthA~(\cons~A~a~l)~(\nS~(\length~l)))$. + +One conceptually simple way to do that, following the basic scheme +proposed by Martin-L\"of in his Intuitionistic Type Theory, is to +introduce for each inductive definition an elimination operator. At +the logical level it is a proof of the usual induction principle and +at the computational level it implements a generic operator for doing +primitive recursion over the structure. + +But this operator is rather tedious to implement and use. We choose in +this version of Coq to factorize the operator for primitive recursion +into two more primitive operations as was first suggested by Th. Coquand +in~\cite{Coq92}. One is the definition by pattern-matching. The second one is a definition by guarded fixpoints. + +\subsubsection{The {\tt match\ldots with \ldots end} construction.} +\label{Caseexpr} +\index{match@{\tt match\ldots with\ldots end}} + +The basic idea of this destructor operation is that we have an object +$m$ in an inductive type $I$ and we want to prove a property $(P~m)$ +which in general depends on $m$. For this, it is enough to prove the +property for $m = (c_i~u_1\ldots u_{p_i})$ for each constructor of $I$. + +The \Coq{} term for this proof will be written~: +\[\kw{match}~m~\kw{with}~ (c_1~x_{11}~...~x_{1p_1}) \Ra f_1 ~|~\ldots~|~ + (c_n~x_{n1}...x_{np_n}) \Ra f_n~ \kw{end}\] +In this expression, if +$m$ is a term built from a constructor $(c_i~u_1\ldots u_{p_i})$ then +the expression will behave as it is specified with $i$-th branch and +will reduce to $f_i$ where the $x_{i1}$\ldots $x_{ip_i}$ are replaced +by the $u_1\ldots u_p$ according to the $\iota$-reduction. + +Actually, for type-checking a \kw{match\ldots with\ldots end} +expression we also need to know the predicate $P$ to be proved by case +analysis. \Coq{} can sometimes infer this predicate but sometimes +not. The concrete syntax for describing this predicate uses the +\kw{as\ldots return} construction. +The predicate is made explicit using the syntax~: +\[\kw{match}~m~\kw{as}~ x~ \kw{return}~ (P~ x) ~\kw{with}~ (c_1~x_{11}~...~x_{1p_1}) \Ra f_1 ~|~\ldots~|~ + (c_n~x_{n1}...x_{np_n}) \Ra f_n \kw{end}\] +For the purpose of presenting the inference rules, we use a more +compact notation~: +\[ \Case{(\lb x \mto P)}{m}{ \lb x_{11}~...~x_{1p_1} \mto f_1 ~|~\ldots~|~ + \lb x_{n1}...x_{np_n} \mto f_n}\] + +This is the basic idea which is generalized to the case where $I$ is +an inductively defined $n$-ary relation (in which case the property +$P$ to be proved will be a $n+1$-ary relation). + + +\paragraph{Non-dependent elimination.} +When defining a function by case analysis, we build an object of type $I +\ra C$ and the minimality principle on an inductively defined logical +predicate of type $A \ra \Prop$ is often used to prove a property +$\forall x:A,(I~x)\ra (C~x)$. This is a particular case of the dependent +principle that we stated before with a predicate which does not depend +explicitly on the object in the inductive definition. + +For instance, a function testing whether a list is empty +can be +defined as: + +\[\lb~l:\ListA \mto\Case{\bool}{l}{\Nil~ \Ra~\true~ |~ (\cons~a~m)~ \Ra~\false}\] +%\noindent {\bf Remark. } + +% In the system \Coq\ the expression above, can be +% written without mentioning +% the dummy abstraction: +% \Case{\bool}{l}{\Nil~ \mbox{\tt =>}~\true~ |~ (\cons~a~m)~ +% \mbox{\tt =>}~ \false} + +\paragraph{Allowed elimination sorts.} +\index{Elimination sorts} + +An important question for building the typing rule for \kw{match} is +what can be the type of $P$ with respect to the type of the inductive +definitions. + +We define now a relation \compat{I:A}{B} between an inductive +definition $I$ of type $A$, an arity $B$ which says that an object in +the inductive definition $I$ can be eliminated for proving a property +$P$ of type $B$. + +The case of inductive definitions in sorts \Set\ or \Type{} is simple. +There is no restriction on the sort of the predicate to be +eliminated. + +\paragraph{Notations.} +The \compat{I:A}{B} is defined as the smallest relation satisfying the +following rules: +We write \compat{I}{B} for \compat{I:A}{B} where $A$ is the type of +$I$. + +\begin{description} +\item[Prod] \inference{\frac{\compat{(I~x):A'}{B'}} + {\compat{I:(x:A)A'}{(x:A)B'}}} +\item[\Set \& \Type] \inference{\frac{ + s_1 \in \{\Set,\Type(j)\}, + s_2 \in \Sort}{\compat{I:s_1}{I\ra s_2}}} +\end{description} + +The case of Inductive Definitions of sort \Prop{} is a bit more +complicated, because of our interpretation of this sort. The only +harmless allowed elimination, is the one when predicate $P$ is also of +sort \Prop. +\begin{description} +\item[\Prop] \inference{\compat{I:\Prop}{I\ra\Prop}} +\end{description} +\Prop{} is the type of logical propositions, the proofs of properties +$P$ in \Prop{} could not be used for computation and are consequently +ignored by the extraction mechanism. +Assume $A$ and $B$ are two propositions, and the logical disjunction +$A\vee B$ is defined inductively by~: +\begin{coq_example*} +Inductive or (A B:Prop) : Prop := + lintro : A -> or A B | rintro : B -> or A B. +\end{coq_example*} +The following definition which computes a boolean value by case over +the proof of \texttt{or A B} is not accepted~: +\begin{coq_eval} +(***************************************************************) +(*** This example should fail with ``Incorrect elimination'' ***) +\end{coq_eval} +\begin{coq_example} +Definition choice (A B: Prop) (x:or A B) := + match x with lintro a => true | rintro b => false end. +\end{coq_example} +From the computational point of view, the structure of the proof of +\texttt{(or A B)} in this term is needed for computing the boolean +value. + +In general, if $I$ has type \Prop\ then $P$ cannot have type $I\ra +\Set$, because it will mean to build an informative proof of type +$(P~m)$ doing a case analysis over a non-computational object that +will disappear in the extracted program. But the other way is safe +with respect to our interpretation we can have $I$ a computational +object and $P$ a non-computational one, it just corresponds to proving +a logical property of a computational object. + +% Also if $I$ is in one of the sorts \{\Prop, \Set\}, one cannot in +% general allow an elimination over a bigger sort such as \Type. But +% this operation is safe whenever $I$ is a {\em small inductive} type, +% which means that all the types of constructors of +% $I$ are small with the following definition:\\ +% $(I~t_1\ldots t_s)$ is a {\em small type of constructor} and +% $\forall~x:T,C$ is a small type of constructor if $C$ is and if $T$ +% has type \Prop\ or \Set. \index{Small inductive type} + +% We call this particular elimination which gives the possibility to +% compute a type by induction on the structure of a term, a {\em strong +% elimination}\index{Strong elimination}. + +In the same spirit, elimination on $P$ of type $I\ra +\Type$ cannot be allowed because it trivially implies the elimination +on $P$ of type $I\ra \Set$ by cumulativity. It also implies that there +is two proofs of the same property which are provably different, +contradicting the proof-irrelevance property which is sometimes a +useful axiom~: +\begin{coq_example} +Axiom proof_irrelevance : forall (P : Prop) (x y : P), x=y. +\end{coq_example} +\begin{coq_eval} +Reset proof_irrelevance. +\end{coq_eval} +The elimination of an inductive definition of type \Prop\ on a +predicate $P$ of type $I\ra \Type$ leads to a paradox when applied to +impredicative inductive definition like the second-order existential +quantifier \texttt{exProp} defined above, because it give access to +the two projections on this type. + +%\paragraph{Warning: strong elimination} +%\index{Elimination!Strong elimination} +%In previous versions of Coq, for a small inductive definition, only the +%non-informative strong elimination on \Type\ was allowed, because +%strong elimination on \Typeset\ was not compatible with the current +%extraction procedure. In this version, strong elimination on \Typeset\ +%is accepted but a dummy element is extracted from it and may generate +%problems if extracted terms are explicitly used such as in the +%{\tt Program} tactic or when extracting ML programs. + +\paragraph{Empty and singleton elimination} +\index{Elimination!Singleton elimination} +\index{Elimination!Empty elimination} + +There are special inductive definitions in \Prop\ for which more +eliminations are allowed. +\begin{description} +\item[\Prop-extended] +\inference{ + \frac{I \mbox{~is an empty or singleton + definition}~~~s\in\Sort}{\compat{I:\Prop}{I\ra s}} +} +\end{description} + +% A {\em singleton definition} has always an informative content, +% even if it is a proposition. + +A {\em singleton +definition} has only one constructor and all the arguments of this +constructor have type \Prop. In that case, there is a canonical +way to interpret the informative extraction on an object in that type, +such that the elimination on any sort $s$ is legal. Typical examples are +the conjunction of non-informative propositions and the equality. +If there is an hypothesis $h:a=b$ in the context, it can be used for +rewriting not only in logical propositions but also in any type. +% In that case, the term \verb!eq_rec! which was defined as an axiom, is +% now a term of the calculus. +\begin{coq_example} +Print eq_rec. +Extraction eq_rec. +\end{coq_example} +An empty definition has no constructors, in that case also, +elimination on any sort is allowed. + +\paragraph{Type of branches.} +Let $c$ be a term of type $C$, we assume $C$ is a type of constructor +for an inductive definition $I$. Let $P$ be a term that represents the +property to be proved. +We assume $r$ is the number of parameters. + +We define a new type \CI{c:C}{P} which represents the type of the +branch corresponding to the $c:C$ constructor. +\[ +\begin{array}{ll} +\CI{c:(I_i~p_1\ldots p_r\ t_1 \ldots t_p)}{P} &\equiv (P~t_1\ldots ~t_p~c) \\[2mm] +\CI{c:\forall~x:T,C}{P} &\equiv \forall~x:T,\CI{(c~x):C}{P} +\end{array} +\] +We write \CI{c}{P} for \CI{c:C}{P} with $C$ the type of $c$. + +\paragraph{Examples.} +For $\ListA$ the type of $P$ will be $\ListA\ra s$ for $s \in \Sort$. \\ +$ \CI{(\cons~A)}{P} \equiv +\forall a:A, \forall l:\ListA,(P~(\cons~A~a~l))$. + +For $\LengthA$, the type of $P$ will be +$\forall l:\ListA,\forall n:\nat, (\LengthA~l~n)\ra \Prop$ and the expression +\CI{(\LCons~A)}{P} is defined as:\\ +$\forall a:A, \forall l:\ListA, \forall n:\nat, \forall +h:(\LengthA~l~n), (P~(\cons~A~a~l)~(\nS~n)~(\LCons~A~a~l~n~l))$.\\ +If $P$ does not depend on its third argument, we find the more natural +expression:\\ +$\forall a:A, \forall l:\ListA, \forall n:\nat, +(\LengthA~l~n)\ra(P~(\cons~A~a~l)~(\nS~n))$. + +\paragraph{Typing rule.} + +Our very general destructor for inductive definition enjoys the +following typing rule +% , where we write +% \[ +% \Case{P}{c}{[x_{11}:T_{11}]\ldots[x_{1p_1}:T_{1p_1}]g_1\ldots +% [x_{n1}:T_{n1}]\ldots[x_{np_n}:T_{np_n}]g_n} +% \] +% for +% \[ +% \Case{P}{c}{(c_1~x_{11}~...~x_{1p_1}) \Ra g_1 ~|~\ldots~|~ +% (c_n~x_{n1}...x_{np_n}) \Ra g_n } +% \] + +\begin{description} +\item[match] \label{elimdep} \index{Typing rules!match} +\inference{ +\frac{\WTEG{c}{(I~q_1\ldots q_r~t_1\ldots t_s)}~~ + \WTEG{P}{B}~~\compat{(I~q_1\ldots q_r)}{B} + ~~ +(\WTEG{f_i}{\CI{(c_{p_i}~q_1\ldots q_r)}{P}})_{i=1\ldots l}} +{\WTEG{\Case{P}{c}{f_1\ldots f_l}}{(P\ t_1\ldots t_s\ c)}}}%\\[3mm] + +provided $I$ is an inductive type in a declaration +\Ind{\Delta}{\Gamma_P}{\Gamma_I}{\Gamma_C} with $|\Gamma_P| = r$, +$\Gamma_C = [c_1:C_1;\ldots;c_n:C_n]$ and $c_{p_1}\ldots c_{p_l}$ are the +only constructors of $I$. +\end{description} + +\paragraph{Example.} +For \List\ and \Length\ the typing rules for the {\tt match} expression +are (writing just $t:M$ instead of \WTEG{t}{M}, the environment and +context being the same in all the judgments). + +\[\frac{l:\ListA~~P:\ListA\ra s~~~f_1:(P~(\Nil~A))~~ + f_2:\forall a:A, \forall l:\ListA, (P~(\cons~A~a~l))} + {\Case{P}{l}{f_1~f_2}:(P~l)}\] + +\[\frac{ + \begin{array}[b]{@{}c@{}} +H:(\LengthA~L~N) \\ P:\forall l:\ListA, \forall n:\nat, (\LengthA~l~n)\ra + \Prop\\ + f_1:(P~(\Nil~A)~\nO~\LNil) \\ + f_2:\forall a:A, \forall l:\ListA, \forall n:\nat, \forall + h:(\LengthA~l~n), (P~(\cons~A~a~n)~(\nS~n)~(\LCons~A~a~l~n~h)) + \end{array}} + {\Case{P}{H}{f_1~f_2}:(P~L~N~H)}\] + +\paragraph{Definition of $\iota$-reduction.}\label{iotared} +\index{iota-reduction@$\iota$-reduction} +We still have to define the $\iota$-reduction in the general case. + +A $\iota$-redex is a term of the following form: +\[\Case{P}{(c_{p_i}~q_1\ldots q_r~a_1\ldots a_m)}{f_1\ldots + f_l}\] +with $c_{p_i}$ the $i$-th constructor of the inductive type $I$ with $r$ +parameters. + +The $\iota$-contraction of this term is $(f_i~a_1\ldots a_m)$ leading +to the general reduction rule: +\[ \Case{P}{(c_{p_i}~q_1\ldots q_r~a_1\ldots a_m)}{f_1\ldots + f_n} \triangleright_{\iota} (f_i~a_1\ldots a_m) \] + +\subsection{Fixpoint definitions} +\label{Fix-term} \index{Fix@{\tt Fix}} +The second operator for elimination is fixpoint definition. +This fixpoint may involve several mutually recursive definitions. +The basic concrete syntax for a recursive set of mutually recursive +declarations is (with $\Gamma_i$ contexts)~: +\[\kw{fix}~f_1 (\Gamma_1) :A_1:=t_1~\kw{with} \ldots \kw{with}~ f_n +(\Gamma_n) :A_n:=t_n\] +The terms are obtained by projections from this set of declarations +and are written +\[\kw{fix}~f_1 (\Gamma_1) :A_1:=t_1~\kw{with} \ldots \kw{with}~ f_n +(\Gamma_n) :A_n:=t_n~\kw{for}~f_i\] +In the inference rules, we represent such a +term by +\[\Fix{f_i}{f_1:A_1':=t_1' \ldots f_n:A_n':=t_n'}\] +with $t_i'$ (resp. $A_i'$) representing the term $t_i$ abstracted +(resp. generalized) with +respect to the bindings in the context $\Gamma_i$, namely +$t_i'=\lb \Gamma_i \mto t_i$ and $A_i'=\forall \Gamma_i, A_i$. + +\subsubsection{Typing rule} +The typing rule is the expected one for a fixpoint. + +\begin{description} +\item[Fix] \index{Typing rules!Fix} +\inference{\frac{(\WTEG{A_i}{s_i})_{i=1\ldots n}~~~~ + (\WTE{\Gamma,f_1:A_1,\ldots,f_n:A_n}{t_i}{A_i})_{i=1\ldots n}} + {\WTEG{\Fix{f_i}{f_1:A_1:=t_1 \ldots f_n:A_n:=t_n}}{A_i}}} +\end{description} + +Any fixpoint definition cannot be accepted because non-normalizing terms +will lead to proofs of absurdity. + +The basic scheme of recursion that should be allowed is the one needed for +defining primitive +recursive functionals. In that case the fixpoint enjoys a special +syntactic restriction, namely one of the arguments belongs to an +inductive type, the function starts with a case analysis and recursive +calls are done on variables coming from patterns and representing subterms. + +For instance in the case of natural numbers, a proof of the induction +principle of type +\[\forall P:\nat\ra\Prop, (P~\nO)\ra((n:\nat)(P~n)\ra(P~(\nS~n)))\ra +\forall n:\nat, (P~n)\] +can be represented by the term: +\[\begin{array}{l} +\lb P:\nat\ra\Prop\mto\lb f:(P~\nO)\mto \lb g:(\forall n:\nat, +(P~n)\ra(P~(\nS~n))) \mto\\ +\Fix{h}{h:\forall n:\nat, (P~n):=\lb n:\nat\mto \Case{P}{n}{f~\lb + p:\nat\mto (g~p~(h~p))}} +\end{array} +\] + +Before accepting a fixpoint definition as being correctly typed, we +check that the definition is ``guarded''. A precise analysis of this +notion can be found in~\cite{Gim94}. + +The first stage is to precise on which argument the fixpoint will be +decreasing. The type of this argument should be an inductive +definition. + +For doing this the syntax of fixpoints is extended and becomes + \[\Fix{f_i}{f_1/k_1:A_1:=t_1 \ldots f_n/k_n:A_n:=t_n}\] +where $k_i$ are positive integers. +Each $A_i$ should be a type (reducible to a term) starting with at least +$k_i$ products $\forall y_1:B_1,\ldots \forall y_{k_i}:B_{k_i}, A'_i$ +and $B_{k_i}$ +being an instance of an inductive definition. + +Now in the definition $t_i$, if $f_j$ occurs then it should be applied +to at least $k_j$ arguments and the $k_j$-th argument should be +syntactically recognized as structurally smaller than $y_{k_i}$ + + +The definition of being structurally smaller is a bit technical. +One needs first to define the notion of +{\em recursive arguments of a constructor}\index{Recursive arguments}. +For an inductive definition \Ind{\Gamma}{\Gamma_P}{\Gamma_I}{\Gamma_C}, +the type of a constructor $c$ have the form +$\forall p_1:P_1,\ldots \forall p_r:P_r, +\forall x_1:T_1, \ldots \forall x_r:T_r, (I_j~p_1\ldots +p_r~t_1\ldots t_s)$ the recursive arguments will correspond to $T_i$ in +which one of the $I_l$ occurs. + + +The main rules for being structurally smaller are the following:\\ +Given a variable $y$ of type an inductive +definition in a declaration +\Ind{\Gamma}{\Gamma_P}{\Gamma_I}{\Gamma_C} +where $\Gamma_I$ is $[I_1:A_1;\ldots;I_k:A_k]$, and $\Gamma_C$ is + $[c_1:C_1;\ldots;c_n:C_n]$. +The terms structurally smaller than $y$ are: +\begin{itemize} +\item $(t~u), \lb x:u \mto t$ when $t$ is structurally smaller than $y$ . +\item \Case{P}{c}{f_1\ldots f_n} when each $f_i$ is structurally + smaller than $y$. \\ + If $c$ is $y$ or is structurally smaller than $y$, its type is an inductive + definition $I_p$ part of the inductive + declaration corresponding to $y$. + Each $f_i$ corresponds to a type of constructor $C_q \equiv + \forall y_1:B_1, \ldots \forall y_k:B_k, (I~a_1\ldots a_k)$ + and can consequently be + written $\lb y_1:B'_1\mto \ldots \lb y_k:B'_k\mto g_i$. + ($B'_i$ is obtained from $B_i$ by substituting parameters variables) + the variables $y_j$ occurring + in $g_i$ corresponding to recursive arguments $B_i$ (the ones in + which one of the $I_l$ occurs) are structurally smaller than $y$. +\end{itemize} +The following definitions are correct, we enter them using the +{\tt Fixpoint} command as described in section~\ref{Fixpoint} and show +the internal representation. +\begin{coq_example} +Fixpoint plus (n m:nat) {struct n} : nat := + match n with + | O => m + | S p => S (plus p m) + end. +Print plus. +Fixpoint lgth (A:Set) (l:list A) {struct l} : nat := + match l with + | nil => O + | cons a l' => S (lgth A l') + end. +Print lgth. +Fixpoint sizet (t:tree) : nat := let (f) := t in S (sizef f) + with sizef (f:forest) : nat := + match f with + | emptyf => O + | consf t f => plus (sizet t) (sizef f) + end. +Print sizet. +\end{coq_example} + + +\subsubsection{Reduction rule} +\index{iota-reduction@$\iota$-reduction} +Let $F$ be the set of declarations: $f_1/k_1:A_1:=t_1 \ldots +f_n/k_n:A_n:=t_n$. +The reduction for fixpoints is: +\[ (\Fix{f_i}{F}~a_1\ldots +a_{k_i}) \triangleright_{\iota} \substs{t_i}{f_k}{\Fix{f_k}{F}}{k=1\ldots n}\] +when $a_{k_i}$ starts with a constructor. +This last restriction is needed in order to keep strong normalization +and corresponds to the reduction for primitive recursive operators. + +We can illustrate this behavior on examples. +\begin{coq_example} +Goal forall n m:nat, plus (S n) m = S (plus n m). +reflexivity. +Abort. +Goal forall f:forest, sizet (node f) = S (sizef f). +reflexivity. +Abort. +\end{coq_example} +But assuming the definition of a son function from \tree\ to \forest: +\begin{coq_example} +Definition sont (t:tree) : forest + := let (f) := t in f. +\end{coq_example} +The following is not a conversion but can be proved after a case analysis. +\begin{coq_eval} +(******************************************************************) +(** Error: Impossible to unify .... **) +\end{coq_eval} +\begin{coq_example} +Goal forall t:tree, sizet t = S (sizef (sont t)). +reflexivity. (** this one fails **) +destruct t. +reflexivity. +\end{coq_example} +\begin{coq_eval} +Abort. +\end{coq_eval} + +% La disparition de Program devrait rendre la construction Match obsolete +% \subsubsection{The {\tt Match \ldots with \ldots end} expression} +% \label{Matchexpr} +% %\paragraph{A unary {\tt Match\ldots with \ldots end}.} +% \index{Match...with...end@{\tt Match \ldots with \ldots end}} +% The {\tt Match} operator which was a primitive notion in older +% presentations of the Calculus of Inductive Constructions is now just a +% macro definition which generates the good combination of {\tt Case} +% and {\tt Fix} operators in order to generate an operator for primitive +% recursive definitions. It always considers an inductive definition as +% a single inductive definition. + +% The following examples illustrates this feature. +% \begin{coq_example} +% Definition nat_pr : (C:Set)C->(nat->C->C)->nat->C +% :=[C,x,g,n]Match n with x g end. +% Print nat_pr. +% \end{coq_example} +% \begin{coq_example} +% Definition forest_pr +% : (C:Set)C->(tree->forest->C->C)->forest->C +% := [C,x,g,n]Match n with x g end. +% \end{coq_example} + +% Cet exemple faisait error (HH le 12/12/96), j'ai change pour une +% version plus simple +%\begin{coq_example} +%Definition forest_pr +% : (P:forest->Set)(P emptyf)->((t:tree)(f:forest)(P f)->(P (consf t f))) +% ->(f:forest)(P f) +% := [C,x,g,n]Match n with x g end. +%\end{coq_example} + +\subsubsection{Mutual induction} + +The principles of mutual induction can be automatically generated +using the {\tt Scheme} command described in section~\ref{Scheme}. + +\section{Coinductive types} +The implementation contains also coinductive definitions, which are +types inhabited by infinite objects. +More information on coinductive definitions can be found +in~\cite{Gimenez95b,Gim98}. +%They are described inchapter~\ref{Coinductives}. + +\section{\iCIC : the Calculus of Inductive Construction with + impredicative \Set}\label{impredicativity} + +\Coq{} can be used as a type-checker for \iCIC{}, the original +Calculus of Inductive Constructions with an impredicative sort \Set{} +by using the compiler option \texttt{-impredicative-set}. + +For example, using the ordinary \texttt{coqtop} command, the following +is rejected. +\begin{coq_eval} +(** This example should fail ******************************* + Error: The term forall X:Set, X -> X has type Type + while it is expected to have type Set +***) +\end{coq_eval} +\begin{coq_example} +Definition id: Set := forall X:Set,X->X. +\end{coq_example} +while it will type-check, if one use instead the \texttt{coqtop + -impredicative-set} command. + +The major change in the theory concerns the rule for product formation +in the sort \Set, which is extended to a domain in any sort~: +\begin{description} +\item [Prod] \index{Typing rules!Prod (impredicative Set)} +\inference{\frac{\WTEG{T}{s}~~~~s\in\Sort~~~~~~ + \WTE{\Gamma::(x:T)}{U}{\Set}} + { \WTEG{\forall~x:T,U}{\Set}}} +\end{description} +This extension has consequences on the inductive definitions which are +allowed. +In the impredicative system, one can build so-called {\em large inductive + definitions} like the example of second-order existential +quantifier (\texttt{exSet}). + +There should be restrictions on the eliminations which can be +performed on such definitions. The eliminations rules in the +impredicative system for sort \Set{} become~: +\begin{description} +\item[\Set] \inference{\frac{s \in + \{\Prop, \Set\}}{\compat{I:\Set}{I\ra s}} +~~~~\frac{I \mbox{~is a small inductive definition}~~~~s \in + \{\Type(i)\}} + {\compat{I:\Set}{I\ra s}}} +\end{description} + + + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: + + diff --git a/doc/refman/RefMan-coi.tex b/doc/refman/RefMan-coi.tex new file mode 100755 index 0000000000..120c1201ef --- /dev/null +++ b/doc/refman/RefMan-coi.tex @@ -0,0 +1,406 @@ +%\documentstyle[11pt,../tools/coq-tex/coq]{article} +%\input{title} + +%\include{macros} +%\begin{document} + +%\coverpage{Co-inductive types in Coq}{Eduardo Gim\'enez} +\chapter{Co-inductive types in Coq}\label{Coinductives} + +%\begin{abstract} +{\it Co-inductive} types are types whose elements may not be well-founded. +A formal study of the Calculus of Constructions extended by +co-inductive types has been presented +in \cite{Gim94}. It is based on the notion of +{\it guarded definitions} introduced by Th. Coquand +in \cite{Coquand93}. The implementation is by E. Gim\'enez. +%\end{abstract} + +\section{A short introduction to co-inductive types} + +We assume that the reader is rather familiar with inductive types. +These types are characterized by their {\it constructors}, which can be +regarded as the basic methods from which the elements +of the type can be built up. It is implicit in the definition +of an inductive type that +its elements are the result of a {\it finite} number of +applications of its constructors. Co-inductive types arise from +relaxing this implicit condition and admitting that an element of +the type can also be introduced by a non-ending (but effective) process +of construction defined in terms of the basic methods which characterize the +type. So we could think in the wider notion of types defined by +constructors (let us call them {\it recursive types}) and classify +them into inductive and co-inductive ones, depending on whether or not +we consider non-ending methods as admissible for constructing elements +of the type. Note that in both cases we obtain a ``closed type'', all whose +elements are pre-determined in advance (by the constructors). When we +know that $a$ is an element of a recursive type (no matter if it is +inductive or co-inductive) what we know is that it is the result of applying +one of the basic forms of construction allowed for the type. +So the more primitive way of eliminating an element of a recursive type is +by case analysis, i.e. by considering through which constructor it could have +been introduced. In the case of inductive sets, the additional knowledge that +constructors can be applied only a finite number of times provide +us with a more powerful way of eliminating their elements, say, +the principle of +induction. This principle is obviously not valid for co-inductive types, +since it is just the expression of this extra knowledge attached to inductive +types. + + +An example of a co-inductive type is the type of infinite sequences formed with +elements of type $A$, or streams for shorter. In Coq, +it can be introduced using the \verb!CoInductive! command~: +\begin{coq_example} +CoInductive Stream (A:Set) : Set := + cons : A -> Stream A -> Stream A. +\end{coq_example} + +The syntax of this command is the same as the +command \verb!Inductive! (cf. section +\ref{gal_Inductive_Definitions}). +Definition of mutually coinductive types are possible. + +As was already said, there are not principles of +induction for co-inductive sets, the only way of eliminating these +elements is by case analysis. +In the example of streams, this elimination principle can be +used for instance to define the well known +destructors on streams $\hd : (\Str\;A)\rightarrow A$ +and $\tl: (\Str\;A)\rightarrow (\Str\;A)$ : +\begin{coq_example} +Section Destructors. +Variable A : Set. +Definition hd (x:Stream A) := match x with + | cons a s => a + end. +Definition tl (x:Stream A) := match x with + | cons a s => s + end. +\end{coq_example} +\begin{coq_example*} +End Destructors. +\end{coq_example*} + +\subsection{Non-ending methods of construction} + +At this point the reader should have realized that we have left unexplained +what is a ``non-ending but effective process of +construction'' of a stream. In the widest sense, a +method is a non-ending process of construction if we can eliminate the +stream that it introduces, in other words, if we can reduce +any case analysis on it. In this sense, the following ways of +introducing a stream are not acceptable. +\begin{center} +$\zeros = (\cons\;\nat\;\nO\;(\tl\;\zeros))\;\;:\;\;(\Str\;\nat)$\\[12pt] +$\filter\;(\cons\;A\;a\;s) = \si\;\;(P\;a)\;\;\alors\;\;(\cons\;A\;a\;(\filter\;s))\;\;\sinon\;\;(\filter\;s) )\;\;:\;\;(\Str\;A)$ +\end{center} +\noindent The former it is not valid since the stream can not be eliminated +to obtain its tail. In the latter, a stream is naively defined as +the result of erasing from another (arbitrary) stream +all the elements which does not verify a certain property $P$. This +does not always makes sense, for example it does not when all the elements +of the stream verify $P$, in which case we can not eliminate it to +obtain its head\footnote{Note that there is no notion of ``the empty +stream'', a stream is always infinite and build by a \texttt{cons}.}. +On the contrary, the following definitions are acceptable methods for +constructing a stream~: +\begin{center} +$\zeros = (\cons\;\nat\;\nO\;\zeros)\;\;:\;\;(\Str\;\nat)\;\;\;(*)$\\[12pt] +$(\from\;n) = (\cons\;\nat\;n\;(\from\;(\nS\;n)))\;:\;(\Str\;\nat)$\\[12pt] +$\alter = (\cons\;\bool\;\true\;(\cons\;\bool\;\false\;\alter))\;:\;(\Str\;\bool)$. +\end{center} +\noindent The first one introduces a stream containing all the natural numbers +greater than a given one, and the second the stream which infinitely +alternates the booleans true and false. + +In general it is not evident to realise when a definition can +be accepted or not. However, there is a class of definitions that +can be easily recognised as being valid : those +where (1) all the recursive calls of the method are done +after having explicitly mentioned which is (at least) the first constructor +to start building the element, and (2) no other +functions apart from constructors are applied to recursive calls. +This class of definitions is usually +referred as {\it guarded-by-constructors} +definitions \cite{Coquand93,Gim94}. +The methods $\from$ +and $\alter$ are examples of definitions which are guarded by constructors. +The definition of function $\filter$ is not, because there is no +constructor to guard +the recursive call in the {\it else} branch. Neither is the one of +$\zeros$, since there is function applied to the recursive call +which is not a constructor. However, there is a difference between +the definition of $\zeros$ and $\filter$. The former may be seen as a +wrong way of characterising an object which makes sense, and it can +be reformulated in an admissible way using the equation (*). On the contrary, +the definition of +$\filter$ can not be patched, since is the idea itself +of traversing an infinite +construction searching for an element whose existence is not ensured +which does not make sense. + + + +Guarded definitions are exactly the kind of non-ending process of +construction which are allowed in Coq. The way of introducing +a guarded definition in Coq is using the special command +{\tt CoFixpoint}. This command verifies that the definition introduces an +element of a co-inductive type, and checks if it is guarded by constructors. +If we try to +introduce the definitions above, $\from$ and $\alter$ will be accepted, +while $\zeros$ and $\filter$ will be rejected giving some explanation +about why. +\begin{coq_example} +CoFixpoint zeros : Stream nat := cons nat 0%N (tl nat zeros). +CoFixpoint zeros : Stream nat := cons nat 0%N zeros. +CoFixpoint from (n:nat) : Stream nat := cons nat n (from (S n)). +\end{coq_example} + +As in the \verb!Fixpoint! command (cf. section~\ref{Fixpoint}), it is possible +to introduce a block of mutually dependent methods. The general syntax +for this case is : + +{\tt CoFixpoint {\ident$_1$} :{\term$_1$} := {\term$_1'$}\\ + with\\ + \mbox{}\hspace{0.1cm} $\ldots$ \\ + with {\ident$_m$} : {\term$_m$} := {\term$_m'$}} + + +\subsection{Non-ending methods and reduction} + +The elimination of a stream introduced by a \verb!CoFixpoint! definition +is done lazily, i.e. its definition can be expanded only when it occurs +at the head of an application which is the argument of a case expression. +Isolately it is considered as a canonical expression which +is completely evaluated. We can test this using the command \verb!compute! +to calculate the normal forms of some terms~: +\begin{coq_example} +Eval compute in (from 0). +Eval compute in (hd nat (from 0)). +Eval compute in (tl nat (from 0)). +\end{coq_example} +\noindent Thus, the equality +$(\from\;n)\equiv(\cons\;\nat\;n\;(\from \; (\S\;n)))$ +does not hold as definitional one. Nevertheless, it can be proved +as a propositional equality, in the sense of Leibniz's equality. +The version {\it à la Leibniz} of the equality above follows from +a general lemma stating that eliminating and then re-introducing a stream +yields the same stream. +\begin{coq_example} +Lemma unfold_Stream : + forall x:Stream nat, x = match x with + | cons a s => cons nat a s + end. +\end{coq_example} + +\noindent The proof is immediate from the analysis of +the possible cases for $x$, which transforms +the equality in a trivial one. + +\begin{coq_example} +olddestruct x. +trivial. +\end{coq_example} +\begin{coq_eval} +Qed. +\end{coq_eval} +The application of this lemma to $(\from\;n)$ puts this +constant at the head of an application which is an argument +of a case analysis, forcing its expansion. +We can test the type of this application using Coq's command \verb!Check!, +which infers the type of a given term. +\begin{coq_example} +Check (fun n:nat => unfold_Stream (from n)). +\end{coq_example} + \noindent Actually, The elimination of $(\from\;n)$ has actually +no effect, because it is followed by a re-introduction, +so the type of this application is in fact +definitionally equal to the +desired proposition. We can test this computing +the normal form of the application above to see its type. +\begin{coq_example} +Transparent unfold_Stream. +Eval compute in (fun n:nat => unfold_Stream (from n)). +\end{coq_example} + + +\section{Reasoning about infinite objects} + +At a first sight, it might seem that +case analysis does not provide a very powerful way +of reasoning about infinite objects. In fact, what we can prove about +an infinite object using +only case analysis is just what we can prove unfolding its method +of construction a finite number of times, which is not always +enough. Consider for example the following method for appending +two streams~: +\begin{coq_example} +Variable A : Set. +CoFixpoint conc (s1 s2:Stream A) : Stream A := + cons A (hd A s1) (conc (tl A s1) s2). +\end{coq_example} + +Informally speaking, we expect that for all pair of streams $s_1$ and $s_2$, +$(\conc\;s_1\;s_2)$ +defines the ``the same'' stream as $s_1$, +in the sense that if we would be able to unfold the definition +``up to the infinite'', we would obtain definitionally equal normal forms. +However, no finite unfolding of the definitions gives definitionally +equal terms. Their equality can not be proved just using case analysis. + + +The weakness of the elimination principle proposed for infinite objects +contrast with the power provided by the inductive +elimination principles, but it is not actually surprising. It just means +that we can not expect to prove very interesting things about infinite +objects doing finite proofs. To take advantage of infinite objects we +have to consider infinite proofs as well. For example, +if we want to catch up the equality between $(\conc\;s_1\;s_2)$ and +$s_1$ we have to introduce first the type of the infinite proofs +of equality between streams. This is a +co-inductive type, whose elements are build up from a +unique constructor, requiring a proof of the equality of the +heads of the streams, and an (infinite) proof of the equality +of their tails. + +\begin{coq_example} +CoInductive EqSt : Stream A -> Stream A -> Prop := + eqst : + forall s1 s2:Stream A, + hd A s1 = hd A s2 -> EqSt (tl A s1) (tl A s2) -> EqSt s1 s2. +\end{coq_example} +\noindent Now the equality of both streams can be proved introducing +an infinite object of type + +\noindent $(\EqSt\;s_1\;(\conc\;s_1\;s_2))$ by a \verb!CoFixpoint! +definition. +\begin{coq_example} +CoFixpoint eqproof (s1 s2:Stream A) : EqSt s1 (conc s1 s2) := + eqst s1 (conc s1 s2) (refl_equal (hd A (conc s1 s2))) + (eqproof (tl A s1) s2). +\end{coq_example} +\begin{coq_eval} +Reset eqproof. +\end{coq_eval} +\noindent Instead of giving an explicit definition, +we can use the proof editor of Coq to help us in +the construction of the proof. +A tactic \verb!Cofix! allows to place a \verb!CoFixpoint! definition +inside a proof. +This tactic introduces a variable in the context which has +the same type as the current goal, and its application stands +for a recursive call in the construction of the proof. If no name is +specified for this variable, the name of the lemma is chosen by +default. +%\pagebreak + +\begin{coq_example} +Lemma eqproof : forall s1 s2:Stream A, EqSt s1 (conc s1 s2). +cofix. +\end{coq_example} + +\noindent An easy (and wrong!) way of finishing the proof is just to apply the +variable \verb!eqproof!, which has the same type as the goal. + +\begin{coq_example} +intros. +apply eqproof. +\end{coq_example} + +\noindent The ``proof'' constructed in this way +would correspond to the \verb!CoFixpoint! definition +\begin{coq_example*} +CoFixpoint eqproof : forall s1 s2:Stream A, EqSt s1 (conc s1 s2) := + eqproof. +\end{coq_example*} + +\noindent which is obviously non-guarded. This means that +we can use the proof editor to +define a method of construction which does not make sense. However, +the system will never accept to include it as part of the theory, +because the guard condition is always verified before saving the proof. + +\begin{coq_example} +Qed. +\end{coq_example} + +\noindent Thus, the user must be careful in the +construction of infinite proofs +with the tactic \verb!Cofix!. Remark that once it has been used +the application of tactics performing automatic proof search in +the environment (like for example \verb!Auto!) +could introduce unguarded recursive calls in the proof. +The command \verb!Guarded! allows to verify +if the guarded condition has been violated +during the construction of the proof. This command can be +applied even if the proof term is not complete. + + + +\begin{coq_example} +Restart. +cofix. +auto. +Guarded. +Undo. +Guarded. +\end{coq_example} + +\noindent To finish with this example, let us restart from the +beginning and show how to construct an admissible proof~: + +\begin{coq_example} +Restart. + cofix. +\end{coq_example} + +%\pagebreak + +\begin{coq_example} +intros. +apply eqst. +trivial. +simpl. +apply eqproof. +Qed. +\end{coq_example} + + +\section{Experiments with co-inductive types} + +Some examples involving co-inductive types are available with +the distributed system, in the theories library and in the contributions +of the Lyon site. Here we present a short description of their contents~: +\begin{itemize} +\item Directory \verb!theories/LISTS! : + \begin{itemize} + \item File \verb!Streams.v! : The type of streams and the +extensional equality between streams. + \end{itemize} + +\item Directory \verb!contrib/Lyon/COINDUCTIVES! : + \begin{itemize} + \item Directory \verb!ARITH! : An arithmetic where $\infty$ +is an explicit constant of the language instead of a metatheoretical notion. + \item Directory \verb!STREAM! : + \begin{itemize} + \item File \verb!Examples! : +Several examples of guarded definitions, as well as +of frequent errors in the introduction of a stream. A different +way of defining the extensional equality of two streams, +and the proofs showing that it is equivalent to the one in \verb!theories!. + \item File \verb!Alter.v! : An example showing how +an infinite proof introduced by a guarded definition can be also described +using an operator of co-recursion \cite{Gimenez95b}. + \end{itemize} +\item Directory \verb!PROCESSES! : A proof of the alternating +bit protocol based on Pra\-sad's Calculus of Broadcasting Systems \cite{Prasad93}, +and the verification of an interpreter for this calculus. +See \cite{Gimenez95b} for a complete description about this development. + \end{itemize} +\end{itemize} + +%\end{document} + +% $Id$ diff --git a/doc/refman/RefMan-com.tex b/doc/refman/RefMan-com.tex new file mode 100755 index 0000000000..730100eedd --- /dev/null +++ b/doc/refman/RefMan-com.tex @@ -0,0 +1,280 @@ +\chapter{The \Coq~commands}\label{Addoc-coqc} +\ttindex{coqtop} +\ttindex{coqc} + +There are two \Coq~commands: +\begin{itemize} +\item {\tt coqtop}: The \Coq\ toplevel (interactive mode) ; +\item {\tt coqc} : The \Coq\ compiler (batch compilation). +\end{itemize} +The options are (basically) the same for the two commands, and +roughly described below. You can also look at the \verb!man! pages of +\verb!coqtop! and \verb!coqc! for more details. + + +\section{Interactive use ({\tt coqtop})} + +In the interactive mode, also known as the \Coq~toplevel, the user can +develop his theories and proofs step by step. The \Coq~toplevel is +run by the command {\tt coqtop}. + +\index{byte-code} +\index{native code} +\label{binary-images} +They are two different binary images of \Coq: the byte-code one and +the native-code one (if Objective Caml provides a native-code compiler +for your platform, which is supposed in the following). When invoking +\verb!coqtop! or \verb!coqc!, the native-code version of the system is +used. The command-line options \verb!-byte! and \verb!-opt! explicitly +select the byte-code and the native-code versions, respectively. + +The byte-code toplevel is based on a Caml +toplevel (to allow the dynamic link of tactics). You can switch to +the Caml toplevel with the command \verb!Drop.!, and come back to the +\Coq~toplevel with the command \verb!Toplevel.loop();;!. + +% The command \verb!coqtop -searchisos! runs the search tool {\sf +% Coq\_SearchIsos} (see section~\ref{coqsearchisos}, +% page~\pageref{coqsearchisos}) and, as the \Coq~system, can be combined +% with the option \verb!-opt!. + +\section{Batch compilation ({\tt coqc})} +The {\tt coqc} command takes a name {\em file} as argument. Then it +looks for a vernacular file named {\em file}{\tt .v}, and tries to +compile it into a {\em file}{\tt .vo} file (See ~\ref{compiled}). + +\Warning The name {\em file} must be a regular {\Coq} identifier, as +defined in the section \ref{lexical}. It +must only contain letters, digits or underscores +(\_). Thus it can be \verb+/bar/foo/toto.v+ but cannot be +\verb+/bar/foo/to-to.v+ . + +Notice that the \verb!-byte! and \verb!-opt! options are still +available with \verb!coqc! and allow you to select the byte-code or +native-code versions of the system. + + +\section{Resource file} +\index{Resource file} + +When \Coq\ is launched, with either {\tt coqtop} or {\tt coqc}, the +resource file \verb:$HOME/.coqrc.7.0: is loaded, where \verb:$HOME: is +the home directory of the user. If this file is not found, then the +file \verb:$HOME/.coqrc: is searched. You can also specify an +arbitrary name for the resource file (see option \verb:-init-file: +below), or the name of another user to load the resource file of +someone else (see option \verb:-user:). + +This file may contain, for instance, \verb:Add LoadPath: commands to add +directories to the load path of \Coq. +It is possible to skip the loading of the resource file with the +option \verb:-q:. + +\section{Environment variables} +\label{EnvVariables} +\index{Environment variables} + +There are three environment variables used by the \Coq\ system. +\verb:$COQBIN: for the directory where the binaries are, +\verb:$COQLIB: for the directory whrer the standard library is, and +\verb:$COQTOP: for the directory of the sources. The latter is useful +only for developers that are writing their own tactics and are using +\texttt{coq\_makefile} (see \ref{Makefile}). If \verb:$COQBIN: or +\verb:$COQLIB: are not defined, \Coq\ will use the default values +(defined at installation time). So these variables are useful only if +you move the \Coq\ binaries and library after installation. + +\section{Options} +\index{Options of the command line} + +The following command-line options are recognized by the commands {\tt + coqc} and {\tt coqtop}, unless stated otherwise: + +\begin{description} +\item[{\tt -byte}]\ + + Run the byte-code version of \Coq{}. + +\item[{\tt -opt}]\ + + Run the native-code version of \Coq{}. + +\item[{\tt -I} {\em directory}, {\tt -include} {\em directory}]\ + + Add {\em directory} to the searched directories when looking for a + file. + +\item[{\tt -R} {\em directory} {\dirpath}]\ + + This maps the subdirectory structure of physical {\em directory} to + logical {\dirpath} and adds {\em directory} and its subdirectories + to the searched directories when looking for a file. + +\item[{\tt -top} {\dirpath}]\ + + This sets the toplevel module name to {\dirpath} instead of {\tt + Top}. Not valid for {\tt coqc}. + +\item[{\tt -is} {\em file}, {\tt -inputstate} {\em file}]\ + + Cause \Coq~to use the state put in the file {\em file} as its input + state. The default state is {\em initial.coq}. + Mainly useful to build the standard input state. + +\item[{\tt -outputstate} {\em file}]\ + + Cause \Coq~to dump its state to file {\em file}.coq just after finishing + parsing and evaluating all the arguments from the command line. + +\item[{\tt -nois}]\ + + Cause \Coq~to begin with an empty state. Mainly useful to build the + standard input state. + +%Obsolete? +% +%\item[{\tt -notactics}]\ +% +% Forbid the dynamic loading of tactics in the bytecode version of {\Coq}. + +\item[{\tt -init-file} {\em file}]\ + + Take {\em file} as the resource file. + +\item[{\tt -q}]\ + + Cause \Coq~not to load the resource file. + +\item[{\tt -user} {\em username}]\ + + Take resource file of user {\em username} (that is + \verb+~+{\em username}{\tt /.coqrc.7.0}) instead of yours. + +\item[{\tt -load-ml-source} {\em file}]\ + + Load the Caml source file {\em file}. + +\item[{\tt -load-ml-object} {\em file}]\ + + Load the Caml object file {\em file}. + +\item[{\tt -l} {\em file}, {\tt -load-vernac-source} {\em file}]\ + + Load \Coq~file {\em file}{\tt .v} + +\item[{\tt -lv} {\em file}, {\tt -load-vernac-source-verbose} {\em file}]\ + + Load \Coq~file {\em file}{\tt .v} with + a copy of the contents of the file on standard input. + +\item[{\tt -load-vernac-object} {\em file}]\ + + Load \Coq~compiled file {\em file}{\tt .vo} + +%\item[{\tt -preload} {\em file}]\ \\ +%Add {\em file}{\tt .vo} to the files to be loaded and opened +%before making the initial state. +% +\item[{\tt -require} {\em file}]\ + + Load \Coq~compiled file {\em file}{\tt .vo} and import it ({\tt + Require} {\em file}). + +\item[{\tt -compile} {\em file}]\ + + This compiles file {\em file}{\tt .v} into {\em file}{\tt .vo}. + This option implies options {\tt -batch} and {\tt -silent}. It is + only available for {\tt coqtop}. + +\item[{\tt -compile-verbose} {\em file}]\ + + This compiles file {\em file}{\tt .v} into {\em file}{\tt .vo} with + a copy of the contents of the file on standard input. + This option implies options {\tt -batch} and {\tt -silent}. It is + only available for {\tt coqtop}. + +\item[{\tt -verbose}]\ + + This option is only for {\tt coqc}. It tells to compile the file with + a copy of its contents on standard input. + +\item[{\tt -batch}]\ + + Batch mode : exit just after arguments parsing. This option is only + used by {\tt coqc}. + +%Mostly unused in the code +%\item[{\tt -debug}]\ +% +% Switch on the debug flag. + +\item[{\tt -xml}]\ + + This option is for use with {\tt coqc}. It tells \Coq\ to export on + the standard output the content of the compiled file into XML format. + +\item[{\tt -quality}] + + Improve the legibility of the proof terms produced by some tactics. + +\item[{\tt -emacs}]\ + + Tells \Coq\ it is executed under Emacs. + +\item[{\tt -impredicative-set}]\ + + Change the logical theory of {\Coq} by declaring the sort {\tt Set} + impredicative; warning: this is known to be inconsistent with + some standard axioms of classical mathematics such as the functional + axiom of choice or the principle of description + +\item[{\tt -dump-glob} {\em file}]\ + + This dumps references for global names in file {\em file} + (to be used by coqdoc, see~\ref{coqdoc}) + +\item[{\tt -dont-load-proofs}]\ + + This avoids loading in memory the proofs of opaque theorems + resulting in a smaller memory requirement and faster compilation; + warning: this invalidates some features such as the extraction tool. + +\item[{\tt -image} {\em file}]\ + + This option sets the binary image to be used to be {\em file} + instead of the standard one. Not of general use. + +\item[{\tt -bindir} {\em directory}]\ + + Set for {\tt coqc} the directory containing \Coq\ binaries. + It is equivalent to do \texttt{export COQBIN=}{\em directory} + before lauching {\tt coqc}. + +\item[{\tt -where}]\ + + Print the \Coq's standard library location and exit. + +\item[{\tt -v}]\ + + Print the \Coq's version and exit. + +\item[{\tt -h}, {\tt --help}]\ + + Print a short usage and exit. + +\end{description} + +% {\tt coqtop} has an additional option: + +% \begin{description} +% \item[{\tt -searchisos}]\ \\ +% Launch the {\sf Coq\_SearchIsos} toplevel +% (see section~\ref{coqsearchisos}, page~\pageref{coqsearchisos}). +% \end{description} + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-ext.tex b/doc/refman/RefMan-ext.tex new file mode 100644 index 0000000000..503d7571d6 --- /dev/null +++ b/doc/refman/RefMan-ext.tex @@ -0,0 +1,1173 @@ +\chapter{Extensions of \Gallina{}} +\label{Gallina-extension}\index{Gallina} + +{\gallina} is the kernel language of {\Coq}. We describe here extensions of +the Gallina's syntax. + +\section{Record types +\comindex{Record} +\label{Record}} + +The \verb+Record+ construction is a macro allowing the definition of +records as is done in many programming languages. Its syntax is +described on figure \ref{record-syntax}. In fact, the \verb+Record+ +macro is more general than the usual record types, since it allows +also for ``manifest'' expressions. In this sense, the \verb+Record+ +construction allows to define ``signatures''. + +\begin{figure}[h] +\begin{centerframe} +\begin{tabular}{lcl} +{\sentence} & ++= & {\record}\\ + & & \\ +{\record} & ::= & + {\tt Record} {\ident} \sequence{\binderlet}{} {\tt :} {\sort} \verb.:=. \\ +&& ~~~~\zeroone{\ident} + \verb!{! \zeroone{\nelist{\field}{;}} \verb!}! \verb:.:\\ + & & \\ +{\field} & ::= & {\name} : {\type} \\ + & $|$ & {\name} {\typecstr} := {\term} +\end{tabular} +\end{centerframe} +\caption{Syntax for the definition of {\tt Record}} +\label{record-syntax} +\end{figure} + +\noindent In the expression + +\smallskip +{\tt Record} {\ident} {\params} \texttt{:} + {\sort} := {\ident$_0$} \verb+{+ + {\ident$_1$} \texttt{:} {\term$_1$}; + \dots + {\ident$_n$} \texttt{:} {\term$_n$} \verb+}+. +\smallskip + +\noindent the identifier {\ident} is the name of the defined record +and {\sort} is its type. The identifier {\ident$_0$} is the name of +its constructor. If {\ident$_0$} is omitted, the default name {\tt +Build\_{\ident}} is used. The identifiers {\ident$_1$}, .., +{\ident$_n$} are the names of fields and {\term$_1$}, .., {\term$_n$} +their respective types. Remark that the type of {\ident$_i$} may +depend on the previous {\ident$_j$} (for $j bottom; + Rat_irred_cond : + forall x y z:nat, (x * y) = top /\ (x * z) = bottom -> x = 1}. +\end{coq_example} + +Remark here that the field +\verb+Rat_cond+ depends on the field \verb+bottom+. + +%Let us now see the work done by the {\tt Record} macro. +%First the macro generates an inductive definition +%with just one constructor: +% +%\medskip +%\noindent +%{\tt Inductive {\ident} {\binderlet} : {\sort} := \\ +%\mbox{}\hspace{0.4cm} {\ident$_0$} : forall ({\ident$_1$}:{\term$_1$}) .. +%({\ident$_n$}:{\term$_n$}), {\ident} {\rm\sl params}.} +%\medskip + +Let us now see the work done by the {\tt Record} macro. First the +macro generates an inductive definition with just one constructor: +\begin{quote} +{\tt Inductive {\ident} {\params} :{\sort} :=} \\ +\qquad {\tt + {\ident$_0$} ({\ident$_1$}:{\term$_1$}) .. ({\ident$_n$}:{\term$_n$}).} +\end{quote} +To build an object of type {\ident}, one should provide the +constructor {\ident$_0$} with $n$ terms filling the fields of +the record. + +As an example, let us define the rational $1/2$: +\begin{coq_example*} +Require Import Arith. +Theorem one_two_irred : + forall x y z:nat, x * y = 1 /\ x * z = 2 -> x = 1. +\end{coq_example*} +\begin{coq_eval} +Lemma mult_m_n_eq_m_1 : forall m n:nat, m * n = 1 -> m = 1. +destruct m; trivial. +intros; apply f_equal with (f := S). +destruct m; trivial. +destruct n; simpl in H. + rewrite <- mult_n_O in H. + discriminate. + rewrite <- plus_n_Sm in H. + discriminate. +Qed. + +intros x y z [H1 H2]. + apply mult_m_n_eq_m_1 with (n := y); trivial. +\end{coq_eval} +\ldots +\begin{coq_example*} +Qed. +\end{coq_example*} +\begin{coq_example} +Definition half := mkRat true 1 2 (O_S 1) one_two_irred. +\end{coq_example} +\begin{coq_example} +Check half. +\end{coq_example} + +The macro generates also, when it is possible, the projection +functions for destructuring an object of type {\ident}. These +projection functions have the same name that the corresponding +fields. If a field is named ``\verb=_='' then no projection is built +for it. In our example: + +\begin{coq_example} +Eval compute in half.(top). +Eval compute in half.(bottom). +Eval compute in half.(Rat_bottom_cond). +\end{coq_example} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} + +\begin{Warnings} +\item {\tt Warning: {\ident$_i$} cannot be defined.} + + It can happen that the definition of a projection is impossible. + This message is followed by an explanation of this impossibility. + There may be three reasons: + \begin{enumerate} + \item The name {\ident$_i$} already exists in the environment (see + Section~\ref{Axiom}). + \item The body of {\ident$_i$} uses an incorrect elimination for + {\ident} (see Sections~\ref{Fixpoint} and~\ref{Caseexpr}). + \item The type of the projections {\ident$_i$} depends on previous + projections which themselves couldn't be defined. + \end{enumerate} +\end{Warnings} + +\begin{ErrMsgs} + +\item \errindex{A record cannot be recursive} + + The record name {\ident} appears in the type of its fields. + +\item During the definition of the one-constructor inductive + definition, all the errors of inductive definitions, as described in + Section~\ref{gal_Inductive_Definitions}, may also occur. + +\end{ErrMsgs} + +\SeeAlso Coercions and records in Section~\ref{Coercions-and-records} +of the chapter devoted to coercions. + +\Rem {\tt Structure} is a synonym of the keyword {\tt Record}. + +\Rem An experimental syntax for projections based on a dot notation is +available. The command to activate it is +\begin{quote} +{\tt Set Printing Projections.} +\end{quote} + +\begin{figure}[t] +\begin{centerframe} +\begin{tabular}{lcl} +{\term} & ++= & {\term} {\tt .(} {\qualid} {\tt )}\\ + & $|$ & {\term} {\tt .(} {\qualid} \nelist{\termarg}{} {\tt )}\\ + & $|$ & {\term} {\tt .(} {@}{\qualid} \nelist{\term}{} {\tt )} +\end{tabular} +\end{centerframe} +\caption{Syntax of \texttt{Record} projections} +\label{fig:projsyntax} +\end{figure} + +The corresponding grammar rules are given Figure~\ref{fig:projsyntax}. +When {\qualid} denotes a projection, the syntax {\tt + {\term}.({\qualid})} is equivalent to {\qualid~\term}, the syntax +{\tt {\term}.({\qualid}~{\termarg}$_1$~ \ldots~ {\termarg}$_n$)} to +{\qualid~{\termarg}$_1$ \ldots {\termarg}$_n$~\term}, and the syntax +{\tt {\term}.(@{\qualid}~{\term}$_1$~\ldots~{\term}$_n$)} to +{@\qualid~{\term}$_1$ \ldots {\term}$_n$~\term}. In each case, {\term} +is the object projected and the other arguments are the parameters of +the inductive type. + +To deactivate the printing of projections, use +{\tt Unset Printing Projections}. + + +\section{Variants and extensions of {\tt match} +\label{Extensions-of-match} +\index{match@{\tt match\ldots with\ldots end}}} + +\subsection{Multiple and nested pattern-matching +\index{ML-like patterns} +\label{Mult-match}} + +The basic version of \verb+match+ allows pattern-matching on simple +patterns. As an extension, multiple and nested patterns are +allowed, as in ML-like languages. + +The extension just acts as a macro that is expanded during parsing +into a sequence of {\tt match} on simple patterns. Especially, a +construction defined using the extended {\tt match} is printed under +its expanded form. + +\SeeAlso chapter \ref{Mult-match-full}. + +\subsection{Pattern-matching on boolean values: the {\tt if} expression +\index{if@{\tt if ... then ... else}}} + +For inductive types with exactly two constructors and for +pattern-matchings expressions which do not depend on the arguments of +the constructors, it is possible to use a {\tt if ... then ... else} +notation. For instance, the definition + +\begin{coq_example} +Definition not (b:bool) := + match b with + | true => false + | false => true + end. +\end{coq_example} + +can be alternatively written + +\begin{coq_eval} +Reset not. +\end{coq_eval} +\begin{coq_example} +Definition not (b:bool) := if b then false else true. +\end{coq_example} + +More generally, for an inductive type with constructors {\tt C$_1$} +and {\tt C$_2$}, we have the following equivalence + +\smallskip + +{\tt if {\term} \zeroone{\ifitem} then {\term}$_1$ else {\term}$_2$} $\equiv$ +\begin{tabular}[c]{l} +{\tt match {\term} \zeroone{\ifitem} with}\\ +{\tt \verb!|! C$_1$ \_ {\ldots} \_ \verb!=>! {\term}$_1$} \\ +{\tt \verb!|! C$_2$ \_ {\ldots} \_ \verb!=>! {\term}$_2$} \\ +{\tt end} +\end{tabular} + +Here is an example. + +\begin{coq_example} +Check (fun x (H:{x=0}+{x<>0}) => + match H with + | left _ => true + | right _ => false + end). +\end{coq_example} + +Notice that the printing uses the {\tt if} syntax because {\tt sumbool} is +declared as such (see section \ref{printing-options}). + +\subsection{Irrefutable patterns: the destructuring {\tt let} +\index{let in@{\tt let ... in}} +\label{Letin}} + + + +Closed terms (that is not relying on any axiom or variable) in an +inductive type having only one constructor, say {\tt foo}, have +necessarily the form \texttt{(foo ...)}. In this case, the {\tt match} +construction can be written with a syntax close to the {\tt let ... in +...} construction. Expression {\tt let +(}~{\ident$_1$},\ldots,{\ident$_n$}~{\tt ) :=}~{\term$_0$}~{\tt +in}~{\term$_1$} performs case analysis on {\term$_0$} which must be in +an inductive type with one constructor with $n$ arguments. Variables +{\ident$_1$}\ldots{\ident$_n$} are bound to the $n$ arguments of the +constructor in expression {\term$_1$}. For instance, the definition + +\begin{coq_example} +Definition fst (A B:Set) (H:A * B) := match H with + | pair x y => x + end. +\end{coq_example} + +can be alternatively written + +\begin{coq_eval} +Reset fst. +\end{coq_eval} +\begin{coq_example} +Definition fst (A B:Set) (p:A * B) := let (x, _) := p in x. +\end{coq_example} +Note however that reduction is slightly different from regular {\tt +let ... in ...} construction since it can occur only if {\term$_0$} +can be put in constructor form. Otherwise, reduction is blocked. + +The pretty-printing of a definition by matching on a +irrefutable pattern can either be done using {\tt match} or the {\tt +let} construction (see Section~\ref{printing-options}). + +The general equivalence for an inductive type with one constructors {\tt C} is + +\smallskip +{\tt let ({\ident}$_1$,\ldots,{\ident}$_n$) \zeroone{\ifitem} := {\term} in {\term}'} \\ +$\equiv$~ +{\tt match {\term} \zeroone{\ifitem} with C {\ident}$_1$ {\ldots} {\ident}$_n$ \verb!=>! {\term}' end} + +\subsection{Options for pretty-printing of {\tt match} +\label{printing-options}} + +There are three options controlling the pretty-printing of {\tt match} +expressions. + +\subsubsection{Printing of wildcard pattern +\comindex{Set Printing Wildcard} +\comindex{Unset Printing Wildcard} +\comindex{Test Printing Wildcard}} + +Some variables in a pattern may not occur in the right-hand side of +the pattern-matching clause. There are options to control the +display of these variables. + +\begin{quote} +{\tt Set Printing Wildcard.} +\end{quote} +The variables having no occurrences in the right-hand side of the +pattern-matching clause are just printed using the wildcard symbol +``{\tt \_}''. + +\begin{quote} +{\tt Unset Printing Wildcard.} +\end{quote} +The variables, even useless, are printed using their usual name. But some +non dependent variables have no name. These ones are still printed +using a ``{\tt \_}''. + +\begin{quote} +{\tt Test Printing Wildcard.} +\end{quote} +This tells if the wildcard printing mode is on or off. The default is +to print wildcard for useless variables. + +\subsubsection{Printing of the elimination predicate +\comindex{Set Printing Synth} +\comindex{Unset Printing Synth} +\comindex{Test Printing Synth}} + +In most of the cases, the type of the result of a matched term is +mechanically synthesisable. Especially, if the result type does not +depend of the matched term. + +\begin{quote} +{\tt Set Printing Synth.} +\end{quote} +The result type is not printed when {\Coq} knows that it can +re-synthesise it. + +\begin{quote} +{\tt Unset Printing Synth.} +\end{quote} +This forces the result type to be always printed. + +\begin{quote} +{\tt Test Printing Synth.} +\end{quote} +This tells if the non-printing of synthesisable types is on or off. +The default is to not print synthesisable types. + +\subsubsection{Printing matching on irrefutable pattern +\comindex{Add Printing Let {\ident}} +\comindex{Remove Printing Let {\ident}} +\comindex{Test Printing Let {\ident}} +\comindex{Print Table Printing Let}} + +If an inductive type has just one constructor, +pattern-matching can be written using {\tt let} ... {\tt :=} +... {\tt in}~... + +\begin{quote} +{\tt Add Printing Let {\ident}.} +\end{quote} +This adds {\ident} to the list of inductive types for which +pattern-matching is written using a {\tt let} expression. + +\begin{quote} +{\tt Remove Printing Let {\ident}.} +\end{quote} +This removes {\ident} from this list. + +\begin{quote} +{\tt Test Printing Let {\ident}.} +\end{quote} +This tells if {\ident} belongs to the list. + +\begin{quote} +{\tt Print Table Printing Let.} +\end{quote} +This prints the list of inductive types for which pattern-matching is +written using a {\tt let} expression. + +The list of inductive types for which pattern-matching is written +using a {\tt let} expression is managed synchronously. This means that +it is sensible to the command {\tt Reset}. + +\subsubsection{Printing matching on booleans +\comindex{Add Printing If {\ident}} +\comindex{Remove Printing If {\ident}} +\comindex{Test Printing If {\ident}} +\comindex{Print Table Printing If}} + +If an inductive type is isomorphic to the boolean type, +pattern-matching can be written using {\tt if} ... {\tt then} ... {\tt + else} ... + +\begin{quote} +{\tt Add Printing If {\ident}.} +\end{quote} +This adds {\ident} to the list of inductive types for which +pattern-matching is written using an {\tt if} expression. + +\begin{quote} +{\tt Remove Printing If {\ident}.} +\end{quote} +This removes {\ident} from this list. + +\begin{quote} +{\tt Test Printing If {\ident}.} +\end{quote} +This tells if {\ident} belongs to the list. + +\begin{quote} +{\tt Print Table Printing If.} +\end{quote} +This prints the list of inductive types for which pattern-matching is +written using an {\tt if} expression. + +The list of inductive types for which pattern-matching is written +using an {\tt if} expression is managed synchronously. This means that +it is sensible to the command {\tt Reset}. + +\subsubsection{Example} + +This example emphasizes what the printing options offer. + +\begin{coq_example} +Test Printing Let prod. +Print fst. +Remove Printing Let prod. +Unset Printing Synth. +Unset Printing Wildcard. +Print fst. +\end{coq_example} + +% \subsection{Still not dead old notations} + +% The following variant of {\tt match} is inherited from older version +% of {\Coq}. + +% \medskip +% \begin{tabular}{lcl} +% {\term} & ::= & {\annotation} {\tt Match} {\term} {\tt with} {\terms} {\tt end}\\ +% \end{tabular} +% \medskip + +% This syntax is a macro generating a combination of {\tt match} with {\tt +% Fix} implementing a combinator for primitive recursion equivalent to +% the {\tt Match} construction of \Coq\ V5.8. It is provided only for +% sake of compatibility with \Coq\ V5.8. It is recommended to avoid it. +% (see section~\ref{Matchexpr}). + +% There is also a notation \texttt{Case} that is the +% ancestor of \texttt{match}. Again, it is still in the code for +% compatibility with old versions but the user should not use it. + +% Explained in RefMan-gal.tex +%% \section{Forced type} + +%% In some cases, one may wish to assign a particular type to a term. The +%% syntax to force the type of a term is the following: + +%% \medskip +%% \begin{tabular}{lcl} +%% {\term} & ++= & {\term} {\tt :} {\term}\\ +%% \end{tabular} +%% \medskip + +%% It forces the first term to be of type the second term. The +%% type must be compatible with +%% the term. More precisely it must be either a type convertible to +%% the automatically inferred type (see chapter \ref{Cic}) or a type +%% coercible to it, (see \ref{Coercions}). When the type of a +%% whole expression is forced, it is usually not necessary to give the types of +%% the variables involved in the term. + +%% Example: + +%% \begin{coq_example} +%% Definition ID := forall X:Set, X -> X. +%% Definition id := (fun X x => x):ID. +%% Check id. +%% \end{coq_example} + +\section{Section mechanism +\index{Sections} +\label{Section}} + +The sectioning mechanism allows to organise a proof in structured +sections. Then local declarations become available (see +Section~\ref{Simpl-definitions}). + +\subsection{\tt Section {\ident}\comindex{Section}} + +This command is used to open a section named {\ident}. + +%% Discontinued ? +%% \begin{Variants} +%% \comindex{Chapter} +%% \item{\tt Chapter {\ident}}\\ +%% Same as {\tt Section {\ident}} +%% \end{Variants} + +\subsection{\tt End {\ident} +\comindex{End}} + +This command closes the section named {\ident}. When a section is +closed, all local declarations (variables and local definitions) are +{\em discharged}. This means that all global objects defined in the +section are generalised with respect to all variables and local +definitions it depends on in the section. None of the local +declarations (considered as autonomous declarations) survive the end +of the section. + +Here is an example : +\begin{coq_example} +Section s1. +Variables x y : nat. +Let y' := y. +Definition x' := S x. +Definition x'' := x' + y'. +Print x'. +End s1. +Print x'. +Print x''. +\end{coq_example} +Notice the difference between the value of {\tt x'} and {\tt x''} +inside section {\tt s1} and outside. + +\begin{ErrMsgs} +\item \errindex{This is not the last opened section} +\end{ErrMsgs} + +\begin{Remarks} +\item Most commands, like {\tt Hint}, {\tt Notation}, option management, ... +which appear inside a section are cancelled when the +section is closed. +% cf section \ref{LongNames} +%\item Usually all identifiers must be distinct. +%However, a name already used in a closed section (see \ref{Section}) +%can be reused. In this case, the old name is no longer accessible. + +% Obsolète +%\item A module implicitly open a section. Be careful not to name a +%module with an identifier already used in the module (see \ref{compiled}). +\end{Remarks} + +\input{RefMan-mod.v} + +\section{Libraries and qualified names} + +\subsection{Names of libraries and files +\label{Libraries} +\index{Libraries} +\index{Logical paths}} + +\paragraph{Libraries} + +The theories developed in {\Coq} are stored in {\em libraries}. A +library is characterised by a name called {\it root} of the +library. The standard library of {\Coq} has root name {\tt Coq} and is +known by default when a {\Coq} session starts. + +Libraries have a tree structure. E.g., the {\tt Coq} library +contains the sub-libraries {\tt Init}, {\tt Logic}, {\tt Arith}, {\tt +Lists}, ... The ``dot notation'' is used to separate the different +component of a library name. For instance, the {\tt Arith} library of +{\Coq} standard library is written ``{\tt Coq.Arith}''. + +\medskip +\Rem no blank is allowed between the dot and the identifier on its +right, otherwise the dot is interpreted as the full stop (period) of +the command! +\medskip + +\paragraph{Physical paths vs logical paths} + +Libraries and sub-libraries are denoted by {\em logical directory +paths} (written {\dirpath} and of which the syntax is the same as +{\qualid}, see \ref{qualid}). Logical directory +paths can be mapped to physical directories of the +operating system using the command (see \ref{AddLoadPath}) +\begin{quote} +{\tt Add LoadPath {\it physical\_path} as {\dirpath}}. +\end{quote} +A library can inherit the tree structure of a physical directory by +using the {\tt -R} option to {\tt coqtop} or the +command (see \ref{AddRecLoadPath}) +\begin{quote} +{\tt Add Rec LoadPath {\it physical\_path} as {\dirpath}}. +\end{quote} + +\Rem When used interactively with {\tt coqtop} command, {\Coq} opens a +library called {\tt Top}. + +\paragraph{The file level} + +At some point, (sub-)libraries contain {\it modules} which coincide +with files at the physical level. As for sublibraries, the dot +notation is used to denote a specific module of a library. Typically, +{\tt Coq.Init.Logic} is the logical path associated to the file {\tt + Logic.v} of {\Coq} standard library. Notice that compilation (see +\ref{Addoc-coqc}) is done at the level of files. + +If the physical directory where a file {\tt File.v} lies is mapped to +the empty logical directory path (which is the default when using the +simple form of {\tt Add LoadPath} or {\tt -I} option to coqtop), then +the name of the module it defines is {\tt File}. + +\subsection{Qualified names +\label{LongNames} +\index{Qualified identifiers} +\index{Absolute names}} + +Modules contain constructions (sub-modules, axioms, parameters, +definitions, lemmas, theorems, remarks or facts). The (full) name of a +construction starts with the logical name of the module in which it is defined +followed by the (short) name of the construction. +Typically, the full name {\tt Coq.Init.Logic.eq} denotes Leibniz' equality +defined in the module {\tt Logic} in the sublibrary {\tt Init} of the +standard library of \Coq. + +\paragraph{Absolute, partially qualified and short names} + +The full name of a library, module, section, definition, theorem, +... is its {\it absolute name}. The last identifier ({\tt eq} in the +previous example) is its {\it short name} (or sometimes {\it base +name}). Any suffix of the absolute name is a {\em partially qualified +name} (e.g. {\tt Logic.eq} is a partially qualified name for {\tt +Coq.Init.Logic.eq}). Partially qualified names (shortly {\em +qualified name}) are also built from identifiers separated by dots. +They are written {\qualid} in the documentation. + +{\Coq} does not accept two constructions (definition, theorem, ...) +with the same absolute name but different constructions can have the +same short name (or even same partially qualified names as soon as the +full names are different). + +\paragraph{Visibility} + +{\Coq} maintains a {\it name table} mapping qualified names to absolute +names. This table is modified by the commands {\tt Require} (see +\ref{Require}), {\tt Import} and {\tt Export} (see \ref{Import}) and +also each time a new declaration is added to the context. + +An absolute name is called {\it visible} from a given short or +partially qualified name when this name suffices to denote it. This +means that the short or partially qualified name is mapped to the absolute +name in {\Coq} name table. + +It may happen that a visible name is hidden by the short name or a +qualified name of another construction. In this case, the name that +has been hidden must be referred to using one more level of +qualification. Still, to ensure that a construction always remains +accessible, absolute names can never be hidden. + +Examples: +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Check 0. +Definition nat := bool. +Check 0. +Check Datatypes.nat. +Locate nat. +\end{coq_example} + +\Rem There is also a name table for sublibraries, modules and sections. + +\Rem In versions prior to {\Coq} 7.4, lemmas declared with {\tt +Remark} and {\tt Fact} kept in their full name the names of the +sections in which they were defined. Since {\Coq} 7.4, they strictly +behaves as {\tt Theorem} and {\tt Lemma} do. + +\SeeAlso Command {\tt Locate} in Section~\ref{Locate}. + +%% \paragraph{The special case of remarks and facts} +%% +%% In contrast with definitions, lemmas, theorems, axioms and parameters, +%% the absolute name of remarks includes the segment of sections in which +%% it is defined. Concretely, if a remark {\tt R} is defined in +%% subsection {\tt S2} of section {\tt S1} in module {\tt M}, then its +%% absolute name is {\tt M.S1.S2.R}. The same for facts, except that the +%% name of the innermost section is dropped from the full name. Then, if +%% a fact {\tt F} is defined in subsection {\tt S2} of section {\tt S1} +%% in module {\tt M}, then its absolute name is {\tt M.S1.F}. + + +\paragraph{Requiring a file} + +A module compiled in a ``.vo'' file comes with a logical names (e.g. +physical file \verb!theories/Init/Datatypes.vo! in the {\Coq} installation directory is bound to the logical module {\tt Coq.Init.Datatypes}). +When requiring the file, the mapping between physical directories and logical library should be consistent with the mapping used to compile the file (for modules of the standard library, this is automatic -- check it by typing {\tt Print LoadPath}). + +The command {\tt Add Rec LoadPath} is also available from {\tt coqtop} +and {\tt coqc} by using option~\verb=-R=. + +\section{Implicit arguments +\index{Implicit arguments} +\label{Implicit Arguments}} + +An implicit argument of a function is an argument which can be +inferred from the knowledge of the type of other arguments of the +function, or of the type of the surrounding context of the application. +Especially, an implicit argument corresponds to a parameter +dependent in the type of the function. Typical implicit +arguments are the type arguments in polymorphic functions. +More precisely, there are several kinds of implicit arguments. + +\paragraph{Strict Implicit Arguments.} +An implicit argument can be either strict or non strict. An implicit +argument is said {\em strict} if, whatever the other arguments of the +function are, it is still inferable from the type of some other +argument. Technically, an implicit argument is strict if it +corresponds to a parameter which is not applied to a variable which +itself is another parameter of the function (since this parameter +may erase its arguments), not in the body of a {\tt match}, and not +itself applied or matched against patterns (since the original +form of the argument can be lost by reduction). + +For instance, the first argument of +\begin{quote} +\verb|cons: forall A:Set, A -> list A -> list A| +\end{quote} +in module {\tt List.v} is strict because {\tt list} is an inductive +type and {\tt A} will always be inferable from the type {\tt +list A} of the third argument of {\tt cons}. +On the opposite, the second argument of a term of type +\begin{quote} +\verb|forall P:nat->Prop, forall n:nat, P n -> ex nat P| +\end{quote} +is implicit but not strict, since it can only be inferred from the +type {\tt P n} of the the third argument and if {\tt P} is e.g. {\tt +fun \_ => True}, it reduces to an expression where {\tt n} does not +occur any longer. The first argument {\tt P} is implicit but not +strict either because it can only be inferred from {\tt P n} and {\tt +P} is not canonically inferable from an arbitrary {\tt n} and the +normal form of {\tt P n} (consider e.g. that {\tt n} is {\tt 0} and +the third argument has type {\tt True}, then any {\tt P} of the form +{\tt fun n => match n with 0 => True | \_ => \mbox{\em anything} end} would +be a solution of the inference problem. + +\paragraph{Contextual Implicit Arguments.} +An implicit argument can be {\em contextual} or non. An implicit +argument is said {\em contextual} if it can be inferred only from the +knowledge of the type of the context of the current expression. For +instance, the only argument of +\begin{quote} +\verb|nil : forall A:Set, list A| +\end{quote} +is contextual. Similarly, both arguments of a term of type +\begin{quote} +\verb|forall P:nat->Prop, forall n:nat, P n \/ n = 0| +\end{quote} +are contextual (moreover, {\tt n} is strict and {\tt P} is not). + +\subsection{Casual use of implicit arguments} + +In a given expression, if it is clear that some argument of a function +can be inferred from the type of the other arguments, the user can +force the given argument to be guessed by replacing it by ``{\tt \_}''. If +possible, the correct argument will be automatically generated. + +\begin{ErrMsgs} + +\item \errindex{Cannot infer a term for this placeholder} + + {\Coq} was not able to deduce an instantiation of a ``{\tt \_}''. + +\end{ErrMsgs} + +\subsection{Declaration of implicit arguments for a constant +\comindex{Implicit Arguments}} + +In case one wants that some arguments of a given object (constant, +inductive types, constructors, assumptions, local or not) are always +inferred by Coq, one may declare once for all which are the expected +implicit arguments of this object. The syntax is +\begin{quote} +\tt Implicit Arguments {\qualid} [ \nelist{\ident}{} ] +\end{quote} +where the list of {\ident} is the list of parameters to be declared +implicit. After this, implicit arguments can just (and have to) be +skipped in any expression involving an application of {\qualid}. + +\Example +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Inductive list (A:Set) : Set := + | nil : list A + | cons : A -> list A -> list A. +\end{coq_example*} +\begin{coq_example} +Check (cons nat 3 (nil nat)). +Implicit Arguments cons [A]. +Implicit Arguments nil [A]. +Check (cons 3 nil). +\end{coq_example} + +\Rem To know which are the implicit arguments of an object, use command +{\tt Print Implicit} (see \ref{PrintImplicit}). + +\Rem If the list of arguments is empty, the command removes the +implicit arguments of {\qualid}. + +\subsection{Automatic declaration of implicit arguments for a constant} + +{\Coq} can also automatically detect what are the implicit arguments +of a defined object. The command is just +\begin{quote} +\tt Implicit Arguments {\qualid}. +\end{quote} +The auto-detection is governed by options telling if strict and +contextual implicit arguments must be considered or not (see +Sections~\ref{SetStrictImplicit} and~\ref{SetContextualImplicit}). + +\Example +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Inductive list (A:Set) : Set := + | nil : list A + | cons : A -> list A -> list A. +\end{coq_example*} +\begin{coq_example} +Implicit Arguments cons. +Print Implicit cons. +Implicit Arguments nil. +Print Implicit nil. +Set Contextual Implicit. +Implicit Arguments nil. +Print Implicit nil. +\end{coq_example} + +The computation of implicit arguments takes account of the +unfolding of constants. For instance, the variable {\tt p} below has +type {\tt (Transitivity R)} which is reducible to {\tt forall x,y:U, R x +y -> forall z:U, R y z -> R x z}. As the variables {\tt x}, {\tt y} and +{\tt z} appear strictly in body of the type, they are implicit. + +\begin{coq_example*} +Variable X : Type. +Definition Relation := X -> X -> Prop. +Definition Transitivity (R:Relation) := + forall x y:X, R x y -> forall z:X, R y z -> R x z. +Variables (R : Relation) (p : Transitivity R). +Implicit Arguments p. +\end{coq_example*} +\begin{coq_example} +Print p. +Print Implicit p. +\end{coq_example} +\begin{coq_example*} +Variables (a b c : X) (r1 : R a b) (r2 : R b c). +\end{coq_example*} +\begin{coq_example} +Check (p r1 r2). +\end{coq_example} + +\subsection{Mode for automatic declaration of implicit arguments +\label{Auto-implicit} +\comindex{Set Implicit Arguments} +\comindex{Unset Implicit Arguments}} + +In case one wants to systematically declare implicit the arguments +detectable as such, one may switch to the automatic declaration of +implicit arguments mode by using the command +\begin{quote} +\tt Set Implicit Arguments. +\end{quote} +Conversely, one may unset the mode by using {\tt Unset Implicit +Arguments}. The mode is off by default. Auto-detection of implicit +arguments is governed by options controlling whether strict and +contextual implicit arguments have to be considered or not. + +\subsection{Controlling strict implicit arguments +\comindex{Set Strict Implicit} +\comindex{Unset Strict Implicit} +\label{SetStrictImplicit}} + +By default, {\Coq} automatically set implicit only the strict implicit +arguments. To relax this constraint, use command +\begin{quote} +\tt Unset Strict Implicit. +\end{quote} +Conversely, use command {\tt Set Strict Implicit} to +restore the strict implicit mode. + +\Rem In versions of {\Coq} prior to version 8.0, the default was to +declare the strict implicit arguments as implicit. + +\subsection{Controlling contextual implicit arguments +\comindex{Set Contextual Implicit} +\comindex{Unset Contextual Implicit} +\label{SetContextualImplicit}} + +By default, {\Coq} does not automatically set implicit the contextual +implicit arguments. To tell {\Coq} to infer also contextual implicit +argument, use command +\begin{quote} +\tt Set Contextual Implicit. +\end{quote} +Conversely, use command {\tt Unset Contextual Implicit} to +unset the contextual implicit mode. + +\subsection{Explicit Applications +\index{Explicitation of implicit arguments} +\label{Implicits-explicitation} +\index{qualid@{\qualid}}} + +In presence of non strict or contextual argument, or in presence of +partial applications, the synthesis of implicit arguments may fail, so +one may have to give explicitly certain implicit arguments of an +application. The syntax for this is {\tt (\ident:=\term)} where {\ident} +is the name of the implicit argument and {\term} is its corresponding +explicit term. Alternatively, one can locally deactivate the hidding of +implicit arguments of a function by using the notation +{\tt @{\qualid}~{\term}$_1$..{\term}$_n$}. This syntax extension is +given Figure~\ref{fig:explicitations}. +\begin{figure} +\begin{centerframe} +\begin{tabular}{lcl} +{\term} & ++= & @ {\qualid} \nelist{\term}{}\\ +& $|$ & @ {\qualid}\\ +& $|$ & {\qualid} \nelist{\textrm{\textsl{argument}}}{}\\ +\\ +{\textrm{\textsl{argument}}} & ::= & {\term} \\ +& $|$ & {\tt ({\ident}:={\term})}\\ +\end{tabular} +\end{centerframe} +\caption{Syntax for explicitations of implicit arguments} +\label{fig:explicitations} +\end{figure} + +\noindent {\bf Example (continued): } +\begin{coq_example} +Check (p r1 (z:=c)). +Check (p (x:=a) (y:=b) r1 (z:=c) r2). +\end{coq_example} + +\subsection{Displaying what the implicit arguments are +\comindex{Print Implicit} +\label{PrintImplicit}} + +To display the implicit arguments associated to an object use command +\begin{quote} +\tt Print Implicit {\qualid}. +\end{quote} + +\subsection{Explicitation of implicit arguments for pretty-printing +\comindex{Set Printing Implicit} +\comindex{Unset Printing Implicit}} + +By default the basic pretty-printing rules hide the inferable implicit +arguments of an application. To force printing all implicit arguments, +use command +\begin{quote} +{\tt Set Printing Implicit.} +\end{quote} +Conversely, to restore the hidding of implicit arguments, use command +\begin{quote} +{\tt Unset Printing Implicit.} +\end{quote} + +\SeeAlso {\tt Set Printing All} in section \ref{SetPrintingAll}. + +\subsection{Interaction with subtyping} + +When an implicit argument can be inferred from the type of more than +one of the other arguments, then only the type of the first of these +arguments is taken into account, and not an upper type of all of +them. As a consequence, the inference of the implicit argument of +``='' fails in + +\begin{coq_example*} +Check nat = Prop. +\end{coq_example*} + +but succeeds in + +\begin{coq_example*} +Check Prop = nat. +\end{coq_example*} + +\subsection{Canonical structures +\comindex{Canonical Structure}} + +A canonical structure is an instance of a record/structure type that +can be used to solve equations involving implicit arguments. Assume +that {\qualid} denotes an object $(Build\_struc~ c_1~ \ldots~ c_n)$ in the +structure {\em struct} of which the fields are $x_1$, ..., +$x_n$. Assume that {\qualid} is declared as a canonical structure +using the command +\begin{quote} +{\tt Canonical Structure {\qualid}.} +\end{quote} +Then, each time an equation of the form $(x_i~ +\_)=_{\beta\delta\iota\zeta}c_i$ has to be solved during the +type-checking process, {\qualid} is used as a solution. Otherwise +said, {\qualid} is canonically used to extend the field $c_i$ into a +complete structure built on $c_i$. + +Canonical structures are particularly useful when mixed with +coercions and strict implicit arguments. Here is an example. +\begin{coq_example*} +Require Import Relations. +Require Import EqNat. +Set Implicit Arguments. +Unset Strict Implicit. +Structure Setoid : Type := + {Carrier :> Set; + Equal : relation Carrier; + Prf_equiv : equivalence Carrier Equal}. +Definition is_law (A B:Setoid) (f:A -> B) := + forall x y:A, Equal x y -> Equal (f x) (f y). +Axiom eq_nat_equiv : equivalence nat eq_nat. +Definition nat_setoid : Setoid := Build_Setoid eq_nat_equiv. +Canonical Structure nat_setoid. +\end{coq_example*} + +Thanks to \texttt{nat\_setoid} declared as canonical, the implicit +arguments {\tt A} and {\tt B} can be synthesised in the next statement. +\begin{coq_example} +Lemma is_law_S : is_law S. +\end{coq_example} + +\Rem If a same field occurs in several canonical structure, then +only the structure declared first as canonical is considered. + +\begin{Variants} +\item {\tt Canonical Structure {\ident} := {\term} : {\type}.}\\ + {\tt Canonical Structure {\ident} := {\term}.}\\ + {\tt Canonical Structure {\ident} : {\type} := {\term}.} + +These are equivalent to a regular definition of {\ident} followed by +the declaration + +{\tt Canonical Structure {\ident}}. +\end{Variants} + +\SeeAlso more examples in user contribution \texttt{category} +(\texttt{Rocq/ALGEBRA}). + +\subsection{Implicit types of variables} + +It is possible to bind variable names to a given type (e.g. in a +development using arithmetic, it may be convenient to bind the names +{\tt n} or {\tt m} to the type {\tt nat} of natural numbers). The +command for that is +\begin{quote} +\tt Implicit Types \nelist{\ident}{} : {\type} +\end{quote} +The effect of the command is to automatically set the type of bound +variables starting with {\ident} (either {\ident} itself or +{\ident} followed by one or more single quotes, underscore or digits) +to be {\type} (unless the bound variable is already declared with an +explicit type in which case, this latter type is considered). + +\Example +\begin{coq_example} +Require Import List. +Implicit Types m n : nat. +Lemma cons_inj_nat : forall m n l, n :: l = m :: l -> n = m. +intros m n. +Lemma cons_inj_bool : forall (m n:bool) l, n :: l = m :: l -> n = m. +\end{coq_example} + +\begin{Variants} +\item {\tt Implicit Type {\ident} : {\type}}\\ +This is useful for declaring the implicit type of a single variable. +\end{Variants} + +\section{Coercions +\label{Coercions} +\index{Coercions}} + +Coercions can be used to implicitly inject terms from one {\em class} in +which they reside into another one. A {\em class} is either a sort +(denoted by the keyword {\tt Sortclass}), a product type (denoted by the +keyword {\tt Funclass}), or a type constructor (denoted by its name), +e.g. an inductive type or any constant with a type of the form +\texttt{forall} $(x_1:A_1) .. (x_n:A_n),~s$ where $s$ is a sort. + +Then the user is able to apply an +object that is not a function, but can be coerced to a function, and +more generally to consider that a term of type A is of type B provided +that there is a declared coercion between A and B. The main command is +\comindex{Coercion} +\begin{quote} +\tt Coercion {\qualid} : {\class$_1$} >-> {\class$_2$}. +\end{quote} +which declares the construction denoted by {\qualid} as a +coercion between {\class$_1$} and {\class$_2$}. + +More details and examples, and a description of the commands related +to coercions are provided in chapter \ref{Coercions-full}. + +\section{Printing constructions in full} +\label{SetPrintingAll} +\comindex{Set Printing All} +\comindex{Unset Printing All} + +Coercions, implicit arguments, the type of pattern-matching, but also +notations (see chapter \ref{Addoc-syntax}) can obfuscate the behavior +of some tactics (typically the tactics applying to occurrences of +subterms are sensitive to the implicit arguments). The command +\begin{quote} +{\tt Set Printing All.} +\end{quote} +deactivates all high-level printing features such as coercions, +implicit arguments, returned type of pattern-matching, notations and +various syntactic sugar for pattern-matching or record projections. +Otherwise said, {\tt Set Printing All} includes the effects +of the commands {\tt Set Printing Implicit}, {\tt Set Printing +Coercions}, {\tt Set Printing Synth}, {\tt Unset Printing Projections} +and {\tt Unset Printing Notations}. To reactivate the high-level +printing features, use the command +\begin{quote} +{\tt Unset Printing All.} +\end{quote} + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-gal.tex b/doc/refman/RefMan-gal.tex new file mode 100644 index 0000000000..543b3fb829 --- /dev/null +++ b/doc/refman/RefMan-gal.tex @@ -0,0 +1,1451 @@ +\chapter{The \gallina{} specification language +\label{Gallina}\index{Gallina}} + +This chapter describes \gallina, the specification language of {\Coq}. +It allows to develop mathematical theories and to prove specifications +of programs. The theories are built from axioms, hypotheses, +parameters, lemmas, theorems and definitions of constants, functions, +predicates and sets. The syntax of logical objects involved in +theories is described in section \ref{term}. The language of +commands, called {\em The Vernacular} is described in section +\ref{Vernacular}. + +In {\Coq}, logical objects are typed to ensure their logical +correctness. The rules implemented by the typing algorithm are described in +chapter \ref{Cic}. + +\subsection*{About the grammars in the manual +\label{BNF-syntax}\index{BNF metasyntax}} + +Grammars are presented in Backus-Naur form (BNF). Terminal symbols are +set in {\tt typewriter font}. In addition, there are special +notations for regular expressions. + +An expression enclosed in square brackets \zeroone{\ldots} means at +most one occurrence of this expression (this corresponds to an +optional component). + +The notation ``\nelist{\entry}{sep}'' stands for a non empty +sequence of expressions parsed by {\entry} and +separated by the literal ``{\tt sep}''\footnote{This is similar to the +expression ``{\entry} $\{$ {\tt sep} {\entry} $\}$'' in +standard BNF, or ``{\entry} $($ {\tt sep} {\entry} $)$*'' in +the syntax of regular expressions.}. + +Similarly, the notation ``\nelist{\entry}{}'' stands for a non +empty sequence of expressions parsed by the ``{\entry}'' entry, +without any separator between. + +At the end, the notation ``\sequence{\entry}{\tt sep}'' stands for a +possibly empty sequence of expressions parsed by the ``{\entry}'' entry, +separated by the literal ``{\tt sep}''. + +\section{Lexical conventions +\label{lexical}\index{Lexical conventions}} + +\paragraph{Blanks} +Space, newline and horizontal tabulation are considered as blanks. +Blanks are ignored but they separate tokens. + +\paragraph{Comments} + +Comments in {\Coq} are enclosed between {\tt (*} and {\tt + *)}\index{Comments}, and can be nested. They can contain any +character. However, string literals must be correctly closed. Comments +are treated as blanks. + +\paragraph{Identifiers and access identifiers} + +Identifiers, written {\ident}, are sequences of letters, digits, +\verb!_! and \verb!'!, that do not start with a digit or \verb!'!. +That is, they are recognized by the following lexical class: + +\index{ident@\ident} +\begin{center} +\begin{tabular}{rcl} +{\firstletter} & ::= & {\tt a..z} $\mid$ {\tt A..Z} $\mid$ {\tt \_} +% $\mid$ {\tt unicode-letter} +\\ +{\subsequentletter} & ::= & {\tt a..z} $\mid$ {\tt A..Z} $\mid$ {\tt 0..9} +$\mid$ {\tt \_} % $\mid$ {\tt \$} +$\mid$ {\tt '} \\ +{\ident} & ::= & {\firstletter} \sequencewithoutblank{\subsequentletter}{} +\end{tabular} +\end{center} +All characters are meaningful. In particular, identifiers are case-sensitive. +Access identifiers, written {\accessident}, are identifiers prefixed +by \verb!.! (dot) without blank. They are used in the syntax of qualified +identifiers. + +\paragraph{Natural numbers and integers} +Numerals are sequences of digits. Integers are numerals optionally preceded by a minus sign. + +\index{num@{\num}} +\index{integer@{\integer}} +\begin{center} +\begin{tabular}{r@{\quad::=\quad}l} +{\digit} & {\tt 0..9} \\ +{\num} & \nelistwithoutblank{\digit}{} \\ +{\integer} & \zeroone{\tt -}{\num} \\ +\end{tabular} +\end{center} + +\paragraph{Strings} +\label{strings} +\index{string@{\qstring}} +Strings are delimited by \verb!"! (double quote), and enclose a +sequence of any characters different from \verb!"! or the sequence +\verb!""! to denote the double quote character. In grammars, the +entry for quoted strings is {\qstring}. + +\paragraph{Keywords} +The following identifiers are reserved keywords, and cannot be +employed otherwise: +\begin{center} +\begin{tabular}{llllll} +\verb!_! & +\verb!as! & +\verb!at! & +\verb!cofix! & +\verb!else! & +\verb!end! \\ +% +\verb!exists! & +\verb!exists2! & +\verb!fix! & +\verb!for! & +\verb!forall! & +\verb!fun! \\ +% +\verb!if! & +\verb!IF! & +\verb!in! & +\verb!let! & +\verb!match! & +\verb!mod! \\ +% +\verb!Prop! & +\verb!return! & +\verb!Set! & +\verb!then! & +\verb!Type! & +\verb!using! \\ +% +\verb!where! & +\verb!with! & +\end{tabular} +\end{center} + + +\paragraph{Special tokens} +The following sequences of characters are special tokens: +\begin{center} +\begin{tabular}{lllllll} +\verb/!/ & +\verb!%! & +\verb!&! & +\verb!&&! & +\verb!(! & +\verb!()! & +\verb!)! \\ +% +\verb!*! & +\verb!+! & +\verb!++! & +\verb!,! & +\verb!-! & +\verb!->! & +\verb!.! \\ +% +\verb!.(! & +\verb!..! & +\verb!/! & +\verb!/\! & +\verb!:! & +\verb!::! & +\verb!:! & +\verb!;! & +\verb!! & +\verb!<:! \\ +% +\verb!<=! & +\verb!<>! & +\verb!=! & +\verb!=>! & +\verb!=_D! & +\verb!>! & +\verb!>->! \\ +% +\verb!>=! & +\verb!?! & +\verb!?=! & +\verb!@! & +\verb![! & +\verb!\/! & +\verb!]! \\ +% +\verb!^! & +\verb!{! & +\verb!|! & +\verb!|-! & +\verb!||! & +\verb!}! & +\verb!~! \\ +\end{tabular} +\end{center} + +Lexical ambiguities are resolved according to the ``longest match'' +rule: when a sequence of non alphanumerical characters can be decomposed +into several different ways, then the first token is the longest +possible one (among all tokens defined at this moment), and so on. + +\section{Terms \label{term}\index{Terms}} + +\subsection{Syntax of terms} + +Figures \ref{term-syntax} and \ref{term-syntax-aux} describe the basic +set of terms which form the {\em Calculus of Inductive Constructions} +(also called \CIC). The formal presentation of {\CIC} is given in +chapter \ref{Cic}. Extensions of this syntax are given in chapter +\ref{Gallina-extension}. How to customize the syntax is described in +chapter \ref{Addoc-syntax}. + +\begin{figure}[htbp] +\begin{centerframe} +\begin{tabular}{lcl@{\qquad}r} +{\term} & ::= & + {\tt forall} {\binderlist} {\tt ,} {\term} &(\ref{products})\\ + & $|$ & {\tt fun} {\binderlist} {\tt =>} {\term} &(\ref{abstractions})\\ + & $|$ & {\tt fix} {\fixpointbodies} &(\ref{fixpoints})\\ + & $|$ & {\tt cofix} {\cofixpointbodies} &(\ref{fixpoints})\\ + & $|$ & {\tt let} {\idparams} {\tt :=} {\term} + {\tt in} {\term} &(\ref{let-in})\\ + & $|$ & {\tt let fix} {\fixpointbody} {\tt in} {\term} &(\ref{fixpoints})\\ + & $|$ & {\tt let cofix} {\cofixpointbody} + {\tt in} {\term} &(\ref{fixpoints})\\ + & $|$ & {\tt let} {\tt (} \sequence{\name}{,} {\tt )} \zeroone{\ifitem} + {\tt :=} {\term} + {\tt in} {\term} &(\ref{caseanalysis}, \ref{Mult-match})\\ + & $|$ & {\tt if} {\term} \zeroone{\ifitem} {\tt then} {\term} + {\tt else} {\term} &(\ref{caseanalysis}, \ref{Mult-match})\\ + & $|$ & {\term} {\tt :} {\term} &(\ref{typecast})\\ + & $|$ & {\term} {\tt ->} {\term} &(\ref{products})\\ + & $|$ & {\term} \nelist{\termarg}{}&(\ref{applications})\\ + & $|$ & {\tt @} {\qualid} \sequence{\term}{} + &(\ref{Implicits-explicitation})\\ + & $|$ & {\term} {\tt \%} {\ident} &(\ref{scopechange})\\ + & $|$ & {\tt match} \nelist{\caseitem}{\tt ,} + \zeroone{\returntype} {\tt with} &\\ + && ~~~\zeroone{\zeroone{\tt |} \nelist{\eqn}{|}} {\tt end} + &(\ref{caseanalysis})\\ + & $|$ & {\qualid} &(\ref{qualid})\\ + & $|$ & {\sort} &(\ref{Gallina-sorts})\\ + & $|$ & {\num} &(\ref{numerals})\\ + & $|$ & {\_} &(\ref{hole})\\ + & & &\\ +{\termarg} & ::= & {\term} &\\ + & $|$ & {\tt (} {\ident} {\tt :=} {\term} {\tt )} + &(\ref{Implicits-explicitation})\\ +%% & $|$ & {\tt (} {\num} {\tt :=} {\term} {\tt )} +%% &(\ref{Implicits-explicitation})\\ +&&&\\ +{\binderlist} & ::= & \nelist{\name}{} {\typecstr} & \ref{Binders} \\ + & $|$ & {\binder} \nelist{\binderlet}{} &\\ +&&&\\ +{\binder} & ::= & {\name} & \ref{Binders} \\ + & $|$ & {\tt (} \nelist{\name}{} {\tt :} {\term} {\tt )} &\\ +&&&\\ +{\binderlet} & ::= & {\binder} & \ref{Binders} \\ + & $|$ & {\tt (} {\name} {\typecstr} {\tt :=} {\term} {\tt )} &\\ +& & &\\ +{\name} & ::= & {\ident} &\\ + & $|$ & {\tt \_} &\\ +&&&\\ +{\qualid} & ::= & {\ident} & \\ + & $|$ & {\qualid} {\accessident} &\\ + & & &\\ +{\sort} & ::= & {\tt Prop} ~$|$~ {\tt Set} ~$|$~ {\tt Type} & +\end{tabular} +\end{centerframe} +\caption{Syntax of terms} +\label{term-syntax} +\index{term@{\term}} +\index{sort@{\sort}} +\end{figure} + + + +\begin{figure}[htb] +\begin{centerframe} +\begin{tabular}{lcl} +{\idparams} & ::= & {\ident} \sequence{\binderlet}{} {\typecstr} \\ +&&\\ +{\fixpointbodies} & ::= & + {\fixpointbody} \\ + & $|$ & {\fixpointbody} {\tt with} \nelist{\fixpointbody}{{\tt with}} + {\tt for} {\ident} \\ +{\cofixpointbodies} & ::= & + {\cofixpointbody} \\ + & $|$ & {\cofixpointbody} {\tt with} \nelist{\cofixpointbody}{{\tt with}} + {\tt for} {\ident} \\ +&&\\ +{\fixpointbody} & ::= & + {\ident} \nelist{\binderlet}{} \zeroone{\annotation} {\typecstr} + {\tt :=} {\term} \\ +{\cofixpointbody} & ::= & {\idparams} {\tt :=} {\term} \\ + & &\\ +{\annotation} & ::= & {\tt \{ struct} {\ident} {\tt \}} \\ +&&\\ +{\caseitem} & ::= & {\term} \zeroone{{\tt as} \name} + \zeroone{{\tt in} \term} \\ +&&\\ +{\ifitem} & ::= & \zeroone{{\tt as} {\name}} {\returntype} \\ +&&\\ +{\returntype} & ::= & {\tt return} {\term} \\ +&&\\ +{\eqn} & ::= & \nelist{\pattern}{\tt ,} {\tt =>} {\term}\\ +&&\\ +{\pattern} & ::= & {\qualid} \nelist{\pattern}{} \\ + & $|$ & {\pattern} {\tt as} {\ident} \\ + & $|$ & {\pattern} {\tt \%} {\ident} \\ + & $|$ & {\qualid} \\ + & $|$ & {\tt \_} \\ + & $|$ & {\num} \\ + & $|$ & {\tt (} \nelist{\pattern}{,} {\tt )} +\end{tabular} +\end{centerframe} +\caption{Syntax of terms (continued)} +\label{term-syntax-aux} +\end{figure} + + +%%%%%%% + +\subsection{Types} + +{\Coq} terms are typed. {\Coq} types are recognized by the same +syntactic class as {\term}. We denote by {\type} the semantic subclass +of types inside the syntactic class {\term}. +\index{type@{\type}} + + +\subsection{Qualified identifiers and simple identifiers +\label{qualid} +\label{ident}} + +{\em Qualified identifiers} ({\qualid}) denote {\em global constants} +(definitions, lemmas, theorems, remarks or facts), {\em global +variables} (parameters or axioms), {\em inductive +types} or {\em constructors of inductive types}. +{\em Simple identifiers} (or shortly {\ident}) are a +syntactic subset of qualified identifiers. Identifiers may also +denote local {\em variables}, what qualified identifiers do not. + +\subsection{Numerals +\label{numerals}} + +Numerals have no definite semantics in the calculus. They are mere +notations that can be bound to objects through the notation mechanism +(see chapter~\ref{Addoc-syntax} for details). Initially, numerals are +bound to Peano's representation of natural numbers +(see~\ref{libnats}). + +Note: negative integers are not at the same level as {\num}, for this +would make precedence unnatural. + +\subsection{Sorts +\index{Sorts} +\index{Type@{\Type}} +\index{Set@{\Set}} +\index{Prop@{\Prop}} +\index{Sorts} +\label{Gallina-sorts}} + +There are three sorts \Set, \Prop\ and \Type. +\begin{itemize} +\item \Prop\ is the universe of {\em logical propositions}. +The logical propositions themselves are typing the proofs. +We denote propositions by {\form}. This constitutes a semantic +subclass of the syntactic class {\term}. +\index{form@{\form}} +\item \Set\ is is the universe of {\em program +types} or {\em specifications}. +The specifications themselves are typing the programs. +We denote specifications by {\specif}. This constitutes a semantic +subclass of the syntactic class {\term}. +\index{specif@{\specif}} +\item {\Type} is the type of {\Set} and {\Prop} +\end{itemize} +\noindent More on sorts can be found in section \ref{Sorts}. + +\subsection{Binders +\label{Binders} +\index{binders}} + +Various constructions introduce variables which scope is some of its +sub-expressions. There is a uniform syntax for this. A binder may be +an (unqualified) identifier: the name to use to refer to this +variable. If the variable is not to be used, its name can be {\tt +\_}. When its type cannot be synthesized by the system, it can be +specified with notation {\tt (}\,{\ident}\,{\tt :}\,{\type}\,{\tt +)}. There is a notation for several variables sharing the same type: +{\tt (}\,{\ident$_1$}\ldots{\ident$_n$}\,{\tt :}\,{\type}\,{\tt )}. + +Some constructions allow ``let-binders'', that is either a binder as +defined above, or a variable with a value. The notation is {\tt +(}\,{\ident}\,{\tt :=}\,{\term}\,{\tt )}. Only one variable can be +introduced at the same time. It is also possible to give the type of +the variable before the symbol {\tt :=}. + +The last kind of binders is the ``binder list''. It is either a list +of let-binders (the first one not being a variable with value), or +{\ident$_1$}\ldots{\ident$_n$}\,{\tt :}\,{\type} if all variables +share the same type. + +{\Coq} terms are typed. {\Coq} types are recognized by the same +syntactic class as {\term}. We denote by {\type} the semantic subclass +of types inside the syntactic class {\term}. +\index{type@{\type}} + +\subsection{Abstractions +\label{abstractions} +\index{abstractions}} + +The expression ``{\tt fun} {\ident} {\tt :} \type {\tt =>}~{\term}'' +denotes the {\em abstraction} of the variable {\ident} of type +{\type}, over the term {\term}. Put in another way, it is function of +formal parameter {\ident} of type {\type} returning {\term}. + +Keyword {\tt fun} is followed by a ``binder list'', so any of the +binders of Section~\ref{Binders} apply. Internally, abstractions are +only over one variable. Multiple variable binders are an iteration of +the single variable abstraction: notation {\tt +fun}~{\ident$_{1}$}~{\ldots}~{\ident$_{n}$}~{\tt :}~\type~{\tt +=>}~{\term} stands for {\tt fun}~{\ident$_{1}$}~{\tt :}~\type~{\tt +=>}~{\ldots}~{\tt fun}~{\ident$_{n}$}~{\tt :}~\type~{\tt =>}~{\term}. +Variables with a value expand to a local definition (see +Section~\ref{let-in}). + +\subsection{Products +\label{products} +\index{products}} + +The expression ``{\tt forall}~{\ident}~{\tt :}~\type~{\tt ,}~{\term}'' +denotes the {\em product} of the variable {\ident} of type {\type}, +over the term {\term}. As for abstractions, {\tt forall} is followed +by a binder list, and it is represented by an iteration of single +variable products. + +Non dependent product types have a special notation ``$A$ {\tt ->} +$B$'' stands for ``{\tt forall \_:}$A${\tt ,}~$B$''. This is to stress +on the fact that non dependent product types are usual functional types. + +\subsection{Applications +\label{applications} +\index{applications}} + +The expression \term$_0$ \term$_1$ denotes the application of + term \term$_0$ to \term$_1$. + +The expression {\tt }\term$_0$ \term$_1$ ... \term$_n${\tt} +denotes the application of the term \term$_0$ to the arguments +\term$_1$ ... then \term$_n$. It is equivalent to {\tt } {\ldots} +{\tt (} {\term$_0$} {\term$_1$} {\tt )} {\ldots} {\term$_n$} {\tt }: +associativity is to the left. + +When using implicit arguments mechanism, implicit positions can be +forced a value with notation {\tt (}\,{\ident}\,{\tt +:=}\,{\term}\,{\tt )} or {\tt (}\,{\num}\,{\tt +:=}\,{\term}\,{\tt )}. See Section~\ref{Implicits-explicitation} for +details. + +\subsection{Type cast +\label{typecast} +\index{Cast}} + +The expression ``{\term}~{\tt :}~{\type}'' is a type cast +expression. It enforces the type of {\term} to be {\type}. + +\subsection{Inferable subterms +\label{hole} +\index{\_}} + +Since there are redundancies, a term can be type-checked without +giving it in totality. Subterms that are left to guess by the +type-checker are replaced by ``\_''. + + +\subsection{Local definitions (let-in) +\label{let-in} +\index{Local definitions} +\index{let-in}} + + +{\tt let}~{\ident}~{\tt :=}~{\term$_1$}~{\tt in}~{\term$_2$} denotes +the local binding of \term$_1$ to the variable $\ident$ in +\term$_2$. + +There is a syntactic sugar for local definition of functions: {\tt +let} {\ident} {\binder$_1$} \ldots {\binder$_n$} {\tt :=} {\term$_1$} +{\tt in} {\term$_2$} stands for {\tt let} {\ident} {\tt := fun} +{\binder$_1$} \ldots {\binder$_n$} {\tt in} {\term$_2$}. + + +\subsection{Definition by case analysis +\label{caseanalysis} +\index{match@{\tt match\ldots with\ldots end}}} + + +This paragraph only shows simple variants of case analysis. See +Section~\ref{Mult-match} and Chapter~\ref{Mult-match-full} for +explanations of the general form. + +Objects of inductive types can be destructurated by a case-analysis +construction, also called pattern-matching in functional languages. In +its simple form, a case analysis expression is used to analyze the +structure of an inductive objects (upon which constructor it is +built). + +The expression {\tt match} {\term$_0$} {\returntype} {\tt with} +{\pattern$_1$} {\tt =>} {\term$_1$} {\tt $|$} {\ldots} {\tt $|$} +{\pattern$_n$} {\tt =>} {\term$_n$} {\tt end}, denotes a {\em +pattern-matching} over the term {\term$_0$} (expected to be of an +inductive type $I$). {\term$_1$}\ldots{\term$_n$} are called branches. In +a simple pattern \qualid~\nelist{\ident}{}, the qualified identifier +{\qualid} is intended to +be a constructor. There should be a branch for every constructor of +$I$. + +The {\returntype} is used to compute the resulting type of the whole +{\tt match} expression and the type of the branches. Most of the time, +when this type is the same as the types of all the {\term$_i$}, the +annotation is not needed\footnote{except if no equation is given, to +match the term in an empty type, e.g. the type \texttt{False}}. This +annotation has to be given when the resulting type of the whole {\tt +match} depends on the actual {\term$_0$} matched. + +There are specific notations for case analysis on types with one or +two constructors: {\tt if / then / else} and +{\tt let (}\ldots{\tt ) :=} \ldots {\tt in}\ldots. \SeeAlso +section~\ref{Mult-match} for details and examples. + +\SeeAlso Section~\ref{Mult-match} for details and examples. + +\subsection{Recursive functions +\label{fixpoints} +\index{fix@{fix \ident$_i$\{\dots\}}}} + +Expression ``{\tt fix} \ident$_1$ \binder$_1$ {\tt :} {\type$_1$} +\texttt{:=} \term$_1$ {\tt with} {\ldots} {\tt with} \ident$_n$ +\binder$_n$~{\tt :} {\type$_n$} \texttt{:=} \term$_n$ {\tt for} +{\ident$_i$}'' denotes the $i$th component of a block of functions +defined by mutual well-founded recursion. It is the local counterpart +of the {\tt Fixpoint} command. See Section~\ref{Fixpoint} for more +details. When $n=1$, the {\tt for}~{\ident$_i$} is omitted. + +The expression ``{\tt cofix} \ident$_1$~\binder$_1$ {\tt :} +{\type$_1$} {\tt with} {\ldots} {\tt with} \ident$_n$ \binder$_n$ {\tt +:} {\type$_n$}~{\tt for} {\ident$_i$}'' denotes the $i$th component of +a block of terms defined by a mutual guarded co-recursion. It is the +local counterpart of the {\tt CoFixpoint} command. See +Section~\ref{CoFixpoint} for more details. When $n=1$, the {\tt +for}~{\ident$_i$} is omitted. + +The association of a single fixpoint and a local +definition have a special syntax: ``{\tt let fix}~$f$~{\ldots}~{\tt + :=}~{\ldots}~{\tt in}~{\ldots}'' stands for ``{\tt let}~$f$~{\tt := + fix}~$f$~\ldots~{\tt :=}~{\ldots}~{\tt in}~{\ldots}''. The same + applies for co-fixpoints. + + +\section{The Vernacular +\label{Vernacular}} + +\begin{figure}[tbp] +\begin{centerframe} +\begin{tabular}{lcl} +{\sentence} & ::= & {\declaration} \\ + & $|$ & {\definition} \\ + & $|$ & {\inductive} \\ + & $|$ & {\fixpoint} \\ + & $|$ & {\statement} \zeroone{\proof} \\ +&&\\ +%% Declarations +{\declaration} & ::= & {\declarationkeyword} {\assums} {\tt .} \\ +&&\\ +{\declarationkeyword} & ::= & {\tt Axiom} $|$ {\tt Conjecture} \\ + & $|$ & {\tt Parameter} $|$ {\tt Parameters} \\ + & $|$ & {\tt Variable} $|$ {\tt Variables} \\ + & $|$ & {\tt Hypothesis} $|$ {\tt Hypotheses}\\ +&&\\ +{\assums} & ::= & \nelist{\ident}{} {\tt :} {\term} \\ + & $|$ & \nelist{\binder}{} \\ +&&\\ +%% Definitions +{\definition} & ::= & + {\tt Definition} {\idparams} {\tt :=} {\term} {\tt .} \\ + & $|$ & {\tt Let} {\idparams} {\tt :=} {\term} {\tt .} \\ +&&\\ +%% Inductives +{\inductive} & ::= & + {\tt Inductive} \nelist{\inductivebody}{with} {\tt .} \\ + & $|$ & {\tt CoInductive} \nelist{\inductivebody}{with} {\tt .} \\ + & & \\ +{\inductivebody} & ::= & + {\ident} \sequence{\binderlet}{} {\tt :} {\term} {\tt :=} \\ + && ~~~\zeroone{\zeroone{\tt |} \nelist{\idparams}{|}} \\ + & & \\ %% TODO: where ... +%% Fixpoints +{\fixpoint} & ::= & {\tt Fixpoint} \nelist{\fixpointbody}{with} {\tt .} \\ + & $|$ & {\tt CoFixpoint} \nelist{\cofixpointbody}{with} {\tt .} \\ +&&\\ +%% Lemmas & proofs +{\statement} & ::= & + {\statkwd} {\ident} \sequence{\binderlet}{} {\tt :} {\term} {\tt .} \\ +&&\\ + {\statkwd} & ::= & {\tt Theorem} $|$ {\tt Lemma} $|$ {\tt Definition} \\ +&&\\ +{\proof} & ::= & {\tt Proof} {\tt .} {\dots} {\tt Qed} {\tt .}\\ + & $|$ & {\tt Proof} {\tt .} {\dots} {\tt Defined} {\tt .}\\ + & $|$ & {\tt Proof} {\tt .} {\dots} {\tt Admitted} {\tt .} +\end{tabular} +\end{centerframe} +\caption{Syntax of sentences} +\label{sentences-syntax} +\end{figure} + +Figure \ref{sentences-syntax} describes {\em The Vernacular} which is the +language of commands of \gallina. A sentence of the vernacular +language, like in many natural languages, begins with a capital letter +and ends with a dot. + +The different kinds of command are described hereafter. They all suppose +that the terms occurring in the sentences are well-typed. + +%% +%% Axioms and Parameters +%% +\subsection{Declarations +\index{Declarations} +\label{Declarations}} + +The declaration mechanism allows the user to specify his own basic +objects. Declared objects play the role of axioms or parameters in +mathematics. A declared object is an {\ident} associated to a \term. A +declaration is accepted by {\Coq} if and only if this {\term} is a +correct type in the current context of the declaration and \ident\ was +not previously defined in the same module. This {\term} is considered +to be the type, or specification, of the \ident. + +\subsubsection{{\tt Axiom {\ident} :{\term} .} +\comindex{Axiom} +\comindex{Parameter}\comindex{Parameters} +\comindex{Conjecture} +\label{Axiom}} + +This command links {\term} to the name {\ident} as its specification +in the global context. The fact asserted by {\term} is thus assumed as +a postulate. + +\begin{ErrMsgs} +\item \errindex{{\ident} already exists} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Parameter {\ident} :{\term}.} \\ + Is equivalent to {\tt Axiom {\ident} : {\term}} + +\item {\tt Parameter {\ident$_1$}\ldots{\ident$_n$} {\tt :}{\term}.}\\ + Adds $n$ parameters with specification {\term} + +\item + {\tt Parameter\,% +(\,{\ident$_{1,1}$}\ldots{\ident$_{1,k_1}$}\,{\tt :}\,{\term$_1$} {\tt )}\,% +\ldots\,{\tt (}\,{\ident$_{n,1}$}\ldots{\ident$_{n,k_n}$}\,{\tt :}\,% +{\term$_n$} {\tt )}.}\\ + Adds $n$ blocks of parameters with different specifications. + +\item {\tt Conjecture {\ident} :{\term}.}\\ + Is equivalent to {\tt Axiom {\ident} : {\term}}. +\end{Variants} + +\noindent {\bf Remark: } It is possible to replace {\tt Parameter} by +{\tt Parameters}. + + +\subsubsection{{\tt Variable {\ident} :{\term}}. +\comindex{Variable} +\comindex{Variables} +\comindex{Hypothesis} +\comindex{Hypotheses}} + +This command links {\term} to the name {\ident} in the context of the +current section (see Section~\ref{Section} for a description of the section +mechanism). When the current section is closed, name {\ident} will be +unknown and every object using this variable will be explicitly +parameterized (the variable is {\em discharged}). Using the {\tt +Variable} command out of any section is equivalent to {\tt Axiom}. + +\begin{ErrMsgs} +\item \errindex{{\ident} already exists} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Variable {\ident$_1$}\ldots{\ident$_n$} {\tt :}{\term}.}\\ + Links {\term} to names {\ident$_1$}\ldots{\ident$_n$}. +\item + {\tt Variable\,% +(\,{\ident$_{1,1}$}\ldots{\ident$_{1,k_1}$}\,{\tt :}\,{\term$_1$} {\tt )}\,% +\ldots\,{\tt (}\,{\ident$_{n,1}$}\ldots{\ident$_{n,k_n}$}\,{\tt :}\,% +{\term$_n$} {\tt )}.}\\ + Adds $n$ blocks of variables with different specifications. +\item {\tt Hypothesis {\ident} {\tt :}{\term}.} \\ + \texttt{Hypothesis} is a synonymous of \texttt{Variable} +\end{Variants} + +\noindent {\bf Remark: } It is possible to replace {\tt Variable} by +{\tt Variables} and {\tt Hypothesis} by {\tt Hypotheses}. + +It is advised to use the keywords \verb:Axiom: and \verb:Hypothesis: +for logical postulates (i.e. when the assertion {\term} is of sort +\verb:Prop:), and to use the keywords \verb:Parameter: and +\verb:Variable: in other cases (corresponding to the declaration of an +abstract mathematical entity). + +%% +%% Definitions +%% +\subsection{Definitions +\index{Definitions} +\label{Simpl-definitions}} + +Definitions differ from declarations in allowing to give a name to a +term whereas declarations were just giving a type to a name. That is +to say that the name of a defined object can be replaced at any time +by its definition. This replacement is called +$\delta$-conversion\index{delta-reduction@$\delta$-reduction} (see +Section~\ref{delta}). A defined object is accepted by the system if +and only if the defining term is well-typed in the current context of +the definition. Then the type of the name is the type of term. The +defined name is called a {\em constant}\index{Constant} and one says +that {\it the constant is added to the +environment}\index{Environment}. + +A formal presentation of constants and environments is given in +Section~\ref{Typed-terms}. + +\subsubsection{\tt Definition {\ident} := {\term}. +\comindex{Definition}} + +This command binds the value {\term} to the name {\ident} in the +environment, provided that {\term} is well-typed. + +\begin{ErrMsgs} +\item \errindex{{\ident} already exists} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Definition {\ident} {\tt :}{\term$_1$} := {\term$_2$}.}\\ + It checks that the type of {\term$_2$} is definitionally equal to + {\term$_1$}, and registers {\ident} as being of type {\term$_1$}, + and bound to value {\term$_2$}. +\item {\tt Definition {\ident} {\binder$_1$}\ldots{\binder$_n$} + {\tt :}\term$_1$ {\tt :=} {\term$_2$}.}\\ + This is equivalent to \\ + {\tt Definition\,{\ident}\,{\tt :\,forall}\,% + {\binder$_1$}\ldots{\binder$_n$}{\tt ,}\,\term$_1$\,{\tt :=}}\,% + {\tt fun}\,{\binder$_1$}\ldots{\binder$_n$}\,{\tt =>}\,{\term$_2$}\,% + {\tt .} +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{In environment {\dots} the term: {\term$_2$} does not have type + {\term$_1$}}.\\ + \texttt{Actually, it has type {\term$_3$}}. +\end{ErrMsgs} + +\SeeAlso Sections \ref{Opaque}, \ref{Transparent}, \ref{unfold} + +\subsubsection{\tt Let {\ident} := {\term}. +\comindex{Let}} + +This command binds the value {\term} to the name {\ident} in the +environment of the current section. The name {\ident} disappears +when the current section is eventually closed, and, all +persistent objects (such as theorems) defined within the +section and depending on {\ident} are prefixed by the local definition +{\tt let {\ident} := {\term} in}. + +\begin{ErrMsgs} +\item \errindex{{\ident} already exists} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Let {\ident} : {\term$_1$} := {\term$_2$}.} +\end{Variants} + +\SeeAlso Sections \ref{Section} (section mechanism), \ref{Opaque}, +\ref{Transparent} (opaque/transparent constants), \ref{unfold} + +%% +%% Inductive Types +%% +\subsection{Inductive definitions +\index{Inductive definitions} +\label{gal_Inductive_Definitions} +\comindex{Inductive} +\label{Inductive}} + +We gradually explain simple inductive types, simple +annotated inductive types, simple parametric inductive types, +mutually inductive types. We explain also co-inductive types. + +\subsubsection{Simple inductive types} + +The definition of a simple inductive type has the following form: + +\medskip +{\tt +\begin{tabular}{l} +Inductive {\ident} : {\sort} := \\ +\begin{tabular}{clcl} + & {\ident$_1$} &:& {\type$_1$} \\ + | & {\ldots} && \\ + | & {\ident$_n$} &:& {\type$_n$} +\end{tabular} +\end{tabular} +} +\medskip + +The name {\ident} is the name of the inductively defined type and +{\sort} is the universes where it lives. +The names {\ident$_1$}, {\ldots}, {\ident$_n$} +are the names of its constructors and {\type$_1$}, {\ldots}, +{\type$_n$} their respective types. The types of the constructors have +to satisfy a {\em positivity condition} (see Section~\ref{Positivity}) +for {\ident}. This condition ensures the soundness of the inductive +definition. If this is the case, the constants {\ident}, +{\ident$_1$}, {\ldots}, {\ident$_n$} are added to the environment with +their respective types. Accordingly to the universe where +the inductive type lives ({\it e.g.} its type {\sort}), {\Coq} provides a +number of destructors for {\ident}. Destructors are named +{\ident}{\tt\_ind}, {\ident}{\tt \_rec} or {\ident}{\tt \_rect} which +respectively correspond to elimination principles on {\tt Prop}, {\tt +Set} and {\tt Type}. The type of the destructors expresses structural +induction/recursion principles over objects of {\ident}. We give below +two examples of the use of the {\tt Inductive} definitions. + +The set of natural numbers is defined as: +\begin{coq_example} +Inductive nat : Set := + | O : nat + | S : nat -> nat. +\end{coq_example} + +The type {\tt nat} is defined as the least \verb:Set: containing {\tt + O} and closed by the {\tt S} constructor. The constants {\tt nat}, +{\tt O} and {\tt S} are added to the environment. + +Now let us have a look at the elimination principles. They are three : +{\tt nat\_ind}, {\tt nat\_rec} and {\tt nat\_rect}. The type of {\tt + nat\_ind} is: +\begin{coq_example} +Check nat_ind. +\end{coq_example} + +This is the well known structural induction principle over natural +numbers, i.e. the second-order form of Peano's induction principle. +It allows to prove some universal property of natural numbers ({\tt +forall n:nat, P n}) by induction on {\tt n}. + +The types of {\tt nat\_rec} and {\tt nat\_rect} are similar, except +that they pertain to {\tt (P:nat->Set)} and {\tt (P:nat->Type)} +respectively . They correspond to primitive induction principles +(allowing dependent types) respectively over sorts \verb:Set: and +\verb:Type:. The constant {\ident}{\tt \_ind} is always provided, +whereas {\ident}{\tt \_rec} and {\ident}{\tt \_rect} can be impossible +to derive (for example, when {\ident} is a proposition). + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{Variants} +\item +\begin{coq_example*} +Inductive nat : Set := O | S (_:nat). +\end{coq_example*} +In the case where inductive types have no annotations (next section +gives an example of such annotations), the positivity condition +implies that a constructor can be defined by only giving the type of +its arguments. +\end{Variants} + +\subsubsection{Simple annotated inductive types} + +In an annotated inductive types, the universe where the inductive +type is defined is no longer a simple sort, but what is called an +arity, which is a type whose conclusion is a sort. + +As an example of annotated inductive types, let us define the +$even$ predicate: + +\begin{coq_example} +Inductive even : nat -> Prop := + | even_0 : even O + | even_SS : forall n:nat, even n -> even (S (S n)). +\end{coq_example} + +The type {\tt nat->Prop} means that {\tt even} is a unary predicate +(inductively defined) over natural numbers. The type of its two +constructors are the defining clauses of the predicate {\tt even}. The +type of {\tt even\_ind} is: + +\begin{coq_example} +Check even_ind. +\end{coq_example} + +From a mathematical point of view it asserts that the natural numbers +satisfying the predicate {\tt even} are exactly the naturals satisfying +the clauses {\tt even\_0} or {\tt even\_SS}. This is why, when we want +to prove any predicate {\tt P} over elements of {\tt even}, it is +enough to prove it for {\tt O} and to prove that if any natural number +{\tt n} satisfies {\tt P} its double successor {\tt (S (S n))} +satisfies also {\tt P}. This is indeed analogous to the structural +induction principle we got for {\tt nat}. + +\begin{ErrMsgs} +\item \errindex{Non strictly positive occurrence of {\ident} in {\type}} +\item \errindex{The conclusion of {\type} is not valid; it must be +built from {\ident}} +\end{ErrMsgs} + +\subsubsection{Parameterized inductive types} + +Inductive types may be parameterized. Parameters differ from inductive +type annotations in the fact that recursive invokations of inductive +types must always be done with the same values of parameters as its +specification. + +The general scheme is: +\begin{center} +{\tt Inductive} {\ident} {\binder$_1$}\ldots{\binder$_k$} : {\term} := + {\ident$_1$}: {\term$_1$} | {\ldots} | {\ident$_n$}: \term$_n$ +{\tt .} +\end{center} + +A typical example is the definition of polymorphic lists: +\begin{coq_example*} +Inductive list (A:Set) : Set := + | nil : list A + | cons : A -> list A -> list A. +\end{coq_example*} + +Note that in the type of {\tt nil} and {\tt cons}, we write {\tt + (list A)} and not just {\tt list}.\\ The constants {\tt nil} and +{\tt cons} will have respectively types: + +\begin{coq_example} +Check nil. +Check cons. +\end{coq_example} + +Types of destructors are also quantified with {\tt (A:Set)}. + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{Variants} +\item +\begin{coq_example*} +Inductive list (A:Set) : Set := nil | cons (_:A) (_:list A). +\end{coq_example*} +This is an alternative definition of lists where we specify the +arguments of the constructors rather than their full type. +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{The {\num}th argument of {\ident} must be {\ident'} in {\type}} +\end{ErrMsgs} + +\SeeAlso Sections~\ref{Cic-inductive-definitions} and~\ref{elim}. + + +\subsubsection{Mutually defined inductive types +\comindex{Mutual Inductive} +\label{Mutual-Inductive}} + +The definition of a block of mutually inductive types has the form: + +\medskip +{\tt +\begin{tabular}{l} +Inductive {\ident$_1$} : {\type$_1$} := \\ +\begin{tabular}{clcl} + & {\ident$_1^1$} &:& {\type$_1^1$} \\ + | & {\ldots} && \\ + | & {\ident$_{n_1}^1$} &:& {\type$_{n_1}^1$} +\end{tabular} \\ +with\\ +~{\ldots} \\ +with {\ident$_m$} : {\type$_m$} := \\ +\begin{tabular}{clcl} + & {\ident$_1^m$} &:& {\type$_1^m$} \\ + | & {\ldots} \\ + | & {\ident$_{n_m}^m$} &:& {\type$_{n_m}^m$}. +\end{tabular} +\end{tabular} +} +\medskip + +\noindent It has the same semantics as the above {\tt Inductive} +definition for each \ident$_1$, {\ldots}, \ident$_m$. All names +\ident$_1$, {\ldots}, \ident$_m$ and \ident$_1^1$, \dots, +\ident$_{n_m}^m$ are simultaneously added to the environment. Then +well-typing of constructors can be checked. Each one of the +\ident$_1$, {\ldots}, \ident$_m$ can be used on its own. + +It is also possible to parameterize these inductive definitions. +However, parameters correspond to a local +context in which the whole set of inductive declarations is done. For +this reason, the parameters must be strictly the same for each +inductive types The extended syntax is: + +\medskip +{\tt +Inductive {{\ident$_1$} {\params} : {\type$_1$} := \\ +\mbox{}\hspace{0.4cm} {\ident$_1^1$} : {\type$_1^1$} \\ +\mbox{}\hspace{0.1cm}| .. \\ +\mbox{}\hspace{0.1cm}| {\ident$_{n_1}^1$} : {\type$_{n_1}^1$} \\ +with\\ +\mbox{}\hspace{0.1cm} .. \\ +with {\ident$_m$} {\params} : {\type$_m$} := \\ +\mbox{}\hspace{0.4cm}{\ident$_1^m$} : {\type$_1^m$} \\ +\mbox{}\hspace{0.1cm}| .. \\ +\mbox{}\hspace{0.1cm}| {\ident$_{n_m}^m$} : {\type$_{n_m}^m$}. +}} +\medskip + +\Example +The typical example of a mutual inductive data type is the one for +trees and forests. We assume given two types $A$ and $B$ as variables. +It can be declared the following way. + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Variables A B : Set. +Inductive tree : Set := + node : A -> forest -> tree +with forest : Set := + | leaf : B -> forest + | cons : tree -> forest -> forest. +\end{coq_example*} + +This declaration generates automatically six induction +principles. They are respectively +called {\tt tree\_rec}, {\tt tree\_ind}, {\tt + tree\_rect}, {\tt forest\_rec}, {\tt forest\_ind}, {\tt + forest\_rect}. These ones are not the most general ones but are +just the induction principles corresponding to each inductive part +seen as a single inductive definition. + +To illustrate this point on our example, we give the types of {\tt + tree\_rec} and {\tt forest\_rec}. + +\begin{coq_example} +Check tree_rec. +Check forest_rec. +\end{coq_example} + +Assume we want to parameterize our mutual inductive definitions with +the two type variables $A$ and $B$, the declaration should be done the +following way: + +\begin{coq_eval} +Reset tree. +\end{coq_eval} +\begin{coq_example*} +Inductive tree (A B:Set) : Set := + node : A -> forest A B -> tree A B +with forest (A B:Set) : Set := + | leaf : B -> forest A B + | cons : tree A B -> forest A B -> forest A B. +\end{coq_example*} + +Assume we define an inductive definition inside a section. When the +section is closed, the variables declared in the section and occurring +free in the declaration are added as parameters to the inductive +definition. + +\SeeAlso Section~\ref{Section} + +\subsubsection{Co-inductive types +\label{CoInductiveTypes} +\comindex{CoInductive}} + +The objects of an inductive type are well-founded with respect to the +constructors of the type. In other words, such objects contain only a +{\it finite} number constructors. Co-inductive types arise from +relaxing this condition, and admitting types whose objects contain an +infinity of constructors. Infinite objects are introduced by a +non-ending (but effective) process of construction, defined in terms +of the constructors of the type. + +An example of a co-inductive type is the type of infinite sequences of +natural numbers, usually called streams. It can be introduced in \Coq\ +using the \texttt{CoInductive} command: +\begin{coq_example} +CoInductive Stream : Set := + Seq : nat -> Stream -> Stream. +\end{coq_example} + +The syntax of this command is the same as the command \texttt{Inductive} +(cf. Section~\ref{gal_Inductive_Definitions}). Notice that no +principle of induction is derived from the definition of a +co-inductive type, since such principles only make sense for inductive +ones. For co-inductive ones, the only elimination principle is case +analysis. For example, the usual destructors on streams +\texttt{hd:Stream->nat} and \texttt{tl:Str->Str} can be defined as +follows: +\begin{coq_example} +Definition hd (x:Stream) := let (a,s) := x in a. +Definition tl (x:Stream) := let (a,s) := x in s. +\end{coq_example} + +Definition of co-inductive predicates and blocks of mutually +co-inductive definitions are also allowed. An example of a +co-inductive predicate is the extensional equality on streams: + +\begin{coq_example} +CoInductive EqSt : Stream -> Stream -> Prop := + eqst : + forall s1 s2:Stream, + hd s1 = hd s2 -> EqSt (tl s1) (tl s2) -> EqSt s1 s2. +\end{coq_example} + +In order to prove the extensionally equality of two streams $s_1$ and +$s_2$ we have to construct and infinite proof of equality, that is, +an infinite object of type $(\texttt{EqSt}\;s_1\;s_2)$. We will see +how to introduce infinite objects in Section~\ref{CoFixpoint}. + +%% +%% (Co-)Fixpoints +%% +\subsection{Definition of recursive functions} + +\subsubsection{\tt Fixpoint {\ident} {\params} {\tt \{struct} + \ident$_0$ {\tt \}} : type$_0$ := \term$_0$ +\comindex{Fixpoint} +\label{Fixpoint}} + +This command allows to define inductive objects using a fixed point +construction. The meaning of this declaration is to define {\it ident} +a recursive function with arguments specified by +{\binder$_1$}\ldots{\binder$_n$} such that {\it ident} applied to +arguments corresponding to these binders has type \type$_0$, and is +equivalent to the expression \term$_0$. The type of the {\ident} is +consequently {\tt forall {\params} {\tt,} \type$_0$} +and the value is equivalent to {\tt fun {\params} {\tt =>} \term$_0$}. + +To be accepted, a {\tt Fixpoint} definition has to satisfy some +syntactical constraints on a special argument called the decreasing +argument. They are needed to ensure that the {\tt Fixpoint} definition +always terminates. The point of the {\tt \{struct \ident {\tt \}}} +annotation is to let the user tell the system which argument decreases +along the recursive calls. This annotation may be left implicit for +fixpoints with one argument. For instance, one can define the addition +function as : + +\begin{coq_example} +Fixpoint add (n m:nat) {struct n} : nat := + match n with + | O => m + | S p => S (add p m) + end. +\end{coq_example} + +The {\tt match} operator matches a value (here \verb:n:) with the +various constructors of its (inductive) type. The remaining arguments +give the respective values to be returned, as functions of the +parameters of the corresponding constructor. Thus here when \verb:n: +equals \verb:O: we return \verb:m:, and when \verb:n: equals +\verb:(S p): we return \verb:(S (add p m)):. + +The {\tt match} operator is formally described +in detail in Section~\ref{Caseexpr}. The system recognizes that in +the inductive call {\tt (add p m)} the first argument actually +decreases because it is a {\em pattern variable} coming from {\tt match + n with}. + +\Example The following definition is not correct and generates an +error message: + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is not correct and should produce **********) +(********* Error: Recursive call to wrongplus ... **********) +\end{coq_eval} +\begin{coq_example} +Fixpoint wrongplus (n m:nat) {struct n} : nat := + match m with + | O => n + | S p => S (wrongplus n p) + end. +\end{coq_example} + +because the declared decreasing argument {\tt n} actually does not +decrease in the recursive call. The function computing the addition +over the second argument should rather be written: + +\begin{coq_example*} +Fixpoint plus (n m:nat) {struct m} : nat := + match m with + | O => n + | S p => S (plus n p) + end. +\end{coq_example*} + +The ordinary match operation on natural numbers can be mimicked in the +following way. +\begin{coq_example*} +Fixpoint nat_match + (C:Set) (f0:C) (fS:nat -> C -> C) (n:nat) {struct n} : C := + match n with + | O => f0 + | S p => fS p (nat_match C f0 fS p) + end. +\end{coq_example*} +The recursive call may not only be on direct subterms of the recursive +variable {\tt n} but also on a deeper subterm and we can directly +write the function {\tt mod2} which gives the remainder modulo 2 of a +natural number. +\begin{coq_example*} +Fixpoint mod2 (n:nat) : nat := + match n with + | O => O + | S p => match p with + | O => S O + | S q => mod2 q + end + end. +\end{coq_example*} +In order to keep the strong normalisation property, the fixed point +reduction will only be performed when the argument in position of the +decreasing argument (which type should be in an inductive definition) +starts with a constructor. + +The {\tt Fixpoint} construction enjoys also the {\tt with} extension +to define functions over mutually defined inductive types or more +generally any mutually recursive definitions. + +\begin{Variants} +\item {\tt Fixpoint {\ident$_1$} {\params$_1$} :{\type$_1$} := {\term$_1$}\\ + with {\ldots} \\ + with {\ident$_m$} {\params$_m$} :{\type$_m$} := {\type$_m$}}\\ + Allows to define simultaneously {\ident$_1$}, {\ldots}, + {\ident$_m$}. +\end{Variants} + +\Example +The size of trees and forests can be defined the following way: +\begin{coq_eval} +Reset Initial. +Variables A B : Set. +Inductive tree : Set := + node : A -> forest -> tree +with forest : Set := + | leaf : B -> forest + | cons : tree -> forest -> forest. +\end{coq_eval} +\begin{coq_example*} +Fixpoint tree_size (t:tree) : nat := + match t with + | node a f => S (forest_size f) + end + with forest_size (f:forest) : nat := + match f with + | leaf b => 1 + | cons t f' => (tree_size t + forest_size f') + end. +\end{coq_example*} +A generic command {\tt Scheme} is useful to build automatically various +mutual induction principles. It is described in Section~\ref{Scheme}. + +\subsubsection{\tt CoFixpoint {\ident} : \type$_0$ := \term$_0$. +\comindex{CoFixpoint} +\label{CoFixpoint}} + +The {\tt CoFixpoint} command introduces a method for constructing an +infinite object of a coinduc\-tive type. For example, the stream +containing all natural numbers can be introduced applying the +following method to the number \texttt{O} (see +Section~\ref{CoInductiveTypes} for the definition of {\tt Stream}, +{\tt hd} and {\tt tl}): +\begin{coq_eval} +Reset Initial. +CoInductive Stream : Set := + Seq : nat -> Stream -> Stream. +Definition hd (x:Stream) := match x with + | Seq a s => a + end. +Definition tl (x:Stream) := match x with + | Seq a s => s + end. +\end{coq_eval} +\begin{coq_example} +CoFixpoint from (n:nat) : Stream := Seq n (from (S n)). +\end{coq_example} + +Oppositely to recursive ones, there is no decreasing argument in a +co-recursive definition. To be admissible, a method of construction +must provide at least one extra constructor of the infinite object for +each iteration. A syntactical guard condition is imposed on +co-recursive definitions in order to ensure this: each recursive call +in the definition must be protected by at least one constructor, and +only by constructors. That is the case in the former definition, where +the single recursive call of \texttt{from} is guarded by an +application of \texttt{Seq}. On the contrary, the following recursive +function does not satisfy the guard condition: + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is not correct and should produce **********) +(***************** Error: Unguarded recursive call *******************) +\end{coq_eval} +\begin{coq_example} +CoFixpoint filter (p:nat -> bool) (s:Stream) : Stream := + if p (hd s) then Seq (hd s) (filter p (tl s)) else filter p (tl s). +\end{coq_example} + +The elimination of co-recursive definition is done lazily, i.e. the +definition is expanded only when it occurs at the head of an +application which is the argument of a case analysis expression. In +any other context, it is considered as a canonical expression which is +completely evaluated. We can test this using the command +\texttt{Eval}, which computes the normal forms of a term: + +\begin{coq_example} +Eval compute in (from 0). +Eval compute in (hd (from 0)). +Eval compute in (tl (from 0)). +\end{coq_example} + +\begin{Variants} +\item{\tt CoFixpoint {\ident$_1$} {\params} :{\type$_1$} := + {\term$_1$}}\\ As for most constructions, arguments of co-fixpoints + expressions can be introduced before the {\tt :=} sign. +\item{\tt CoFixpoint {\ident$_1$} :{\type$_1$} := {\term$_1$}\\ + with\\ + \mbox{}\hspace{0.1cm} $\ldots$ \\ + with {\ident$_m$} : {\type$_m$} := {\term$_m$}}\\ +As in the \texttt{Fixpoint} command (cf. Section~\ref{Fixpoint}), it +is possible to introduce a block of mutually dependent methods. +\end{Variants} + +%% +%% Theorems & Lemmas +%% +\subsection{Statement and proofs} + +A statement claims a goal of which the proof is then interactively done +using tactics. More on the proof editing mode, statements and proofs can be +found in chapter \ref{Proof-handling}. + +\subsubsection{\tt Theorem {\ident} : {\type}. +\comindex{Theorem} +\comindex{Lemma} +\comindex{Remark} +\comindex{Fact}} + +This command binds {\type} to the name {\ident} in the +environment, provided that a proof of {\type} is next given. + +After a statement, {\Coq} needs a proof. + +\begin{Variants} +\item {\tt Lemma {\ident} : {\type}.}\\ +It is a synonymous of \texttt{Theorem} +\item {\tt Remark {\ident} : {\type}.}\\ +It is a synonymous of \texttt{Theorem} +% Same as {\tt Theorem} except +% that if this statement is in one or more levels of sections then the +% name {\ident} will be accessible only prefixed by the sections names +% when the sections (see \ref{Section} and \ref{LongNames}) will be +% closed. +% %All proofs of persistent objects (such as theorems) referring to {\ident} +% %within the section will be replaced by the proof of {\ident}. + \item {\tt Fact {\ident} : {\type}.}\\ +It is a synonymous of \texttt{Theorem} +% Same as {\tt Remark} except +% that the innermost section name is dropped from the full name. +\item {\tt Definition {\ident} : {\type}.} \\ +Allow to define a term of type {\type} using the proof editing mode. It +behaves as {\tt Theorem} but is intended for the interactive +definition of expression which computational behaviour will be used by +further commands. \SeeAlso~\ref{Transparent} and \ref{unfold}. +\end{Variants} + +\subsubsection{{\tt Proof} {\tt .} \dots {\tt Qed} {\tt .} +\comindex{Proof} +\comindex{Qed} +\comindex{Defined} +\comindex{Save} +\comindex{Goal} +\comindex{Admitted}} + +A proof starts by the keyword {\tt Proof}. Then {\Coq} enters the +proof editing mode until the proof is completed. The proof editing +mode essentially contains tactics that are described in chapter +\ref{Tactics}. Besides tactics, there are commands to manage the proof +editing mode. They are described in chapter \ref{Proof-handling}. When +the proof is completed it should be validated and put in the +environment using the keyword {\tt Qed}. +\medskip + +\ErrMsg +\begin{enumerate} +\item \errindex{{\ident} already exists} +\end{enumerate} + +\begin{Remarks} +\item Several statements can be simultaneously opened. +\item Not only other statements but any vernacular command can be given +within the proof editing mode. In this case, the command is +understood as if it would have been given before the statements still to be +proved. +\item {\tt Proof} is recommended but can currently be omitted. On the +opposite, {\tt Qed} (or {\tt Defined}, see below) is mandatory to validate a proof. +\item Proofs ended by {\tt Qed} are declared opaque (see \ref{Opaque}) +and cannot be unfolded by conversion tactics (see \ref{Conversion-tactics}). +To be able to unfold a proof, you should end the proof by {\tt Defined} + (see below). +\end{Remarks} + +\begin{Variants} +\item {\tt Proof} {\tt .} \dots {\tt Defined} {\tt .}\\ + Same as {\tt Proof} {\tt .} \dots {\tt Qed} {\tt .} but the proof is + then declared transparent (see \ref{Transparent}), which means it + can be unfolded in conversion tactics (see \ref{Conversion-tactics}). +\item {\tt Proof} {\tt .} \dots {\tt Save.}\\ + Same as {\tt Proof} {\tt .} \dots {\tt Qed} {\tt .} +\item {\tt Goal} \type \dots {\tt Save} \ident \\ + Same as {\tt Lemma} \ident {\tt :} \type \dots {\tt Save.} + This is intended to be used in the interactive mode. Conversely to named + lemmas, anonymous goals cannot be nested. +\item {\tt Proof.} \dots {\tt Admitted.}\\ + Turns the current conjecture into an axiom and exits editing of + current proof. +\end{Variants} + +% Local Variables: +% mode: LaTeX +% TeX-master: "Reference-Manual" +% End: + +% $Id$ diff --git a/doc/refman/RefMan-ide.tex b/doc/refman/RefMan-ide.tex new file mode 100644 index 0000000000..d1497f1620 --- /dev/null +++ b/doc/refman/RefMan-ide.tex @@ -0,0 +1,327 @@ +\chapter{\Coq{} Integrated Development Environment} +\label{Addoc-coqide} +\ttindex{coqide} + +The \Coq{} Integrated Development Environment is a graphical tool, to +be used as a user-friendly replacement to \texttt{coqtop}. Its main +purpose is to allow the user to navigate forward and backward into a +\Coq{} vernacular file, executing corresponding commands or undoing +them respectively. % CREDITS ? Proof general, lablgtk, ... + +\CoqIDE{} is run by typing the command \verb|coqide| on the command +line. Without argument, the main screen is displayed with an ``unnamed +buffer'', and with a file name as argument, another buffer displaying +the contents of that file. Additionally, coqide accepts the same +options as coqtop, given in Chapter~\ref{Addoc-coqc}, the ones having +obviously no meaning for \CoqIDE{} being ignored. + +\begin{figure}[t] +\begin{center} +%HEVEA\imgsrc{coqide.png} +%BEGIN LATEX +\ifx\pdfoutput\undefined % si on est pas en pdflatex +\includegraphics[width=1.0\textwidth]{coqide.eps} +\else +\includegraphics[width=1.0\textwidth]{coqide.png} +\fi +%END LATEX +\end{center} +\caption{\CoqIDE{} main screen} +\label{fig:coqide} +\end{figure} + +A sample \CoqIDE{} main screen, while navigating into a file +\verb|Fermat.v|, is shown on Figure~\ref{fig:coqide}. At +the top is a menu bar, and a tool bar below it. The large window on +the left is displaying the various \emph{script buffers}. The upper right +window is the \emph{goal window}, where goals to +prove are displayed. The lower right window is the \emph{message window}, +where various messages resulting from commands are displayed. At the +bottom is the status bar. + +\section{Managing files and buffers, basic edition} + +In the script window, you may open arbitrarily many buffers to +edit. The \emph{File} menu allows you to open files or create some, +save them, print or export them into various formats. Among all these +buffers, there is always one which is the current \emph{running + buffer}, whose name is displayed on a green background, which is the +one where Coq commands are currently executed. + +Buffers may be edited as in any text editor, and classical basic +editing commands (Copy/Paste, \ldots) are available in the \emph{Edit} +menu. \CoqIDE{} offers only basic editing commands, so if you need +more complex editing commands, you may launch your favorite text +editor on the current buffer, using the \emph{Edit/External Editor} +menu. + +\section{Interactive navigation into \Coq{} scripts} + +The running buffer is the one where navigation takes place. The +toolbar proposes five basic commands for this. The first one, +represented by a down arrow icon, is for going forward executing one +command. If that command is successful, the part of the script that +has been executed is displayed on a green background. If that command +fails, the error message is displayed in the message window, and the +location of the error is emphasized by a red underline. + +On Figure~\ref{fig:coqide}, the running buffer is \verb|Fermat.v|, all +commands until the \verb|Theorem| have been already executed, and the +user tried to go forward executing \verb|Induction n|. That command +failed because no such tactic exist (tactics are now in +lowercase\ldots), and the wrong word is underlined. + +Notice that the green part of the running buffer is not editable. If +you ever want to modify something you have to go backward using the up +arrow tool, or even better, put the cursor where you want to go back +and use the \textsf{goto} button. Unlike with \verb|coqtop|, you +should never use \verb|Undo| to go backward. + +Two additional tool buttons exist, one to go directly to the end and +one to go back to the beginning. If you try to go to the end, or in +general to run several commands using the \textsf{goto} button, the + execution will stop whenever an error is found. + +If you ever try to execute a command which happens to run during a +long time, and would like to abort it before its +termination, you may use the interrupt button (the white cross on a red circle). + +Finally, notice that these navigation buttons are also available in +the menu, where their keyboard shortcuts are given. + +\section{Try tactics automatically} +\label{sec:trytactics} + +The menu \texttt{Try Tactics} provides some features for automatically +trying to solve the current goal using simple tactics. If such a +tactic succeeds in solving the goal, then its text is automatically +inserted into the script. There is finally a combination of these +tactics, called the \emph{proof wizard} which will try each of them in +turn. This wizard is also available as a tool button (the light +bulb). The set of tactics tried by the wizard is customizable in +the preferences. + +These tactics are general ones, in particular they do not refer to +particular hypotheses. You may also try specific tactics related to +the goal or one of the hypotheses, by clicking with the right mouse +button one the goal or the considered hypothesis. This is the +``contextual menu on goals'' feature, that may be disabled in the +preferences if undesirable. + +\section{Vernacular commands, templates} + +The \texttt{Templates} menu allows to use shortcuts to insert +vernacular commands. This is a nice way to proceed if you are not sure +of the spelling of the command you want. + +Moreover, this menu offers some \emph{templates} which will automatic +insert a complex command like Fixpoint with a convenient shape for its +arguments. + +\section{Queries} + +\begin{figure}[t] +\begin{center} +%HEVEA\imgsrc{coqide-queries.png} +%BEGIN LATEX +\ifx\pdfoutput\undefined % si on est pas en pdflatex +\includegraphics[width=1.0\textwidth]{coqide-queries.eps} +\else +\includegraphics[width=1.0\textwidth]{coqide-queries.png} +\fi +%END LATEX +\end{center} +\caption{\CoqIDE{}: the query window} +\label{fig:querywindow} +\end{figure} + + +We call \emph{query} any vernacular command that do not change the +current state, such as \verb|Check|, \verb|SearchAbout|, etc. Those +commands are of course useless during compilation of a file, hence +should not be included in scripts. To run such commands without +writing them in the script, \CoqIDE{} offers another input window +called the \emph{query window}. This window can be displayed on +demand, either by using the \texttt{Window} menu, or directly using +shortcuts given in the \texttt{Queries} menu. Indeed, with \CoqIDE{} +the simplest way to perform a \texttt{SearchAbout} on some identifier +is to select it using the mouse, and pressing \verb|F2|. This will +both make appear the query window and run the \texttt{SearchAbout} in +it, displaying the result. Shortcuts \verb|F3| and \verb|F4| are for +\verb|Check| and \verb|Print| respectively. +Figure~\ref{fig:querywindow} displays the query window after selection +of the word ``mult'' in the script windows, and pressing \verb|F4| to +print its definition. + +\section{Compilation} + +The \verb|Compile| menu offers direct commands to: +\begin{itemize} +\item compile the current buffer +\item run a compilation using \verb|make| +\item go to the last compilation error +\item create a \verb|makefile| using \verb|coq_makefile|. +\end{itemize} + +\section{Customizations} + +You may customize your environment using menu +\texttt{Edit/Preferences}. A new window will be displayed, with +several customization sections presented as a notebook. + +The first section is for selecting the text font used for scripts, goal +and message windows. + +The second section is devoted to file management: you may +configure automatic saving of files, by periodically saving the +contents into files named \verb|#f#| for each opened file +\verb|f|. You may also activate the \emph{revert} feature: in case a +opened file is modified on the disk by a third party, \CoqIDE{} may read +it again for you. Note that in the case you edited that same file, you +will be prompt to choose to either discard your changes or not. The +\texttt{File charset encoding} choice is described below in +Section~\ref{sec:coqidecharencoding} + + +The \verb|Externals| section allows to customize the external commands +for compilation, printing, web browsing. In the browser command, you +may use \verb|%s| to denote the URL to open, for example: % +\verb|mozilla -remote "OpenURL(%s)"|. + +The \verb|Tactics Wizard| section allows to defined the set of tactics +that should be tried, in sequence, to solve the current goal. + +The last section is for miscellaneous boolean settings, such as the +``contextual menu on goals'' feature presented in +Section~\ref{sec:trytactics}. + +Notice that these settings are saved in the file \verb|.coqiderc| of +your home directory. + +A gtk2 accelerator keymap is saved under the name \verb|.coqide.keys|. +This file should not be edited manually: to modify a given menu +shortcut, go to the corresponding menu item without releasing the +mouse button, press the key you want for the new shortcut, and release +the mouse button afterwards. + +For experts: it is also possible to set up a specific gtk resource +file, under the name \verb|.coqide-gtk2rc|, following the gtk2 +resources syntax +\url{http://developer.gnome.org/doc/API/2.0/gtk/gtk-Resource-Files.html}. +Such a default resource file exists in the \Coq{} library, you may +copy this file into your home directory, and edit it using any text +editor, \CoqIDE{} itself for example. + +\section{Using unicode symbols} + +\CoqIDE{} supports unicode character encoding in its text windows, +consequently a large set of symbols is available for notations. + +\subsection{Displaying unicode symbols} + +You just need to define suitable notations as described in +Chapter~\ref{Addoc-syntax}. For example, to use the mathematical symbols +$\forall$ and $\exists$, you may define +\begin{quote}\tt +Notation "$\forall$ x : t, P" := \\ +\qquad (forall x:t, P) (at level 200, x ident).\\ +Notation "$\exists$ x : t, P" := \\ +\qquad (exists x:t, P) (at level 200, x ident). +\end{quote} +There exists a small set of such notations already defined, in the +file \verb|utf8.v| of \Coq{} library, so you may enable them just by +\verb|Require utf8| inside \CoqIDE{}, or equivalently, by starting +\CoqIDE{} with \verb|coqide -l utf8|. + +However, there are some issues when using such unicode symbols: you of +course need to use a character font which supports them. In the Fonts +section of the preferences, the Preview line displays some unicode symbols, so +you could figure out if the selected font is OK. Related to this, one +thing you may need to do is choose whether Gtk should use antialiased +fonts or not, by setting the environment variable \verb|GDK_USE_XFT| +to 1 or 0 respectively. + +\subsection{Defining an input method for non ASCII symbols} + +To input an Unicode symbol, a general method is to press both the +CONTROL and the SHIFT keys, and type the hexadecimal code of the +symbol required, for example \verb|2200| for the $\forall$ symbol. +A list of symbol codes is available at \url{http://www.unicode.org}. + +Of course, this method is painful for symbols you use often. There is +always the possibility to copy-paste a symbol already typed in. +Another method is to bind some key combinations for frequently used +symbols. For example, to bind keys \verb|F11| and \verb|F12| to +$\forall$ and $\exists$ respectively, you may add +\begin{quote}\tt + bind "F11" {"insert-at-cursor" ("$\forall$")}\\ + bind "F12" {"insert-at-cursor" ("$\exists$")} +\end{quote} +to your \verb|binding "text"| section in \verb|.coqiderc-gtk2rc|. + + +% such a binding is system-dependent. We +% give here a solution for X11: +% \begin{itemize} +% \item first, using \verb|xmodmap|, bind some key combination into a +% new key name such as Fxx where xx greater that 12: for example (on a +% french keyboard) +% \begin{quote}\tt +% xmodmap -e "keycode 24 = a A F13 F13" \\ +% xmodmap -e "keycode 26 = e E F14 F14" +% \end{quote} +% will rebind "a" to F13 and "e" to F14. +% \item then add +% \begin{quote}\tt +% bind "F13" {"insert-at-cursor" ("$\forall$")}\\ +% bind "F14" {"insert-at-cursor" ("$\exists$")} +% \end{quote} +% to your \verb|binding "text"| section in \verb|.coqiderc-gtk2rc|. +% The strange \verb|∀| argument is the UTF-8 encoding for +% 0x2200, that is the symbol $\forall$. Computing UTF-8 encoding +% for a unicode can be done in various ways, including +% launching a \verb|lablgtk2| toplevel and use +%\begin{verbatim} +% Glib.Utf8.from_unichar 0x2200;; +%\end{verbatim} +%\end{itemize} + +\subsection{Character encoding for saved files} +\label{sec:coqidecharencoding} + +In the \texttt{Files} section of the preferences, the encoding option +is related to the way files are saved. + +If you have no need to exchange files with non UTF-8 aware +applications, it is better to choose the UTF-8 encoding, since it +guarantees that your files will be read again without problems. (This +is because when \CoqIDE{} reads a file, it tries to automatically +detect its character encoding.) + +If you choose something else than UTF-8, then missing characters will +be written encoded by \verb|\x{....}| or \verb|\x{........}| where +each dot is an hexadecimal digit: the number between braces is the +hexadecimal UNICODE index for the missing character. + + +\section{Building a custom \CoqIDE{} with user \textsc{ML} code} + +You can do this as described in Section~\ref{Coqmktop} for a +custom coq text toplevel, simply by adding +option \verb|-ide| to \verb|coqmktop|, that is something like +\begin{quote} +\texttt{coqmktop -ide -byte $m_1$.cmo \ldots{} $m_n$.cmo} +\end{quote} +or +\begin{quote} +\texttt{coqmktop -ide -opt $m_1$.cmx \ldots{} $m_n$.cmx} +\end{quote} + + + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-ind.tex b/doc/refman/RefMan-ind.tex new file mode 100755 index 0000000000..3389382af0 --- /dev/null +++ b/doc/refman/RefMan-ind.tex @@ -0,0 +1,498 @@ + +%\documentstyle[11pt]{article} +%\input{title} + +%\include{macros} +%\makeindex + +%\begin{document} +%\coverpage{The module {\tt Equality}}{Cristina CORNES} + +%\tableofcontents + +\chapter{Tactics for inductive types and families} +\label{Addoc-equality} + +This chapter details a few special tactics useful for inferring facts +from inductive hypotheses. They can be considered as tools that +macro-generate complicated uses of the basic elimination tactics for +inductive types. + +Sections \ref{inversion_introduction} to \ref{inversion_using} present +inversion tactics and section \ref{scheme} describes +a command {\tt Scheme} for automatic generation of induction schemes +for mutual inductive types. + +%\end{document} +%\documentstyle[11pt]{article} +%\input{title} + +%\begin{document} +%\coverpage{Module Inv: Inversion Tactics}{Cristina CORNES} + +\section{Generalities about inversion} +\label{inversion_introduction} +When working with (co)inductive predicates, we are very often faced to +some of these situations: +\begin{itemize} +\item we have an inconsistent instance of an inductive predicate in the + local context of hypotheses. Thus, the current goal can be trivially + proved by absurdity. + +\item we have a hypothesis that is an instance of an inductive + predicate, and the instance has some variables whose constraints we + would like to derive. +\end{itemize} + +The inversion tactics are very useful to simplify the work in these +cases. Inversion tools can be classified in three groups: +\begin{enumerate} +\item tactics for inverting an instance without stocking the inversion + lemma in the context: + (\texttt{Dependent}) \texttt{Inversion} and + (\texttt{Dependent}) \texttt{Inversion\_clear}. +\item commands for generating and stocking in the context the inversion + lemma corresponding to an instance: \texttt{Derive} + (\texttt{Dependent}) \texttt{Inversion}, \texttt{Derive} + (\texttt{Dependent}) \texttt{Inversion\_clear}. +\item tactics for inverting an instance using an already defined + inversion lemma: \texttt{Inversion \ldots using}. +\end{enumerate} + +These tactics work for inductive types of arity $(\vec{x}:\vec{T})s$ +where $s \in \{Prop,Set,Type\}$. Sections \ref{inversion_primitive}, +\ref{inversion_derivation} and \ref{inversion_using} +describe respectively each group of tools. + +As inversion proofs may be large in size, we recommend the user to +stock the lemmas whenever the same instance needs to be inverted +several times.\\ + +Let's consider the relation \texttt{Le} over natural numbers and the +following variables: + +\begin{coq_eval} +Restore State "Initial". +\end{coq_eval} + +\begin{coq_example*} +Inductive Le : nat -> nat -> Set := + | LeO : forall n:nat, Le 0%N n + | LeS : forall n m:nat, Le n m -> Le (S n) (S m). +Variable P : nat -> nat -> Prop. +Variable Q : forall n m:nat, Le n m -> Prop. +\end{coq_example*} + +For example purposes we defined \verb+Le: nat->nat->Set+ + but we may have defined +it \texttt{Le} of type \verb+nat->nat->Prop+ or \verb+nat->nat->Type+. + + +\section{Inverting an instance} +\label{inversion_primitive} +\subsection{The non dependent case} +\begin{itemize} + +\item \texttt{Inversion\_clear} \ident~\\ +\index{Inversion-clear@{\tt Inversion\_clear}} + Let the type of \ident~ in the local context be $(I~\vec{t})$, + where $I$ is a (co)inductive predicate. Then, + \texttt{Inversion} applied to \ident~ derives for each possible + constructor $c_i$ of $(I~\vec{t})$, {\bf all} the necessary + conditions that should hold for the instance $(I~\vec{t})$ to be + proved by $c_i$. Finally it erases \ident~ from the context. + + + +For example, consider the goal: +\begin{coq_eval} +Lemma ex : forall n m:nat, Le (S n) m -> P n m. +intros. +\end{coq_eval} + +\begin{coq_example} +Show. +\end{coq_example} + +To prove the goal we may need to reason by cases on \texttt{H} and to + derive that \texttt{m} is necessarily of +the form $(S~m_0)$ for certain $m_0$ and that $(Le~n~m_0)$. +Deriving these conditions corresponds to prove that the +only possible constructor of \texttt{(Le (S n) m)} is +\texttt{LeS} and that we can invert the +\texttt{->} in the type of \texttt{LeS}. +This inversion is possible because \texttt{Le} is the smallest set closed by +the constructors \texttt{LeO} and \texttt{LeS}. + + +\begin{coq_example} +inversion_clear H. +\end{coq_example} + +Note that \texttt{m} has been substituted in the goal for \texttt{(S m0)} +and that the hypothesis \texttt{(Le n m0)} has been added to the +context. + +\item \texttt{Inversion} \ident~\\ +\index{Inversion@{\tt Inversion}} + This tactic differs from {\tt Inversion\_clear} in the fact that + it adds the equality constraints in the context and + it does not erase the hypothesis \ident. + + +In the previous example, {\tt Inversion\_clear} +has substituted \texttt{m} by \texttt{(S m0)}. Sometimes it is +interesting to have the equality \texttt{m=(S m0)} in the +context to use it after. In that case we can use \texttt{Inversion} that +does not clear the equalities: + +\begin{coq_example*} +Undo. +\end{coq_example*} +\begin{coq_example} +inversion H. +\end{coq_example} + +\begin{coq_eval} +Undo. +\end{coq_eval} + +Note that the hypothesis \texttt{(S m0)=m} has been deduced and +\texttt{H} has not been cleared from the context. + +\end{itemize} + +\begin{Variants} + +\item \texttt{Inversion\_clear } \ident~ \texttt{in} \ident$_1$ \ldots + \ident$_n$\\ +\index{Inversion_clear...in@{\tt Inversion\_clear...in}} + Let \ident$_1$ \ldots \ident$_n$, be identifiers in the local context. This + tactic behaves as generalizing \ident$_1$ \ldots \ident$_n$, and then performing + {\tt Inversion\_clear}. + +\item \texttt{Inversion } \ident~ \texttt{in} \ident$_1$ \ldots \ident$_n$\\ +\index{Inversion ... in@{\tt Inversion ... in}} + Let \ident$_1$ \ldots \ident$_n$, be identifiers in the local context. This + tactic behaves as generalizing \ident$_1$ \ldots \ident$_n$, and then performing + \texttt{Inversion}. + + +\item \texttt{Simple Inversion} \ident~ \\ +\index{Simple Inversion@{\tt Simple Inversion}} + It is a very primitive inversion tactic that derives all the necessary + equalities but it does not simplify + the constraints as \texttt{Inversion} and + {\tt Inversion\_clear} do. + +\end{Variants} + + +\subsection{The dependent case} +\begin{itemize} +\item \texttt{Dependent Inversion\_clear} \ident~\\ +\index{Dependent Inversion-clear@{\tt Dependent Inversion\_clear}} + Let the type of \ident~ in the local context be $(I~\vec{t})$, + where $I$ is a (co)inductive predicate, and let the goal depend both on + $\vec{t}$ and \ident. Then, + \texttt{Dependent Inversion\_clear} applied to \ident~ derives + for each possible constructor $c_i$ of $(I~\vec{t})$, {\bf all} the + necessary conditions that should hold for the instance $(I~\vec{t})$ to be + proved by $c_i$. It also substitutes \ident~ for the corresponding + term in the goal and it erases \ident~ from the context. + + +For example, consider the goal: +\begin{coq_eval} +Lemma ex_dep : forall (n m:nat) (H:Le (S n) m), Q (S n) m H. +intros. +\end{coq_eval} + +\begin{coq_example} +Show. +\end{coq_example} + +As \texttt{H} occurs in the goal, we may want to reason by cases on its +structure and so, we would like inversion tactics to +substitute \texttt{H} by the corresponding term in constructor form. +Neither \texttt{Inversion} nor {\tt Inversion\_clear} make such a +substitution. To have such a behavior we use the dependent inversion tactics: + +\begin{coq_example} +dependent inversion_clear H. +\end{coq_example} + +Note that \texttt{H} has been substituted by \texttt{(LeS n m0 l)} and +\texttt{m} by \texttt{(S m0)}. + + +\end{itemize} + +\begin{Variants} + +\item \texttt{Dependent Inversion\_clear } \ident~ \texttt{ with } \term\\ +\index{Dependent Inversion_clear...with@{\tt Dependent Inversion\_clear...with}} + \noindent Behaves as \texttt{Dependent Inversion\_clear} but allows to give + explicitly the good generalization of the goal. It is useful when + the system fails to generalize the goal automatically. If + \ident~ has type $(I~\vec{t})$ and $I$ has type + $(\vec{x}:\vec{T})s$, then \term~ must be of type + $I:(\vec{x}:\vec{T})(I~\vec{x})\rightarrow s'$ where $s'$ is the + type of the goal. + + + +\item \texttt{Dependent Inversion} \ident~\\ +\index{Dependent Inversion@{\tt Dependent Inversion}} + This tactic differs from \texttt{Dependent Inversion\_clear} in the fact that + it also adds the equality constraints in the context and + it does not erase the hypothesis \ident~. + +\item \texttt{Dependent Inversion } \ident~ \texttt{ with } \term \\ +\index{Dependent Inversion...with@{\tt Dependent Inversion...with}} + Analogous to \texttt{Dependent Inversion\_clear .. with..} above. +\end{Variants} + + + +\section{Deriving the inversion lemmas} +\label{inversion_derivation} +\subsection{The non dependent case} + +The tactics (\texttt{Dependent}) \texttt{Inversion} and (\texttt{Dependent}) +{\tt Inversion\_clear} work on a +certain instance $(I~\vec{t})$ of an inductive predicate. At each +application, they inspect the given instance and derive the +corresponding inversion lemma. If we have to invert the same +instance several times it is recommended to stock the lemma in the +context and to reuse it whenever we need it. + +The families of commands \texttt{Derive Inversion}, \texttt{Derive +Dependent Inversion}, \texttt{Derive} \\ {\tt Inversion\_clear} and \texttt{Derive Dependent Inversion\_clear} +allow to generate inversion lemmas for given instances and sorts. Next +section describes the tactic \texttt{Inversion}$\ldots$\texttt{using} that refines the +goal with a specified inversion lemma. + +\begin{itemize} + +\item \texttt{Derive Inversion\_clear} \ident~ \texttt{with} + $(\vec{x}:\vec{T})(I~\vec{t})$ \texttt{Sort} \sort~ \\ +\index{Derive Inversion_clear...with@{\tt Derive Inversion\_clear...with}} + Let $I$ be an inductive predicate and $\vec{x}$ the variables + occurring in $\vec{t}$. This command generates and stocks + the inversion lemma for the sort \sort~ corresponding to the instance + $(\vec{x}:\vec{T})(I~\vec{t})$ with the name \ident~ in the {\bf + global} environment. When applied it is equivalent to have + inverted the instance with the tactic {\tt Inversion\_clear}. + + + For example, to generate the inversion lemma for the instance + \texttt{(Le (S n) m)} and the sort \texttt{Prop} we do: +\begin{coq_example} +Derive Inversion_clear leminv with (forall n m:nat, Le (S n) m) Sort + Prop. +\end{coq_example} + +Let us inspect the type of the generated lemma: +\begin{coq_example} +Check leminv. +\end{coq_example} + + + +\end{itemize} + +%\variants +%\begin{enumerate} +%\item \verb+Derive Inversion_clear+ \ident$_1$ \ident$_2$ \\ +%\index{Derive Inversion_clear@{\tt Derive Inversion\_clear}} +% Let \ident$_1$ have type $(I~\vec{t})$ in the local context ($I$ +% an inductive predicate). Then, this command has the same semantics +% as \verb+Derive Inversion_clear+ \ident$_2$~ \verb+with+ +% $(\vec{x}:\vec{T})(I~\vec{t})$ \verb+Sort Prop+ where $\vec{x}$ are the free +% variables of $(I~\vec{t})$ declared in the local context (variables +% of the global context are considered as constants). +%\item \verb+Derive Inversion+ \ident$_1$~ \ident$_2$~\\ +%\index{Derive Inversion@{\tt Derive Inversion}} +% Analogous to the previous command. +%\item \verb+Derive Inversion+ $num$ \ident~ \ident~ \\ +%\index{Derive Inversion@{\tt Derive Inversion}} +% This command behaves as \verb+Derive Inversion+ \ident~ {\it +% namehyp} performed on the goal number $num$. +% +%\item \verb+Derive Inversion_clear+ $num$ \ident~ \ident~ \\ +%\index{Derive Inversion_clear@{\tt Derive Inversion\_clear}} +% This command behaves as \verb+Derive Inversion_clear+ \ident~ +% \ident~ performed on the goal number $num$. +%\end{enumerate} + + + +A derived inversion lemma is adequate for inverting the instance +with which it was generated, \texttt{Derive} applied to +different instances yields different lemmas. In general, if we generate +the inversion lemma with +an instance $(\vec{x}:\vec{T})(I~\vec{t})$ and a sort $s$, the inversion lemma will +expect a predicate of type $(\vec{x}:\vec{T})s$ as first argument. \\ + +\begin{Variant} +\item \texttt{Derive Inversion} \ident~ \texttt{with} + $(\vec{x}:\vec{T})(I~\vec{t})$ \texttt{Sort} \sort\\ +\index{Derive Inversion...with@{\tt Derive Inversion...with}} + Analogous of \texttt{Derive Inversion\_clear .. with ..} but + when applied it is equivalent to having + inverted the instance with the tactic \texttt{Inversion}. +\end{Variant} + +\subsection{The dependent case} +\begin{itemize} +\item \texttt{Derive Dependent Inversion\_clear} \ident~ \texttt{with} + $(\vec{x}:\vec{T})(I~\vec{t})$ \texttt{Sort} \sort~ \\ +\index{Derive Dependent Inversion\_clear...with@{\tt Derive Dependent Inversion\_clear...with}} + Let $I$ be an inductive predicate. This command generates and stocks + the dependent inversion lemma for the sort \sort~ corresponding to the instance + $(\vec{x}:\vec{T})(I~\vec{t})$ with the name \ident~ in the {\bf + global} environment. When applied it is equivalent to having + inverted the instance with the tactic \texttt{Dependent Inversion\_clear}. +\end{itemize} + +\begin{coq_example} +Derive Dependent Inversion_clear leminv_dep with + (forall n m:nat, Le (S n) m) Sort Prop. +\end{coq_example} + +\begin{coq_example} +Check leminv_dep. +\end{coq_example} + +\begin{Variants} +\item \texttt{Derive Dependent Inversion} \ident~ \texttt{with} + $(\vec{x}:\vec{T})(I~\vec{t})$ \texttt{Sort} \sort~ \\ +\index{Derive Dependent Inversion...with@{\tt Derive Dependent Inversion...with}} + Analogous to \texttt{Derive Dependent Inversion\_clear}, but when + applied it is equivalent to having + inverted the instance with the tactic \texttt{Dependent Inversion}. + +\end{Variants} + +\section{Using already defined inversion lemmas} +\label{inversion_using} +\begin{itemize} +\item \texttt{Inversion} \ident \texttt{ using} \ident$'$ \\ +\index{Inversion...using@{\tt Inversion...using}} + Let \ident~ have type $(I~\vec{t})$ ($I$ an inductive + predicate) in the local context, and \ident$'$ be a (dependent) inversion + lemma. Then, this tactic refines the current goal with the specified + lemma. + + +\begin{coq_eval} +Abort. +\end{coq_eval} + +\begin{coq_example} +Show. +\end{coq_example} +\begin{coq_example} +inversion H using leminv. +\end{coq_example} + + +\end{itemize} +\variant +\begin{enumerate} +\item \texttt{Inversion} \ident~ \texttt{using} \ident$'$ \texttt{in} \ident$_1$\ldots \ident$_n$\\ +\index{Inversion...using...in@{\tt Inversion...using...in}} +This tactic behaves as generalizing \ident$_1$\ldots \ident$_n$, +then doing \texttt{Use Inversion} \ident~\ident$'$. +\end{enumerate} + +\section{\tt Scheme ...}\index{Scheme@{\tt Scheme}}\label{Scheme} +\label{scheme} +The {\tt Scheme} command is a high-level tool for generating +automatically (possibly mutual) induction principles for given types +and sorts. Its syntax follows the schema : + +\noindent +{\tt Scheme {\ident$_1$} := Induction for \term$_1$ Sort {\sort$_1$} \\ + with\\ + \mbox{}\hspace{0.1cm} .. \\ + with {\ident$_m$} := Induction for {\term$_m$} Sort + {\sort$_m$}}\\ +\term$_1$ \ldots \term$_m$ are different inductive types belonging to +the same package of mutual inductive definitions. This command +generates {\ident$_1$}\ldots{\ident$_m$} to be mutually recursive +definitions. Each term {\ident$_i$} proves a general principle +of mutual induction for objects in type {\term$_i$}. + +\Example +The definition of principle of mutual induction for {\tt tree} and +{\tt forest} over the sort {\tt Set} is defined by the command: +\begin{coq_eval} +Restore State "Initial". +Variables A B : Set. +Inductive tree : Set := + node : A -> forest -> tree +with forest : Set := + | leaf : B -> forest + | cons : tree -> forest -> forest. +\end{coq_eval} +\begin{coq_example*} +Scheme tree_forest_rec := Induction for tree + Sort Set + with forest_tree_rec := Induction for forest Sort Set. +\end{coq_example*} +You may now look at the type of {\tt tree\_forest\_rec} : +\begin{coq_example} +Check tree_forest_rec. +\end{coq_example} +This principle involves two different predicates for {\tt trees} and +{\tt forests}; it also has three premises each one corresponding to a +constructor of one of the inductive definitions. + +The principle {\tt tree\_forest\_rec} shares exactly the same +premises, only the conclusion now refers to the property of forests. +\begin{coq_example} +Check forest_tree_rec. +\end{coq_example} + +\begin{Variant} +\item {\tt Scheme {\ident$_1$} := Minimality for \term$_1$ Sort {\sort$_1$} \\ + with\\ + \mbox{}\hspace{0.1cm} .. \\ + with {\ident$_m$} := Minimality for {\term$_m$} Sort + {\sort$_m$}}\\ +Same as before but defines a non-dependent elimination principle more +natural in case of inductively defined relations. +\end{Variant} + +\Example +With the predicates {\tt odd} and {\tt even} inductively defined as: +\begin{coq_eval} +Restore State "Initial". +\end{coq_eval} +\begin{coq_example*} +Inductive odd : nat -> Prop := + oddS : forall n:nat, even n -> odd (S n) +with even : nat -> Prop := + | evenO : even 0%N + | evenS : forall n:nat, odd n -> even (S n). +\end{coq_example*} +The following command generates a powerful elimination +principle: +\begin{coq_example*} +Scheme odd_even := Minimality for odd Sort Prop + with even_odd := Minimality for even Sort Prop. +\end{coq_example*} +The type of {\tt odd\_even} for instance will be: +\begin{coq_example} +Check odd_even. +\end{coq_example} +The type of {\tt even\_odd} shares the same premises but the +conclusion is {\tt (n:nat)(even n)->(Q n)}. + + + +%\end{document} + +% $Id$ diff --git a/doc/refman/RefMan-int.tex b/doc/refman/RefMan-int.tex new file mode 100755 index 0000000000..b1f4b26b80 --- /dev/null +++ b/doc/refman/RefMan-int.tex @@ -0,0 +1,147 @@ +\setheaders{Introduction} +\chapter*{Introduction} + +This document is the Reference Manual of version \coqversion{} of the \Coq\ +proof assistant. A companion volume, the \Coq\ Tutorial, is provided +for the beginners. It is advised to read the Tutorial first. +A new book~\cite{CoqArt} on practical uses of the \Coq{} system will +be published in 2004 and is a good support for both the beginner and +the advanced user. + +%The system \Coq\ is designed to develop mathematical proofs. It can be +%used by mathematicians to develop mathematical theories and by +%computer scientists to write formal specifications, +The \Coq{} system is designed to develop mathematical proofs, and +especially to write formal specifications, programs and to verify that +programs are correct with respect to their specification. It provides +a specification language named \gallina. Terms of \gallina\ can +represent programs as well as properties of these programs and proofs +of these properties. Using the so-called \textit{Curry-Howard + isomorphism}, programs, properties and proofs are formalized in the +same language called \textit{Calculus of Inductive Constructions}, +that is a $\lambda$-calculus with a rich type system. All logical +judgments in \Coq\ are typing judgments. The very heart of the Coq +system is the type-checking algorithm that checks the correctness of +proofs, in other words that checks that a program complies to its +specification. \Coq\ also provides an interactive proof assistant to +build proofs using specific programs called \textit{tactics}. + +All services of the \Coq\ proof assistant are accessible by +interpretation of a command language called \textit{the vernacular}. + +\Coq\ has an interactive mode in which commands are interpreted as the +user types them in from the keyboard and a compiled mode where +commands are processed from a file. + +\begin{itemize} +\item The interactive mode may be used as a debugging mode in which + the user can develop his theories and proofs step by step, + backtracking if needed and so on. The interactive mode is run with + the {\tt coqtop} command from the operating system (which we shall + assume to be some variety of UNIX in the rest of this document). +\item The compiled mode acts as a proof checker taking a file + containing a whole development in order to ensure its correctness. + Moreover, \Coq's compiler provides an output file containing a + compact representation of its input. The compiled mode is run with + the {\tt coqc} command from the operating system. + +\end{itemize} +These two modes are documented in chapter \ref{Addoc-coqc}. + +Other modes of interaction with \Coq{} are possible: through an emacs +shell window, an emacs generic user-interface for proof assistant +(ProofGeneral~\cite{ProofGeneral}) or through a customized interface +(PCoq~\cite{Pcoq}). These facilities are not documented here. There +is also a \Coq{} Integrated Development Environment described in +Chapter~\ref{Addoc-coqide}. + +\section*{How to read this book} + +This is a Reference Manual, not a User Manual, then it is not made for a +continuous reading. However, it has some structure that is explained +below. + +\begin{itemize} +\item The first part describes the specification language, + Gallina. Chapters~\ref{Gallina} and~\ref{Gallina-extension} + describe the concrete syntax as well as the meaning of programs, + theorems and proofs in the Calculus of Inductive + Constructions. Chapter~\ref{Theories} describes the standard library + of \Coq. Chapter~\ref{Cic} is a mathematical description of the + formalism. Chapter~\ref{chapter:Modules} describes the module system. + +\item The second part describes the proof engine. It is divided in + five chapters. Chapter~\ref{Vernacular-commands} presents all + commands (we call them \emph{vernacular commands}) that are not + directly related to interactive proving: requests to the + environment, complete or partial evaluation, loading and compiling + files. How to start and stop proofs, do multiple proofs in parallel + is explained in Chapter~\ref{Proof-handling}. In + Chapter~\ref{Tactics}, all commands that realize one or more steps + of the proof are presented: we call them \emph{tactics}. The + language to combine these tactics into complex proof strategies is + given in Chapter~\ref{TacticLanguage}. Examples of tactics are + described in Chapter~\ref{Tactics-examples}. + +%\item The third part describes how to extend the system in two ways: +% adding parsing and pretty-printing rules +% (Chapter~\ref{Addoc-syntax}) and writing new tactics +% (Chapter~\ref{TacticLanguage}). + +\item The third part describes how to extend the syntax of \Coq. It +corresponds to the Chapter~\ref{Addoc-syntax}. + +\item In the fourth part more practical tools are documented. First in + Chapter~\ref{Addoc-coqc}, the usage of \texttt{coqc} (batch mode) + and \texttt{coqtop} (interactive mode) with their options is + described. Then, in Chapter~\ref{Utilities}, + various utilities that come with the \Coq\ distribution are + presented. + Finally, Chapter~\ref{Addoc-coqide} describes the \Coq{} integrated + development environment. +\end{itemize} + +At the end of the document, after the global index, the user can find +specific indexes for tactics, vernacular commands, and error +messages. + +\section*{List of additional documentation} + +This manual does not contain all the documentation the user may need +about \Coq{}. Various informations can be found in the following +documents: +\begin{description} + +\item[Tutorial] + A companion volume to this reference manual, the \Coq{} Tutorial, is + aimed at gently introducing new users to developing proofs in \Coq{} + without assuming prior knowledge of type theory. In a second step, the + user can read also the tutorial on recursive types (document {\tt + RecTutorial.ps}). + +\item[Addendum] The fifth part (the Addendum) of the Reference Manual + is distributed as a separate document. It contains more + detailed documentation and examples about some specific aspects of the + system that may interest only certain users. It shares the indexes, + the page numbers and + the bibliography with the Reference Manual. If you see in one of the + indexes a page number that is outside the Reference Manual, it refers + to the Addendum. + +\item[Installation] A text file INSTALL that comes with the sources + explains how to install \Coq{}. + +\item[The \Coq{} standard library] +A commented version of sources of the \Coq{} standard library +(including only the specifications, the proofs are removed) +is given in the additional document {\tt Library.ps}. + +\end{description} + + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-lib.tex b/doc/refman/RefMan-lib.tex new file mode 100755 index 0000000000..5a022c74e5 --- /dev/null +++ b/doc/refman/RefMan-lib.tex @@ -0,0 +1,1102 @@ +\chapter{The {\Coq} library} +\index{Theories}\label{Theories} + +The \Coq\ library is structured into three parts: + +\begin{description} +\item[The initial library:] it contains + elementary logical notions and datatypes. It constitutes the + basic state of the system directly available when running + \Coq; + +\item[The standard library:] general-purpose libraries containing + various developments of \Coq\ axiomatizations about sets, lists, + sorting, arithmetic, etc. This library comes with the system and its + modules are directly accessible through the \verb!Require! command + (see section~\ref{Require}); + +\item[User contributions:] Other specification and proof developments + coming from the \Coq\ users' community. These libraries are no + longer distributed with the system. They are available by anonymous + FTP (see section~\ref{Contributions}). +\end{description} + +This chapter briefly reviews these libraries. + +\section{The basic library} +\label{Prelude} + +This section lists the basic notions and results which are directly +available in the standard \Coq\ system +\footnote{Most of these constructions are defined in the +{\tt Prelude} module in directory {\tt theories/Init} at the {\Coq} +root directory; this includes the modules +{\tt Notations}, +{\tt Logic}, +{\tt Datatypes}, +{\tt Specif}, +{\tt Peano}, +and {\tt Wf}. +Module {\tt Logic\_Type} also makes it in the initial state}. + +\subsection{Notations} \label{Notations} + +This module defines the parsing and pretty-printing of many symbols +(infixes, prefixes, etc.). However, it does not assign a meaning to these +notations. The purpose of this is to define precedence and +associativity of very common notations, and avoid users to use them +with other precedence, which may be confusing. + +\begin{figure} +\begin{center} +\begin{tabular}{|cll|} +\hline +Notation & Precedence & Associativity \\ +\hline +\verb!_ <-> _! & 95 & no \\ +\verb!_ \/ _! & 85 & right \\ +\verb!_ /\ _! & 80 & right \\ +\verb!~ _! & 75 & right \\ +\verb!_ = _! & 70 & no \\ +\verb!_ = _ = _! & 70 & no \\ +\verb!_ = _ :> _! & 70 & no \\ +\verb!_ <> _! & 70 & no \\ +\verb!_ <> _ :> _! & 70 & no \\ +\verb!_ < _! & 70 & no \\ +\verb!_ > _! & 70 & no \\ +\verb!_ <= _! & 70 & no \\ +\verb!_ >= _! & 70 & no \\ +\verb!_ < _ < _! & 70 & no \\ +\verb!_ < _ <= _! & 70 & no \\ +\verb!_ <= _ < _! & 70 & no \\ +\verb!_ <= _ <= _! & 70 & no \\ +\verb!_ + _! & 50 & left \\ +\verb!_ - _! & 50 & left \\ +\verb!_ * _! & 40 & left \\ +\verb!_ / _! & 40 & left \\ +\verb!- _! & 35 & right \\ +\verb!/ _! & 35 & right \\ +\verb!_ ^ _! & 30 & right \\ +\hline +\end{tabular} +\end{center} +\caption{Notations in the initial state} +\label{init-notations} +\end{figure} + +\subsection{Logic} +\label{Logic} + +\begin{figure} +\begin{centerframe} +\begin{tabular}{lclr} +{\form} & ::= & {\tt True} & ({\tt True})\\ + & $|$ & {\tt False} & ({\tt False})\\ + & $|$ & {\tt\char'176} {\form} & ({\tt not})\\ + & $|$ & {\form} {\tt /$\backslash$} {\form} & ({\tt and})\\ + & $|$ & {\form} {\tt $\backslash$/} {\form} & ({\tt or})\\ + & $|$ & {\form} {\tt ->} {\form} & (\em{primitive implication})\\ + & $|$ & {\form} {\tt <->} {\form} & ({\tt iff})\\ + & $|$ & {\tt forall} {\ident} {\tt :} {\type} {\tt ,} + {\form} & (\em{primitive for all})\\ + & $|$ & {\tt exists} {\ident} \zeroone{{\tt :} {\specif}} {\tt + ,} {\form} & ({\tt ex})\\ + & $|$ & {\tt exists2} {\ident} \zeroone{{\tt :} {\specif}} {\tt + ,} {\form} {\tt \&} {\form} & ({\tt ex2})\\ + & $|$ & {\term} {\tt =} {\term} & ({\tt eq})\\ + & $|$ & {\term} {\tt =} {\term} {\tt :>} {\specif} & ({\tt eq}) +\end{tabular} +\end{centerframe} +\caption{Syntax of formulas} +\label{formulas-syntax} +\end{figure} + +The basic library of {\Coq} comes with the definitions of standard +(intuitionistic) logical connectives (they are defined as inductive +constructions). They are equipped with an appealing syntax enriching the +(subclass {\form}) of the syntactic class {\term}. The syntax +extension is shown on figure \ref{formulas-syntax}. + +% The basic library of {\Coq} comes with the definitions of standard +% (intuitionistic) logical connectives (they are defined as inductive +% constructions). They are equipped with an appealing syntax enriching +% the (subclass {\form}) of the syntactic class {\term}. The syntax +% extension \footnote{This syntax is defined in module {\tt +% LogicSyntax}} is shown on Figure~\ref{formulas-syntax}. + +\Rem Implication is not defined but primitive (it is a non-dependent +product of a proposition over another proposition). There is also a +primitive universal quantification (it is a dependent product over a +proposition). The primitive universal quantification allows both +first-order and higher-order quantification. + +\subsubsection{Propositional Connectives} \label{Connectives} +\index{Connectives} + +First, we find propositional calculus connectives: +\ttindex{True} +\ttindex{I} +\ttindex{False} +\ttindex{not} +\ttindex{and} +\ttindex{conj} +\ttindex{proj1} +\ttindex{proj2} + +\begin{coq_eval} +Set Printing Depth 50. +\end{coq_eval} +\begin{coq_example*} +Inductive True : Prop := I. +Inductive False : Prop := . +Definition not (A: Prop) := A -> False. +Inductive and (A B:Prop) : Prop := conj (_:A) (_:B). +Section Projections. +Variables A B : Prop. +Theorem proj1 : A /\ B -> A. +Theorem proj2 : A /\ B -> B. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\ttindex{or} +\ttindex{or\_introl} +\ttindex{or\_intror} +\ttindex{iff} +\ttindex{IF\_then\_else} +\begin{coq_example*} +End Projections. +Inductive or (A B:Prop) : Prop := + | or_introl (_:A) + | or_intror (_:B). +Definition iff (P Q:Prop) := (P -> Q) /\ (Q -> P). +Definition IF_then_else (P Q R:Prop) := P /\ Q \/ ~ P /\ R. +\end{coq_example*} + +\subsubsection{Quantifiers} \label{Quantifiers} +\index{Quantifiers} + +Then we find first-order quantifiers: +\ttindex{all} +\ttindex{ex} +\ttindex{exists} +\ttindex{ex\_intro} +\ttindex{ex2} +\ttindex{exists2} +\ttindex{ex\_intro2} + +\begin{coq_example*} +Definition all (A:Set) (P:A -> Prop) := forall x:A, P x. +Inductive ex (A: Set) (P:A -> Prop) : Prop := + ex_intro (x:A) (_:P x). +Inductive ex2 (A:Set) (P Q:A -> Prop) : Prop := + ex_intro2 (x:A) (_:P x) (_:Q x). +\end{coq_example*} + +The following abbreviations are allowed: +\begin{center} + \begin{tabular}[h]{|l|l|} + \hline + \verb+exists x:A, P+ & \verb+ex A (fun x:A => P)+ \\ + \verb+exists x, P+ & \verb+ex _ (fun x => P)+ \\ + \verb+exists2 x:A, P & Q+ & \verb+ex2 A (fun x:A => P) (fun x:A => Q)+ \\ + \verb+exists2 x, P & Q+ & \verb+ex2 _ (fun x => P) (fun x => Q)+ \\ + \hline + \end{tabular} +\end{center} + +The type annotation \texttt{:A} can be omitted when \texttt{A} can be +synthesized by the system. + +\subsubsection{Equality} \label{Equality} +\index{Equality} + +Then, we find equality, defined as an inductive relation. That is, +given a \verb:Type: \verb:A: and an \verb:x: of type \verb:A:, the +predicate \verb:(eq A x): is the smallest one which contains \verb:x:. +This definition, due to Christine Paulin-Mohring, is equivalent to +define \verb:eq: as the smallest reflexive relation, and it is also +equivalent to Leibniz' equality. + +\ttindex{eq} +\ttindex{refl\_equal} + +\begin{coq_example*} +Inductive eq (A:Type) (x:A) : A -> Prop := + refl_equal : eq A x x. +\end{coq_example*} + +\subsubsection{Lemmas} +\label{PreludeLemmas} + +Finally, a few easy lemmas are provided. + +\ttindex{absurd} + +\begin{coq_example*} +Theorem absurd : forall A C:Prop, A -> ~ A -> C. +\end{coq_example*} +\begin{coq_eval} +Abort. +\end{coq_eval} +\ttindex{sym\_eq} +\ttindex{trans\_eq} +\ttindex{f\_equal} +\ttindex{sym\_not\_eq} +\begin{coq_example*} +Section equality. +Variables A B : Type. +Variable f : A -> B. +Variables x y z : A. +Theorem sym_eq : x = y -> y = x. +Theorem trans_eq : x = y -> y = z -> x = z. +Theorem f_equal : x = y -> f x = f y. +Theorem sym_not_eq : x <> y -> y <> x. +\end{coq_example*} +\begin{coq_eval} +Abort. +Abort. +Abort. +Abort. +\end{coq_eval} +\ttindex{eq\_ind\_r} +\ttindex{eq\_rec\_r} +\ttindex{eq\_rect} +\ttindex{eq\_rect\_r} +%Definition eq_rect: (A:Set)(x:A)(P:A->Type)(P x)->(y:A)(x=y)->(P y). +\begin{coq_example*} +End equality. +Definition eq_ind_r : + forall (A:Type) (x:A) (P:A -> Prop), P x -> forall y:A, y = x -> P y. +Definition eq_rec_r : + forall (A:Type) (x:A) (P:A -> Set), P x -> forall y:A, y = x -> P y. +Definition eq_rect_r : + forall (A:Type) (x:A) (P:A -> Type), P x -> forall y:A, y = x -> P y. +\end{coq_example*} +\begin{coq_eval} +Abort. +Abort. +Abort. +\end{coq_eval} +%Abort (for now predefined eq_rect) +\begin{coq_example*} +Hint Immediate sym_eq sym_not_eq : core. +\end{coq_example*} +\ttindex{f\_equal$i$} + +The theorem {\tt f\_equal} is extended to functions with two to five +arguments. The theorem are names {\tt f\_equal2}, {\tt f\_equal3}, +{\tt f\_equal4} and {\tt f\_equal5}. +For instance {\tt f\_equal3} is defined the following way. +\begin{coq_example*} +Theorem f_equal3 : + forall (A1 A2 A3 B:Type) (f:A1 -> A2 -> A3 -> B) (x1 y1:A1) (x2 y2:A2) + (x3 y3:A3), x1 = y1 -> x2 = y2 -> x3 = y3 -> f x1 x2 x3 = f y1 y2 y3. +\end{coq_example*} +\begin{coq_eval} +Abort. +\end{coq_eval} + +\subsection{Datatypes} +\label{Datatypes} +\index{Datatypes} + +\begin{figure} +\begin{centerframe} +\begin{tabular}{rclr} +{\specif} & ::= & {\specif} {\tt *} {\specif} & ({\tt prod})\\ + & $|$ & {\specif} {\tt +} {\specif} & ({\tt sum})\\ + & $|$ & {\specif} {\tt + \{} {\specif} {\tt \}} & ({\tt sumor})\\ + & $|$ & {\tt \{} {\specif} {\tt \} + \{} {\specif} {\tt \}} & + ({\tt sumbool})\\ + & $|$ & {\tt \{} {\ident} {\tt :} {\specif} {\tt |} {\form} {\tt \}} + & ({\tt sig})\\ + & $|$ & {\tt \{} {\ident} {\tt :} {\specif} {\tt |} {\form} {\tt \&} + {\form} {\tt \}} & ({\tt sig2})\\ + & $|$ & {\tt \{} {\ident} {\tt :} {\specif} {\tt \&} {\specif} {\tt + \}} & ({\tt sigS})\\ + & $|$ & {\tt \{} {\ident} {\tt :} {\specif} {\tt \&} {\specif} {\tt + \&} {\specif} {\tt \}} & ({\tt sigS2})\\ + & & & \\ +{\term} & ::= & {\tt (} {\term} {\tt ,} {\term} {\tt )} & ({\tt pair}) +\end{tabular} +\end{centerframe} +\caption{Syntax of datatypes and specifications} +\label{specif-syntax} +\end{figure} + + +In the basic library, we find the definition\footnote{They are in {\tt + Datatypes.v}} of the basic data-types of programming, again +defined as inductive constructions over the sort \verb:Set:. Some of +them come with a special syntax shown on Figure~\ref{specif-syntax}. + +\subsubsection{Programming} +\label{Programming} +\index{Programming} +\label{libnats} + +\ttindex{unit} +\ttindex{tt} +\ttindex{bool} +\ttindex{true} +\ttindex{false} +\ttindex{nat} +\ttindex{O} +\ttindex{S} +\ttindex{option} +\ttindex{Some} +\ttindex{None} +\ttindex{identity} +\ttindex{refl\_identity} + +\begin{coq_example*} +Inductive unit : Set := tt. +Inductive bool : Set := true | false. +Inductive nat : Set := O | S (n:nat). +Inductive option (A:Set) : Set := Some (_:A) | None. +Inductive identity (A:Type) (a:A) : A -> Type := + refl_identity : identity A a a. +\end{coq_example*} + +Note that zero is the letter \verb:O:, and {\sl not} the numeral +\verb:0:. + +{\tt identity} is logically equivalent to equality but it lives in +sort {\tt Set}. Computationaly, it behaves like {\tt unit}. + +We then define the disjoint sum of \verb:A+B: of two sets \verb:A: and +\verb:B:, and their product \verb:A*B:. +\ttindex{sum} +\ttindex{A+B} +\ttindex{+} +\ttindex{inl} +\ttindex{inr} +\ttindex{prod} +\ttindex{A*B} +\ttindex{*} +\ttindex{pair} +\ttindex{fst} +\ttindex{snd} + +\begin{coq_example*} +Inductive sum (A B:Set) : Set := inl (_:A) | inr (_:B). +Inductive prod (A B:Set) : Set := pair (_:A) (_:B). +Section projections. +Variables A B : Set. +Definition fst (H: prod A B) := match H with + | pair x y => x + end. +Definition snd (H: prod A B) := match H with + | pair x y => y + end. +End projections. +\end{coq_example*} + +\subsection{Specification} + +The following notions\footnote{They are defined in module {\tt +Specif.v}} allows to build new datatypes and specifications. +They are available with the syntax shown on +Figure~\ref{specif-syntax}\footnote{This syntax can be found in the module +{\tt SpecifSyntax.v}}. + +For instance, given \verb|A:Set| and \verb|P:A->Prop|, the construct +\verb+{x:A | P x}+ (in abstract syntax \verb+(sig A P)+) is a +\verb:Set:. We may build elements of this set as \verb:(exist x p): +whenever we have a witness \verb|x:A| with its justification +\verb|p:P x|. + +From such a \verb:(exist x p): we may in turn extract its witness +\verb|x:A| (using an elimination construct such as \verb:match:) but +{\sl not} its justification, which stays hidden, like in an abstract +data type. In technical terms, one says that \verb:sig: is a ``weak +(dependent) sum''. A variant \verb:sig2: with two predicates is also +provided. + +\index{\{x:A "| (P x)\}} +\index{"|} +\ttindex{sig} +\ttindex{exist} +\ttindex{sig2} +\ttindex{exist2} + +\begin{coq_example*} +Inductive sig (A:Set) (P:A -> Prop) : Set := exist (x:A) (_:P x). +Inductive sig2 (A:Set) (P Q:A -> Prop) : Set := + exist2 (x:A) (_:P x) (_:Q x). +\end{coq_example*} + +A ``strong (dependent) sum'' \verb+{x:A & (P x)}+ may be also defined, +when the predicate \verb:P: is now defined as a \verb:Set: +constructor. + +\ttindex{\{x:A \& (P x)\}} +\ttindex{\&} +\ttindex{sigS} +\ttindex{existS} +\ttindex{projS1} +\ttindex{projS2} +\ttindex{sigS2} +\ttindex{existS2} + +\begin{coq_example*} +Inductive sigS (A:Set) (P:A -> Set) : Set := existS (x:A) (_:P x). +Section sigSprojections. +Variable A : Set. +Variable P : A -> Set. +Definition projS1 (H:sigS A P) := let (x, h) := H in x. +Definition projS2 (H:sigS A P) := + match H return P (projS1 H) with + existS x h => h + end. +End sigSprojections. +Inductive sigS2 (A: Set) (P Q:A -> Set) : Set := + existS2 (x:A) (_:P x) (_:Q x). +\end{coq_example*} + +A related non-dependent construct is the constructive sum +\verb"{A}+{B}" of two propositions \verb:A: and \verb:B:. +\label{sumbool} +\ttindex{sumbool} +\ttindex{left} +\ttindex{right} +\ttindex{\{A\}+\{B\}} + +\begin{coq_example*} +Inductive sumbool (A B:Prop) : Set := left (_:A) | right (_:B). +\end{coq_example*} + +This \verb"sumbool" construct may be used as a kind of indexed boolean +data type. An intermediate between \verb"sumbool" and \verb"sum" is +the mixed \verb"sumor" which combines \verb"A:Set" and \verb"B:Prop" +in the \verb"Set" \verb"A+{B}". +\ttindex{sumor} +\ttindex{inleft} +\ttindex{inright} +\ttindex{A+\{B\}} + +\begin{coq_example*} +Inductive sumor (A:Set) (B:Prop) : Set := inleft (_:A) | inright (_:B). +\end{coq_example*} + +We may define variants of the axiom of choice, like in Martin-Löf's +Intuitionistic Type Theory. +\ttindex{Choice} +\ttindex{Choice2} +\ttindex{bool\_choice} + +\begin{coq_example*} +Lemma Choice : + forall (S S':Set) (R:S -> S' -> Prop), + (forall x:S, {y : S' | R x y}) -> + {f : S -> S' | forall z:S, R z (f z)}. +Lemma Choice2 : + forall (S S':Set) (R:S -> S' -> Set), + (forall x:S, {y : S' & R x y}) -> + {f : S -> S' & forall z:S, R z (f z)}. +Lemma bool_choice : + forall (S:Set) (R1 R2:S -> Prop), + (forall x:S, {R1 x} + {R2 x}) -> + {f : S -> bool | + forall x:S, f x = true /\ R1 x \/ f x = false /\ R2 x}. +\end{coq_example*} +\begin{coq_eval} +Abort. +Abort. +Abort. +\end{coq_eval} + +The next constructs builds a sum between a data type \verb|A:Set| and +an exceptional value encoding errors: + +\ttindex{Exc} +\ttindex{value} +\ttindex{error} + +\begin{coq_example*} +Definition Exc := option. +Definition value := Some. +Definition error := None. +\end{coq_example*} + + +This module ends with theorems, +relating the sorts \verb:Set: and +\verb:Prop: in a way which is consistent with the realizability +interpretation. +\ttindex{False\_rec} +\ttindex{eq\_rec} +\ttindex{Except} +\ttindex{absurd\_set} +\ttindex{and\_rec} + +%Lemma False_rec : (P:Set)False->P. +%Lemma False_rect : (P:Type)False->P. +\begin{coq_example*} +Definition except := False_rec. +Notation Except := (except _). +Theorem absurd_set : forall (A:Prop) (C:Set), A -> ~ A -> C. +Theorem and_rec : + forall (A B:Prop) (P:Set), (A -> B -> P) -> A /\ B -> P. +\end{coq_example*} +%\begin{coq_eval} +%Abort. +%Abort. +%\end{coq_eval} + +\subsection{Basic Arithmetics} + +The basic library includes a few elementary properties of natural +numbers, together with the definitions of predecessor, addition and +multiplication\footnote{This is in module {\tt Peano.v}}. It also +provides a scope {\tt nat\_scope} gathering standard notations for +common operations (+,*) and a decimal notation for numbers. That is he +can write \texttt{3} for \texttt{(S (S (S O)))}. This also works on +the left hand side of a \texttt{match} expression (see for example +section~\ref{refine-example}). This scope is opened by default. + +%Remove the redefinition of nat +\begin{coq_eval} +Reset Initial. +\end{coq_eval} + +The following example is not part of the standard library, but it +shows the usage of the notations: + +\begin{coq_example*} +Fixpoint even (n:nat) : bool := + match n with + | 0 => true + | 1 => false + | S (S n) => even n + end. +\end{coq_example*} + + +\ttindex{eq\_S} +\ttindex{pred} +\ttindex{pred\_Sn} +\ttindex{eq\_add\_S} +\ttindex{not\_eq\_S} +\ttindex{IsSucc} +\ttindex{O\_S} +\ttindex{n\_Sn} +\ttindex{plus} +\ttindex{plus\_n\_O} +\ttindex{plus\_n\_Sm} +\ttindex{mult} +\ttindex{mult\_n\_O} +\ttindex{mult\_n\_Sm} + +\begin{coq_example*} +Theorem eq_S : forall x y:nat, x = y -> S x = S y. +\end{coq_example*} +\begin{coq_eval} +Abort. +\end{coq_eval} +\begin{coq_example*} +Definition pred (n:nat) : nat := + match n with + | 0 => 0 + | S u => u + end. +Theorem pred_Sn : forall m:nat, m = pred (S m). +Theorem eq_add_S : forall n m:nat, S n = S m -> n = m. +Hint Immediate eq_add_S : core. +Theorem not_eq_S : forall n m:nat, n <> m -> S n <> S m. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\begin{coq_example*} +Definition IsSucc (n:nat) : Prop := + match n with + | 0 => False + | S p => True + end. +Theorem O_S : forall n:nat, 0 <> S n. +Theorem n_Sn : forall n:nat, n <> S n. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\begin{coq_example*} +Fixpoint plus (n m:nat) {struct n} : nat := + match n with + | 0 => m + | S p => S (plus p m) + end. +Lemma plus_n_O : forall n:nat, n = plus n 0. +Lemma plus_n_Sm : forall n m:nat, S (plus n m) = plus n (S m). +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\begin{coq_example*} +Fixpoint mult (n m:nat) {struct n} : nat := + match n with + | 0 => 0 + | S p => m + mult p m + end. +Lemma mult_n_O : forall n:nat, 0 = mult n 0. +Lemma mult_n_Sm : forall n m:nat, plus (mult n m) n = mult n (S m). +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} + +Finally, it gives the definition of the usual orderings \verb:le:, +\verb:lt:, \verb:ge:, and \verb:gt:. +\ttindex{le} +\ttindex{le\_n} +\ttindex{le\_S} +\ttindex{lt} +\ttindex{ge} +\ttindex{gt} + +\begin{coq_example*} +Inductive le (n:nat) : nat -> Prop := + | le_n : le n n + | le_S : forall m:nat, le n m -> le n (S m). +Infix "+" := plus : nat_scope. +Definition lt (n m:nat) := S n <= m. +Definition ge (n m:nat) := m <= n. +Definition gt (n m:nat) := m < n. +\end{coq_example*} + +Properties of these relations are not initially known, but may be +required by the user from modules \verb:Le: and \verb:Lt:. Finally, +\verb:Peano: gives some lemmas allowing pattern-matching, and a double +induction principle. + +\ttindex{nat\_case} +\ttindex{nat\_double\_ind} + +\begin{coq_example*} +Theorem nat_case : + forall (n:nat) (P:nat -> Prop), P 0 -> (forall m:nat, P (S m)) -> P n. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\begin{coq_example*} +Theorem nat_double_ind : + forall R:nat -> nat -> Prop, + (forall n:nat, R 0 n) -> + (forall n:nat, R (S n) 0) -> + (forall n m:nat, R n m -> R (S n) (S m)) -> forall n m:nat, R n m. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} + +\subsection{Well-founded recursion} + +The basic library contains the basics of well-founded recursion and +well-founded induction\footnote{This is defined in module {\tt Wf.v}}. +\index{Well foundedness} +\index{Recursion} +\index{Well founded induction} +\ttindex{Acc} +\ttindex{Acc\_inv} +\ttindex{Acc\_rec} +\ttindex{well\_founded} + +\begin{coq_example*} +Section Well_founded. +Variable A : Set. +Variable R : A -> A -> Prop. +Inductive Acc : A -> Prop := + Acc_intro : forall x:A, (forall y:A, R y x -> Acc y) -> Acc x. +Lemma Acc_inv : forall x:A, Acc x -> forall y:A, R y x -> Acc y. +\end{coq_example*} +\begin{coq_eval} +simple destruct 1; trivial. +Defined. +\end{coq_eval} +\begin{coq_example*} +Section AccRec. +Variable P : A -> Set. +Variable F : + forall x:A, + (forall y:A, R y x -> Acc y) -> (forall y:A, R y x -> P y) -> P x. +Fixpoint Acc_rec (x:A) (a:Acc x) {struct a} : P x := + F x (Acc_inv x a) + (fun (y:A) (h:R y x) => Acc_rec y (Acc_inv x a y h)). +End AccRec. +Definition well_founded := forall a:A, Acc a. +Hypothesis Rwf : well_founded. +Theorem well_founded_induction : + forall P:A -> Set, + (forall x:A, (forall y:A, R y x -> P y) -> P x) -> forall a:A, P a. +Theorem well_founded_ind : + forall P:A -> Prop, + (forall x:A, (forall y:A, R y x -> P y) -> P x) -> forall a:A, P a. +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +{\tt Acc\_rec} can be used to define functions by fixpoints using +well-founded relations to justify termination. Assuming +extensionality of the functional used for the recursive call, the +fixpoint equation can be proved. +\ttindex{Fix\_F} +\ttindex{fix\_eq} +\ttindex{Fix\_F\_inv} +\ttindex{Fix\_F\_eq} +\begin{coq_example*} +Section FixPoint. +Variable P : A -> Set. +Variable F : forall x:A, (forall y:A, R y x -> P y) -> P x. +Fixpoint Fix_F (x:A) (r:Acc x) {struct r} : P x := + F x (fun (y:A) (p:R y x) => Fix_F y (Acc_inv x r y p)). +Definition Fix (x:A) := Fix_F x (Rwf x). +Hypothesis F_ext : + forall (x:A) (f g:forall y:A, R y x -> P y), + (forall (y:A) (p:R y x), f y p = g y p) -> F x f = F x g. +Lemma Fix_F_eq : + forall (x:A) (r:Acc x), + F x (fun (y:A) (p:R y x) => Fix_F y (Acc_inv x r y p)) = Fix_F x r. +Lemma Fix_F_inv : forall (x:A) (r s:Acc x), Fix_F x r = Fix_F x s. +Lemma fix_eq : forall x:A, Fix x = F x (fun (y:A) (p:R y x) => Fix y). +\end{coq_example*} +\begin{coq_eval} +Abort All. +\end{coq_eval} +\begin{coq_example*} +End FixPoint. +End Well_founded. +\end{coq_example*} + +\subsection{Accessing the {\Type} level} + +The basic library includes the definitions\footnote{This is in module +{\tt Logic\_Type.v}} of the counterparts of some datatypes and logical +quantifiers at the \verb:Type: level: negation, pair, and properties +of {\tt identity}. + +\ttindex{notT} +\ttindex{prodT} +\ttindex{pairT} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Definition notT (A:Type) := A -> False. +Inductive prodT (A B:Type) : Type := pairT (_:A) (_:B). +\end{coq_example*} + + +At the end, it defines datatypes at the {\Type} level. + +\section{The standard library} + +\subsection{Survey} + +The rest of the standard library is structured into the following +subdirectories: + +\begin{tabular}{lp{12cm}} + {\bf Logic} & Classical logic and dependent equality \\ + {\bf Arith} & Basic Peano arithmetic \\ + {\bf ZArith} & Basic integer arithmetic \\ + {\bf Bool} & Booleans (basic functions and results) \\ + {\bf Lists} & Monomorphic and polymorphic lists (basic functions and + results), Streams (infinite sequences defined with co-inductive + types) \\ + {\bf Sets} & Sets (classical, constructive, finite, infinite, power set, + etc.) \\ + {\bf IntMap} & Representation of finite sets by an efficient + structure of map (trees indexed by binary integers).\\ + {\bf Reals} & Axiomatization of Real Numbers (classical, basic functions, + integer part, fractional part, limit, derivative, Cauchy + series, power series and results,... Requires the + \textbf{ZArith} library).\\ + {\bf Relations} & Relations (definitions and basic results). \\ + {\bf Sorting} & Sorted list (basic definitions and heapsort correctness). \\ + {\bf Wellfounded} & Well-founded relations (basic results). \\ + +\end{tabular} +\medskip + +These directories belong to the initial load path of the system, and +the modules they provide are compiled at installation time. So they +are directly accessible with the command \verb!Require! (see +chapter~\ref{Other-commands}). + +The different modules of the \Coq\ standard library are described in the +additional document \verb!Library.dvi!. They are also accessible on the WWW +through the \Coq\ homepage +\footnote{\texttt{http://coq.inria.fr}}. + +\subsection{Notations for integer arithmetics} +\index{Arithmetical notations} + +On figure \ref{zarith-syntax} is described the syntax of expressions +for integer arithmetics. It is provided by requiring and opening the +module {\tt ZArith} and opening scope {\tt Z\_scope}. + +\ttindex{+} +\ttindex{*} +\ttindex{-} +\ttindex{/} +\ttindex{<=} +\ttindex{>=} +\ttindex{<} +\ttindex{>} +\ttindex{?=} +\ttindex{mod} + +\begin{figure} +\begin{center} +\begin{tabular}{l|l|l|l} +Notation & Interpretation & Precedence & Associativity\\ +\hline +\verb!_ < _! & {\tt Zlt} &&\\ +\verb!x <= y! & {\tt Zle} &&\\ +\verb!_ > _! & {\tt Zgt} &&\\ +\verb!x >= y! & {\tt Zge} &&\\ +\verb!x < y < z! & {\tt x < y \verb!/\! y < z} &&\\ +\verb!x < y <= z! & {\tt x < y \verb!/\! y <= z} &&\\ +\verb!x <= y < z! & {\tt x <= y \verb!/\! y < z} &&\\ +\verb!x <= y <= z! & {\tt x <= y \verb!/\! y <= z} &&\\ +\verb!_ ?= _! & {\tt Zcompare} & 70 & no\\ +\verb!_ + _! & {\tt Zplus} &&\\ +\verb!_ - _! & {\tt Zminus} &&\\ +\verb!_ * _! & {\tt Zmult} &&\\ +\verb!_ / _! & {\tt Zdiv} &&\\ +\verb!_ mod _! & {\tt Zmod} & 40 & no \\ +\verb!- _! & {\tt Zopp} &&\\ +\verb!_ ^ _! & {\tt Zpower} &&\\ +\end{tabular} +\end{center} +\label{zarith-syntax} +\caption{Definition of the scope for integer arithmetics ({\tt Z\_scope})} +\end{figure} + +Figure~\ref{zarith-syntax} shows the notations provided by {\tt +Z\_scope}. It specifies how notations are interpreted and, when not +already reserved, the precedence and associativity. + +\begin{coq_example} +Require Import ZArith. +Check (2 + 3)%Z. +Open Scope Z_scope. +Check 2 + 3. +\end{coq_example} + +\subsection{Peano's arithmetic (\texttt{nat})} +\index{Peano's arithmetic} +\ttindex{nat\_scope} + +While in the initial state, many operations and predicates of Peano's +arithmetic are defined, further operations and results belong to other +modules. For instance, the decidability of the basic predicates are +defined here. This is provided by requiring the module {\tt Arith}. + +Figure~\ref{nat-syntax} describes notation available in scope {\tt +nat\_scope}. + +\begin{figure} +\begin{center} +\begin{tabular}{l|l} +Notation & Interpretation \\ +\hline +\verb!_ < _! & {\tt lt} \\ +\verb!x <= y! & {\tt le} \\ +\verb!_ > _! & {\tt gt} \\ +\verb!x >= y! & {\tt ge} \\ +\verb!x < y < z! & {\tt x < y \verb!/\! y < z} \\ +\verb!x < y <= z! & {\tt x < y \verb!/\! y <= z} \\ +\verb!x <= y < z! & {\tt x <= y \verb!/\! y < z} \\ +\verb!x <= y <= z! & {\tt x <= y \verb!/\! y <= z} \\ +\verb!_ + _! & {\tt plus} \\ +\verb!_ - _! & {\tt minus} \\ +\verb!_ * _! & {\tt mult} \\ +\end{tabular} +\end{center} +\label{nat-syntax} +\caption{Definition of the scope for natural numbers ({\tt nat\_scope})} +\end{figure} + +\subsection{Real numbers library} + +\subsubsection{Notations for real numbers} +\index{Notations for real numbers} + +This is provided by requiring and opening the module {\tt Reals} and +opening scope {\tt R\_scope}. This set of notations is very similar to +the notation for integer arithmetics. The inverse function was added. +\begin{figure} +\begin{center} +\begin{tabular}{l|l} +Notation & Interpretation \\ +\hline +\verb!_ < _! & {\tt Rlt} \\ +\verb!x <= y! & {\tt Rle} \\ +\verb!_ > _! & {\tt Rgt} \\ +\verb!x >= y! & {\tt Rge} \\ +\verb!x < y < z! & {\tt x < y \verb!/\! y < z} \\ +\verb!x < y <= z! & {\tt x < y \verb!/\! y <= z} \\ +\verb!x <= y < z! & {\tt x <= y \verb!/\! y < z} \\ +\verb!x <= y <= z! & {\tt x <= y \verb!/\! y <= z} \\ +\verb!_ + _! & {\tt Rplus} \\ +\verb!_ - _! & {\tt Rminus} \\ +\verb!_ * _! & {\tt Rmult} \\ +\verb!_ / _! & {\tt Rdiv} \\ +\verb!- _! & {\tt Ropp} \\ +\verb!/ _! & {\tt Rinv} \\ +\verb!_ ^ _! & {\tt pow} \\ +\end{tabular} +\end{center} +\label{reals-syntax} +\caption{Definition of the scope for real arithmetics ({\tt R\_scope})} +\end{figure} + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example} +Require Import Reals. +Check (2 + 3)%R. +Open Scope R_scope. +Check 2 + 3. +\end{coq_example} + +\subsubsection{Some tactics} + +In addition to the \verb|ring|, \verb|field| and \verb|fourier| +tactics (see Chapter~\ref{Tactics}) there are: +\begin{itemize} +\item {\tt discrR} \tacindex{discrR} + + Proves that a real integer constant $c_1$ is different from another + real integer constant $c_2$. + +\begin{coq_example*} +Require Import DiscrR. +Goal 5 <> 0. +\end{coq_example*} + +\begin{coq_example} +discrR. +\end{coq_example} + +\begin{coq_eval} +Abort. +\end{coq_eval} + +\item {\tt split\_Rabs} allows to unfold {\tt Rabs} constant and splits +corresponding conjonctions. +\tacindex{split\_Rabs} + +\begin{coq_example*} +Require Import SplitAbsolu. +Goal forall x:R, x <= Rabs x. +\end{coq_example*} + +\begin{coq_example} +intro; split_Rabs. +\end{coq_example} + +\begin{coq_eval} +Abort. +\end{coq_eval} + +\item {\tt split\_Rmult} allows to split a condition that a product is + non null into subgoals corresponding to the condition on each + operand of the product. +\tacindex{split\_Rmult} + +\begin{coq_example*} +Require Import SplitRmult. +Goal forall x y z:R, x * y * z <> 0. +\end{coq_example*} + +\begin{coq_example} +intros; split_Rmult. +\end{coq_example} + +\end{itemize} + +All this tactics has been written with the tactic language Ltac +described in Chapter~\ref{TacticLanguage}. More details are available +in document \url{http://coq.inria.fr/~desmettr/Reals.ps}. + +\begin{coq_eval} +Reset Initial. +\end{coq_eval} + +\subsection{List library} +\index{Notations for lists} +\ttindex{length} +\ttindex{head} +\ttindex{tail} +\ttindex{app} +\ttindex{rev} +\ttindex{nth} +\ttindex{map} +\ttindex{flat\_map} +\ttindex{fold\_left} +\ttindex{fold\_right} + +Some elementary operations on polymorphic lists are defined here. They +can be accessed by requiring module {\tt List}. + +It defines the following notions: +\begin{center} +\begin{tabular}{l|l} +\hline +{\tt length} & length \\ +{\tt head} & first element (with default) \\ +{\tt tail} & all but first element \\ +{\tt app} & concatenation \\ +{\tt rev} & reverse \\ +{\tt nth} & accessing $n$-th element (with default) \\ +{\tt map} & applying a function \\ +{\tt flat\_map} & applying a function returning lists \\ +{\tt fold\_left} & iterator (from head to tail) \\ +{\tt fold\_right} & iterator (from tail to head) \\ +\hline +\end{tabular} +\end{center} + +Table show notations available when opening scope {\tt list\_scope}. + +\begin{figure} +\begin{center} +\begin{tabular}{l|l|l|l} +Notation & Interpretation & Precedence & Associativity\\ +\hline +\verb!_ ++ _! & {\tt app} & 60 & right \\ +\verb!_ :: _! & {\tt cons} & 60 & right \\ +\end{tabular} +\end{center} +\label{list-syntax} +\caption{Definition of the scope for lists ({\tt list\_scope})} +\end{figure} + + +\section{Users' contributions} +\index{Contributions} +\label{Contributions} + +Numerous users' contributions have been collected and are available at +URL \url{coq.inria.fr/contribs/}. On this web page, you have a list +of all contributions with informations (author, institution, quick +description, etc.) and the possibility to download them one by one. +There is a small search engine to look for keywords in all +contributions. You will also find informations on how to submit a new +contribution. + +The users' contributions may also be obtained by anonymous FTP from site +\verb:ftp.inria.fr:, in directory \verb:INRIA/coq/: and +searchable on-line at \url{http://coq.inria.fr/contribs-eng.html} + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-ltac.tex b/doc/refman/RefMan-ltac.tex new file mode 100644 index 0000000000..eced8099aa --- /dev/null +++ b/doc/refman/RefMan-ltac.tex @@ -0,0 +1,1057 @@ +\chapter{The tactic language} +\label{TacticLanguage} + +%\geometry{a4paper,body={5in,8in}} + +This chapter gives a compact documentation of Ltac, the tactic +language available in {\Coq}. We start by giving the syntax, and next, +we present the informal semantics. If you want to know more regarding +this language and especially about its fundations, you can refer +to~\cite{Del00}. Chapter~\ref{Tactics-examples} is devoted to giving +examples of use of this language on small but also with non-trivial +problems. + + +\section{Syntax} + +\def\tacexpr{\textrm{\textsl{expr}}} +\def\tacexprlow{\textrm{\textsl{tacexpr$_1$}}} +\def\tacexprinf{\textrm{\textsl{tacexpr$_2$}}} +\def\tacexprpref{\textrm{\textsl{tacexpr$_3$}}} +\def\atom{\textrm{\textsl{atom}}} +\def\recclause{\textrm{\textsl{rec\_clause}}} +\def\letclause{\textrm{\textsl{let\_clause}}} +\def\matchrule{\textrm{\textsl{match\_rule}}} +\def\contextrule{\textrm{\textsl{context\_rule}}} +\def\contexthyps{\textrm{\textsl{context\_hyps}}} +\def\tacarg{\nterm{tacarg}} +\def\cpattern{\nterm{cpattern}} + +The syntax of the tactic language is given Figures~\ref{ltac} +and~\ref{ltac_aux}. See page~\pageref{BNF-syntax} for a description of +the BNF metasyntax used in these grammar rules. Various already defined +entries will be used in this chapter: entries {\naturalnumber}, +{\integer}, {\ident}, {\qualid}, {\term}, {\cpattern} and {\atomictac} +represent respectively the natural and integer numbers, the authorized +identificators and qualified names, {\Coq}'s terms and patterns and +all the atomic tactics described in chapter~\ref{Tactics}. The syntax +of {\cpattern} is the same as that of terms, but there can be specific +variables like {\tt ?id} where {\tt id} is a {\ident} or {\tt \_}, +which are metavariables for pattern matching. {\tt ?id} allows us to +keep instantiations and to make constraints whereas {\tt \_} shows +that we are not interested in what will be matched. On the right hand +side, they are used without the question mark. + +The main entry of the grammar is {\tacexpr}. This language is used in +proof mode but it can also be used in toplevel definitions as shown in +Figure~\ref{ltactop}. + +\begin{Remarks} +\item The infix tacticals ``\dots\ {\tt ||} \dots'' and ``\dots\ {\tt + ;} \dots'' are associative. + +\item As shown by the figure, tactical {\tt ||} binds more than the +prefix tacticals {\tt try}, {\tt repeat}, {\tt do}, {\tt info} and +{\tt abstract} which themselves bind more than the postfix tactical +``{\tt \dots\ ;[ \dots\ ]}'' which binds more than ``\dots\ {\tt ;} +\dots''. + +For instance +\begin{quote} +{\tt try repeat \tac$_1$ || + \tac$_2$;\tac$_3$;[\tac$_{31}$|\dots|\tac$_{3n}$];\tac$_4$.} +\end{quote} +is understood as +\begin{quote} +{\tt (try (repeat (\tac$_1$ || \tac$_2$)));} \\ +{\tt ((\tac$_3$;[\tac$_{31}$|\dots|\tac$_{3n}$]);\tac$_4$).} +\end{quote} +\end{Remarks} + + +\begin{figure}[htbp] +\begin{centerframe} +\begin{tabular}{lcl} +{\tacexpr} & ::= & + {\tacexpr} {\tt ;} {\tacexpr}\\ +& | & {\tacexpr} {\tt ; [} \nelist{\tacexpr}{|} {\tt ]}\\ +& | & {\tacexprpref}\\ +\\ +{\tacexprpref} & ::= & + {\tt do} {\it (}{\naturalnumber} {\it |} {\ident}{\it )} {\tacexprpref}\\ +& | & {\tt info} {\tacexprpref}\\ +& | & {\tt progress} {\tacexprpref}\\ +& | & {\tt repeat} {\tacexprpref}\\ +& | & {\tt try} {\tacexprpref}\\ +& | & {\tacexprinf} \\ +\\ +{\tacexprinf} & ::= & + {\tacexprlow} {\tt ||} {\tacexprpref}\\ +& | & {\tacexprlow}\\ +\\ +{\tacexprlow} & ::= & +{\tt fun} \nelist{\name}{} {\tt =>} {\atom}\\ +& | & +{\tt let} \nelist{\letclause}{\tt with} {\tt in} +{\atom}\\ +& | & +{\tt let rec} \nelist{\recclause}{\tt with} {\tt in} +{\tacexpr}\\ +& | & +{\tt match goal with} \nelist{\contextrule}{\tt |} {\tt end}\\ +& | & +{\tt match reverse goal with} \nelist{\contextrule}{\tt |} {\tt end}\\ +& | & +{\tt match} {\tacexpr} {\tt with} \nelist{\matchrule}{\tt |} {\tt end}\\ +& | & {\tt abstract} {\atom}\\ +& | & {\tt abstract} {\atom} {\tt using} {\ident} \\ +& | & {\tt first [} \nelist{\tacexpr}{\tt |} {\tt ]}\\ +& | & {\tt solve [} \nelist{\tacexpr}{\tt |} {\tt ]}\\ +& | & {\tt idtac} ~|~ {\tt idtac} {\qstring}\\ +& | & {\tt fail} ~|~ {\tt fail} {\naturalnumber} {\qstring}\\ +& | & {\tt fresh} ~|~ {\tt fresh} {\qstring}\\ +& | & {\tt context} {\ident} {\tt [} {\term} {\tt ]}\\ +& | & {\tt eval} {\nterm{redexpr}} {\tt in} {\term}\\ +& | & {\tt type of} {\term}\\ +& | & {\tt constr :} {\term}\\ +& | & \atomictac\\ +& | & {\qualid} \nelist{\tacarg}{}\\ +& | & {\atom}\\ +\\ +{\atom} & ::= & + {\qualid} \\ +& | & ()\\ +& | & {\tt (} {\tacexpr} {\tt )}\\ +\end{tabular} +\end{centerframe} +\caption{Syntax of the tactic language} +\label{ltac} +\end{figure} + + + +\begin{figure}[htbp] +\begin{centerframe} +\begin{tabular}{lcl} +\tacarg & ::= & + {\qualid}\\ +& $|$ & {\tt ()} \\ +& $|$ & {\tt ltac :} {\atom}\\ +& $|$ & {\term}\\ +\\ +\letclause & ::= & {\ident} \sequence{\name}{} {\tt :=} {\tacexpr}\\ +\\ +\recclause & ::= & {\ident} \nelist{\name}{} {\tt :=} {\tacexpr}\\ +\\ +\contextrule & ::= & + \nelist{\contexthyps}{\tt ,} {\tt |-}{\cpattern} {\tt =>} {\tacexpr}\\ +& $|$ & {\tt |-} {\cpattern} {\tt =>} {\tacexpr}\\ +& $|$ & {\tt \_ =>} {\tacexpr}\\ +\\ +\contexthyps & ::= & {\name} {\tt :} {\cpattern}\\ +\\ +\matchrule & ::= & + {\cpattern} {\tt =>} {\tacexpr}\\ +& $|$ & {\tt context} {\zeroone{\ident}} {\tt [} {\cpattern} {\tt ]} {\tt =>} {\tacexpr}\\ +& $|$ & {\tt \_ =>} {\tacexpr}\\ +\end{tabular} +\end{centerframe} +\caption{Syntax of the tactic language (continued)} +\label{ltac_aux} +\end{figure} + +\begin{figure}[ht] +\begin{centerframe} +\begin{tabular}{lcl} +\nterm{top} & ::= & {\tt Ltac} \nelist{\nterm{ltac\_def}} {\tt with} \\ +\\ +\nterm{ltac\_def} & ::= & {\ident} \sequence{\ident}{} {\tt :=} {\tacexpr} +\end{tabular} +\end{centerframe} +\caption{Tactic toplevel definitions} +\label{ltactop} +\end{figure} + + +%% +%% Semantics +%% +\section{Semantics} +%\index[tactic]{Tacticals} +\index{Tacticals} +%\label{Tacticals} + +Tactic expressions can only be applied in the context of a goal. The +evaluation yields either a term, an integer or a tactic. Intermediary +results can be terms or integers but the final result must be a tactic +which is then applied to the current goal. + +There is a special case for {\tt match goal} expressions of which +the clauses evaluate to tactics. Such expressions can only be used as +end result of a tactic expression (never as argument of a local +definition or of an application). + +The rest of this section explains the semantics of every construction +of Ltac. + + +%% \subsection{Values} + +%% Values are given by Figure~\ref{ltacval}. All these values are tactic values, +%% i.e. to be applied to a goal, except {\tt Fun}, {\tt Rec} and $arg$ values. + +%% \begin{figure}[ht] +%% \noindent{}\framebox[6in][l] +%% {\parbox{6in} +%% {\begin{center} +%% \begin{tabular}{lp{0.1in}l} +%% $vexpr$ & ::= & $vexpr$ {\tt ;} $vexpr$\\ +%% & | & $vexpr$ {\tt ; [} {\it (}$vexpr$ {\tt |}{\it )}$^*$ $vexpr$ {\tt +%% ]}\\ +%% & | & $vatom$\\ +%% \\ +%% $vatom$ & ::= & {\tt Fun} \nelist{\inputfun}{} {\tt ->} {\tacexpr}\\ +%% %& | & {\tt Rec} \recclause\\ +%% & | & +%% {\tt Rec} \nelist{\recclause}{\tt And} {\tt In} +%% {\tacexpr}\\ +%% & | & +%% {\tt Match Context With} {\it (}$context\_rule$ {\tt |}{\it )}$^*$ +%% $context\_rule$\\ +%% & | & {\tt (} $vexpr$ {\tt )}\\ +%% & | & $vatom$ {\tt Orelse} $vatom$\\ +%% & | & {\tt Do} {\it (}{\naturalnumber} {\it |} {\ident}{\it )} $vatom$\\ +%% & | & {\tt Repeat} $vatom$\\ +%% & | & {\tt Try} $vatom$\\ +%% & | & {\tt First [} {\it (}$vexpr$ {\tt |}{\it )}$^*$ $vexpr$ {\tt ]}\\ +%% & | & {\tt Solve [} {\it (}$vexpr$ {\tt |}{\it )}$^*$ $vexpr$ {\tt ]}\\ +%% & | & {\tt Idtac}\\ +%% & | & {\tt Fail}\\ +%% & | & {\primitivetactic}\\ +%% & | & $arg$ +%% \end{tabular} +%% \end{center}}} +%% \caption{Values of ${\cal L}_{tac}$} +%% \label{ltacval} +%% \end{figure} + +%% \subsection{Evaluation} + +\subsubsection{Sequence} +\tacindex{;} +\index{Tacticals!;@{\tt {\tac$_1$};\tac$_2$}} + +A sequence is an expression of the following form: +\begin{quote} +{\tacexpr}$_1$ {\tt ;} {\tacexpr}$_2$ +\end{quote} +{\tacexpr}$_1$ and {\tacexpr}$_2$ are evaluated to $v_1$ and +$v_2$. $v_1$ and $v_2$ must be tactic values. $v_1$ is then applied +and $v_2$ is applied to every subgoal generated by the application of +$v_1$. Sequence is left associating. + +\subsubsection{General sequence} +\tacindex{;[\ldots$\mid$\ldots$\mid$\ldots]} +%\tacindex{; [ | ]} +%\index{; [ | ]@{\tt ;[\ldots$\mid$\ldots$\mid$\ldots]}} +\index{Tacticals!; [ | ]@{\tt {\tac$_0$};[{\tac$_1$}$\mid$\ldots$\mid$\tac$_n$]}} + +We can generalize the previous sequence operator as +\begin{quote} +{\tacexpr}$_0$ {\tt ; [} {\tacexpr}$_1$ {\tt |} $...$ {\tt |} +{\tacexpr}$_n$ {\tt ]} +\end{quote} +{\tacexpr}$_i$ is evaluated to $v_i$, for $i=0,...,n$. $v_0$ is +applied and $v_i$ is applied to the $i$-th generated subgoal by the +application of $v_0$, for $=1,...,n$. It fails if the application of +$v_0$ does not generate exactly $n$ subgoals. + +\subsubsection{For loop} +\tacindex{do} +\index{Tacticals!do@{\tt do}} + +There is a for loop that repeats a tactic {\num} times: +\begin{quote} +{\tt do} {\num} {\tacexpr} +\end{quote} +{\tacexpr} is evaluated to $v$. $v$ must be a tactic value. $v$ is +applied {\num} times. Supposing ${\num}>1$, after the first +application of $v$, $v$ is applied, at least once, to the generated +subgoals and so on. It fails if the application of $v$ fails before +the {\num} applications have been completed. + +\subsubsection{Repeat loop} +\tacindex{repeat} +\index{Tacticals!repeat@{\tt repeat}} + +We have a repeat loop with: +\begin{quote} +{\tt repeat} {\tacexpr} +\end{quote} +{\tacexpr} is evaluated to $v$. $v$ must be a tactic value. $v$ is +applied until it fails. Supposing $n>1$, after the first application +of $v$, $v$ is applied, at least once, to the generated subgoals and +so on. It stops when it fails for all the generated subgoals. It never +fails. + +\subsubsection{Error catching} +\tacindex{try} +\index{Tacticals!try@{\tt try}} + +We can catch the tactic errors with: +\begin{quote} +{\tt try} {\tacexpr} +\end{quote} +{\tacexpr} is evaluated to $v$. $v$ must be a tactic value. $v$ is +applied. If the application of $v$ fails, it catches the error and +leaves the goal unchanged. If the level of the exception is positive, +then the exception is re-raised with its level decremented. + +\subsubsection{Detecting progress} +\tacindex{progress} + +We can check if a tactic made progress with: +\begin{quote} +{\tt progress} {\tacexpr} +\end{quote} +{\tacexpr} is evaluated to $v$. $v$ must be a tactic value. $v$ is +applied. If the application of $v$ produced one subgoal equal to the +initial goal (up to syntactical equality), then an error of level 0 is +raised. + +\ErrMsg \errindex{Failed to progress} + +\subsubsection{Branching} +\tacindex{$\mid\mid$} +\index{Tacticals!orelse@{\tt $\mid\mid$}} + +We can easily branch with the following structure: +\begin{quote} +{\tacexpr}$_1$ {\tt ||} {\tacexpr}$_2$ +\end{quote} +{\tacexpr}$_1$ and {\tacexpr}$_2$ are evaluated to $v_1$ and +$v_2$. $v_1$ and $v_2$ must be tactic values. $v_1$ is applied and if +it fails then $v_2$ is applied. Branching is left associating. + +\subsubsection{First tactic to work} +\tacindex{first} +\index{Tacticals!first@{\tt first}} + +We may consider the first tactic to work (i.e. which does not fail) among a +panel of tactics: +\begin{quote} +{\tt first [} {\tacexpr}$_1$ {\tt |} $...$ {\tt |} {\tacexpr}$_n$ {\tt ]} +\end{quote} +{\tacexpr}$_i$ are evaluated to $v_i$ and $v_i$ must be tactic values, for +$i=1,...,n$. Supposing $n>1$, it applies $v_1$, if it works, it stops else it +tries to apply $v_2$ and so on. It fails when there is no applicable tactic. + +\ErrMsg \errindex{No applicable tactic} + +\subsubsection{Solving} +\tacindex{solve} +\index{Tacticals!solve@{\tt solve}} + +We may consider the first to solve (i.e. which generates no subgoal) among a +panel of tactics: +\begin{quote} +{\tt solve [} {\tacexpr}$_1$ {\tt |} $...$ {\tt |} {\tacexpr}$_n$ {\tt ]} +\end{quote} +{\tacexpr}$_i$ are evaluated to $v_i$ and $v_i$ must be tactic values, for +$i=1,...,n$. Supposing $n>1$, it applies $v_1$, if it solves, it stops else it +tries to apply $v_2$ and so on. It fails if there is no solving tactic. + +\ErrMsg \errindex{Cannot solve the goal} + +\subsubsection{Identity} +\tacindex{idtac} +\index{Tacticals!idtac@{\tt idtac}} + +The constant {\tt idtac} is the identity tactic: it leaves any goal +unchanged but it appears in the proof script. +\begin{quote} +{\tt idtac} and {\tt idtac "message"} +\end{quote} +The latter variant prints the string on the standard output. + + +\subsubsection{Failing} +\tacindex{fail} +\index{Tacticals!fail@{\tt fail}} + +The tactic {\tt fail} is the always-failing tactic: it does not solve +any goal. It is useful for defining other tacticals since it can be +catched by {\tt try} or {\tt match goal}. There are three variants: +\begin{quote} +{\tt fail $n$}, {\tt fail "message"} and {\tt fail $n$ "message"} +\end{quote} +The number $n$ is the failure level. If no level is specified, it +defaults to $0$. The level is used by {\tt try} and {\tt match goal}. +If $0$, it makes {\tt match goal} considering the next clause +(backtracking). If non zero, the current {\tt match goal} block or +{\tt try} command is aborted and the level is decremented. + +\ErrMsg \errindex{Tactic Failure "message" (level $n$)}. + +\subsubsection{Local definitions} +\index{Ltac!let} +\index{Ltac!let rec} +\index{let!in Ltac} +\index{let rec!in Ltac} + +Local definitions can be done as follows: +\begin{quote} +{\tt let} {\ident}$_1$ {\tt :=} {\tacexpr}$_1$\\ +{\tt with} {\ident}$_2$ {\tt :=} {\tacexpr}$_2$\\ +...\\ +{\tt with} {\ident}$_n$ {\tt :=} {\tacexpr}$_n$ {\tt in}\\ +{\tacexpr} +\end{quote} +each {\tacexpr}$_i$ is evaluated to $v_i$, then, {\tacexpr} is +evaluated by substituting $v_i$ to each occurrence of {\ident}$_i$, +for $i=1,...,n$. There is no dependencies between the {\tacexpr}$_i$ +and the {\ident}$_i$. + +Local definitions can be recursive by using {\tt let rec} instead of +{\tt let}. Only functions can be defined by recursion, so at least one +argument is required. + +\subsubsection{Application} + +An application is an expression of the following form: +\begin{quote} +{\qualid} {\tacarg}$_1$ ... {\tacarg}$_n$ +\end{quote} +The reference {\qualid} must be bound to some defined tactic +definition expecting at least $n$ arguments. The expressions +{\tacexpr}$_i$ are evaluated to $v_i$, for $i=1,...,n$. +%If {\tacexpr} is a {\tt Fun} or {\tt Rec} value then the body is evaluated by +%substituting $v_i$ to the formal parameters, for $i=1,...,n$. For recursive +%clauses, the bodies are lazily substituted (when an identifier to be evaluated +%is the name of a recursive clause). + +%\subsection{Application of tactic values} + +\subsubsection{Function construction} +\index{fun!in Ltac} +\index{Ltac!fun} + +A parameterized tactic can be built anonymously (without resorting to +local definitions) with: +\begin{quote} +{\tt fun} {\ident${}_1$} ... {\ident${}_n$} {\tt =>} {\tacexpr} +\end{quote} +Indeed, local definitions of functions are a syntactic sugar for +binding a {\tt fun} tactic to an identifier. + +\subsubsection{Pattern matching on terms} +\index{Ltac!match} +\index{match!in Ltac} + +We can carry out pattern matching on terms with: +\begin{quote} +{\tt match} {\tacexpr} {\tt with}\\ +~~~{\cpattern}$_1$ {\tt =>} {\tacexpr}$_1$\\ +~{\tt |} {\cpattern}$_2$ {\tt =>} {\tacexpr}$_2$\\ +~...\\ +~{\tt |} {\cpattern}$_n$ {\tt =>} {\tacexpr}$_n$\\ +~{\tt |} {\tt \_} {\tt =>} {\tacexpr}$_{n+1}$\\ +{\tt end} +\end{quote} +The {\tacexpr} is evaluated and should yield a term which is matched +(non-linear first order unification) against {\cpattern}$_1$ then +{\tacexpr}$_1$ is evaluated into some value by substituting the +pattern matching instantiations to the metavariables. If the matching +with {\cpattern}$_1$ fails, {\cpattern}$_2$ is used and so on. The +pattern {\_} matches any term and shunts all remaining patterns if +any. If {\tacexpr}$_1$ evaluates to a tactic, this tactic is not +immediately applied to the current goal (in contrast with {\tt match +goal}). If all clauses fail (in particular, there is no pattern {\_}) +then a no-matching error is raised. + +\begin{ErrMsgs} + +\item \errindex{No matching clauses for match} + + No pattern can be used and, in particular, there is no {\tt \_} pattern. + +\item \errindex{Argument of match does not evaluate to a term} + + This happens when {\tacexpr} does not denote a term. + +\end{ErrMsgs} + +\index{context!in pattern} +There is a special form of patterns to match a subterm against the +pattern: +\begin{quote} +{\tt context} {\ident} {\tt [} {\cpattern} {\tt ]} +\end{quote} +It matches any term which one subterm matches {\cpattern}. If there is +a match, the optional {\ident} is assign the ``matched context'', that +is the initial term where the matched subterm is replaced by a +hole. The definition of {\tt context} in expressions below will show +how to use such term contexts. + +This operator never makes backtracking. If there are several subterms +matching the pattern, only the first match is considered. Note that +the order of matching is left unspecified. +%% TODO: clarify this point! It *should* be specified + + +\subsubsection{Pattern matching on goals} +\index{Ltac!match goal} +\index{Ltac!match reverse goal} +\index{match goal!in Ltac} +\index{match reverse goal!in Ltac} + +We can make pattern matching on goals using the following expression: +\begin{quote} +\begin{tabbing} +{\tt match goal with}\\ +~~\={\tt |} $hyp_{1,1}${\tt ,}...{\tt ,}$hyp_{1,m_1}$ + ~~{\tt |-}{\cpattern}$_1${\tt =>} {\tacexpr}$_1$\\ + \>{\tt |} $hyp_{2,1}${\tt ,}...{\tt ,}$hyp_{2,m_2}$ + ~~{\tt |-}{\cpattern}$_2${\tt =>} {\tacexpr}$_2$\\ +~~...\\ + \>{\tt |} $hyp_{n,1}${\tt ,}...{\tt ,}$hyp_{n,m_n}$ + ~~{\tt |-}{\cpattern}$_n${\tt =>} {\tacexpr}$_n$\\ + \>{\tt |\_}~~~~{\tt =>} {\tacexpr}$_{n+1}$\\ +{\tt end} +\end{tabbing} +\end{quote} + +% TODO: specify order of hypothesis and explain reverse... + +If each hypothesis pattern $hyp_{1,i}$, with $i=1,...,m_1$ +is matched (non-linear first order unification) by an hypothesis of +the goal and if {\cpattern}$_1$ is matched by the conclusion of the +goal, then {\tacexpr}$_1$ is evaluated to $v_1$ by substituting the +pattern matching to the metavariables and the real hypothesis names +bound to the possible hypothesis names occurring in the hypothesis +patterns. If $v_1$ is a tactic value, then it is applied to the +goal. If this application fails, then another combination of +hypotheses is tried with the same proof context pattern. If there is +no other combination of hypotheses then the second proof context +pattern is tried and so on. If the next to last proof context pattern +fails then {\tacexpr}$_{n+1}$ is evaluated to $v_{n+1}$ and $v_{n+1}$ +is applied. + +\ErrMsg \errindex{No matching clauses for match goal} + + No goal pattern can be used and, in particular, there is no {\tt + \_} goal pattern. + +\medskip + +It is important to know that each hypothesis of the goal can be +matched by at most one hypothesis pattern. The order of matching is +the following: hypothesis patterns are examined from the right to the +left (i.e. $hyp_{i,m_i}$ before $hyp_{i,1}$). For each hypothesis +pattern, the goal hypothesis are matched in order (fresher hypothesis +first), but it possible to reverse this order (older first) with +the {\tt match reverse goal with} variant. + +\subsubsection{Filling a term context} +\index{context!in expression} + +The following expression is not a tactic in the sense that it does not +produce subgoals but generates a term to be used in tactic +expressions: +\begin{quote} +{\tt context} {\ident} {\tt [} {\tacexpr} {\tt ]} +\end{quote} +{\ident} must denote a context variable bound by a {\tt context} +pattern of a {\tt match} expression. This expression evaluates +replaces the hole of the value of {\ident} by the value of +{\tacexpr}. + +\ErrMsg \errindex{not a context variable} + + +\subsubsection{Generating fresh hypothesis names} +\index{Ltac!fresh} +\index{fresh!in Ltac} + +Tactics sometimes have to generate new names for hypothesis. Letting +the system decide a name with the {\tt intro} tactic is not so good +since it is very awkward to retrieve the name the system gave. + +As before, the following expression returns a term: +\begin{quote} +{\tt fresh} {\qstring} +\end{quote} +It evaluates to an identifier unbound in the goal, which is obtained +by padding {\qstring} with a number if necessary. If no name is given, +the prefix is {\tt H}. + +\subsubsection{{\tt type of} {\term}} +%\tacindex{type of} +\index{Ltac!type of} +\index{type of!in Ltac} + +This tactic computes the type of {\term}. + +\subsubsection{Computing in a constr} +\index{Ltac!eval} +\index{eval!in Ltac} + +Evaluation of a term can be performed with: +\begin{quote} +{\tt eval} {\nterm{redexpr}} {\tt in} {\term} +\end{quote} +where \nterm{redexpr} is a reduction tactic among {\tt red}, {\tt +hnf}, {\tt compute}, {\tt simpl}, {\tt cbv}, {\tt lazy}, {\tt unfold}, +{\tt fold}, {\tt pattern}. + + +\subsubsection{Accessing tactic decomposition} +\tacindex{info} +\index{Tacticals!info@{\tt info}} + +Tactical ``{\tt info} {\tacexpr}'' is not really a tactical. For +elementary tactics, this is equivalent to \tacexpr. For complex tactic +like \texttt{auto}, it displays the operations performed by the +tactic. + +\subsubsection{Proving a subgoal as a separate lemma} +\tacindex{abstract} +\index{Tacticals!abstract@{\tt abstract}} + +From the outside ``\texttt{abstract \tacexpr}'' is the same as +{\tt solve \tacexpr}. Internally it saves an auxiliary lemma called +{\ident}\texttt{\_subproof}\textit{n} where {\ident} is the name of the +current goal and \textit{n} is chosen so that this is a fresh name. + +This tactical is useful with tactics such as \texttt{omega} or +\texttt{discriminate} that generate huge proof terms. With that tool +the user can avoid the explosion at time of the \texttt{Save} command +without having to cut manually the proof in smaller lemmas. + +\begin{Variants} +\item \texttt{abstract {\tacexpr} using {\ident}}.\\ + Give explicitly the name of the auxiliary lemma. +\end{Variants} + +\ErrMsg \errindex{Proof is not complete} + +\section{Tactic toplevel definitions} +\comindex{Ltac} + +Basically, tactics toplevel definitions are made as follows: +%{\tt Tactic Definition} {\ident} {\tt :=} {\tacexpr}\\ +% +%{\tacexpr} is evaluated to $v$ and $v$ is associated to {\ident}. Next, every +%script is evaluated by substituting $v$ to {\ident}. +% +%We can define functional definitions by:\\ +\begin{quote} +{\tt Ltac} {\ident} {\ident}$_1$ ... {\ident}$_n$ {\tt :=} +{\tacexpr} +\end{quote} +This defines a new tactic that can be used in any tactic script or new +tactic toplevel definition. + +\Rem The preceding definition can equivalently be written: +\begin{quote} +{\tt Ltac} {\ident} {\tt := fun} {\ident}$_1$ ... {\ident}$_n$ +{\tt =>} {\tacexpr} +\end{quote} +Recursive and mutual recursive function definitions are also +possible with the syntax: +\begin{quote} +{\tt Ltac} {\ident}$_1$ {\ident}$_{1,1}$ ... +{\ident}$_{1,m_1}$~~{\tt :=} {\tacexpr}$_1$\\ +{\tt with} {\ident}$_2$ {\ident}$_{2,1}$ ... {\ident}$_{2,m_2}$~~{\tt :=} +{\tacexpr}$_2$\\ +...\\ +{\tt with} {\ident}$_n$ {\ident}$_{n,1}$ ... {\ident}$_{n,m_n}$~~{\tt :=} +{\tacexpr}$_n$ +\end{quote} + +%This definition bloc is a set of definitions (use of +%the same previous syntactical sugar) and the other scripts are evaluated as +%usual except that the substitutions are lazily carried out (when an identifier +%to be evaluated is the name of a recursive definition). + +\endinput + + +\subsection{Permutation on closed lists} + +\begin{figure}[b] +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example*} +Require Import List. +Section Sort. +Variable A : Set. +Inductive permut : list A -> list A -> Prop := + | permut_refl : forall l, permut l l + | permut_cons : + forall a l0 l1, permut l0 l1 -> permut (a :: l0) (a :: l1) + | permut_append : forall a l, permut (a :: l) (l ++ a :: nil) + | permut_trans : + forall l0 l1 l2, permut l0 l1 -> permut l1 l2 -> permut l0 l2. +End Sort. +\end{coq_example*} +\end{center} +\caption{Definition of the permutation predicate} +\label{permutpred} +\end{figure} + + +Another more complex example is the problem of permutation on closed +lists. The aim is to show that a closed list is a permutation of +another one. First, we define the permutation predicate as shown on +Figure~\ref{permutpred}. + +\begin{figure}[p] +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example} +Ltac Permut n := + match goal with + | |- (permut _ ?l ?l) => apply permut_refl + | |- (permut _ (?a :: ?l1) (?a :: ?l2)) => + let newn := eval compute in (length l1) in + (apply permut_cons; Permut newn) + | |- (permut ?A (?a :: ?l1) ?l2) => + match eval compute in n with + | 1 => fail + | _ => + let l1' := constr:(l1 ++ a :: nil) in + (apply (permut_trans A (a :: l1) l1' l2); + [ apply permut_append | compute; Permut (pred n) ]) + end + end. +Ltac PermutProve := + match goal with + | |- (permut _ ?l1 ?l2) => + match eval compute in (length l1 = length l2) with + | (?n = ?n) => Permut n + end + end. +\end{coq_example} +\end{minipage}} +\end{center} +\caption{Permutation tactic} +\label{permutltac} +\end{figure} + +\begin{figure}[p] +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example*} +Lemma permut_ex1 : + permut nat (1 :: 2 :: 3 :: nil) (3 :: 2 :: 1 :: nil). +Proof. +PermutProve. +Qed. + +Lemma permut_ex2 : + permut nat + (0 :: 1 :: 2 :: 3 :: 4 :: 5 :: 6 :: 7 :: 8 :: 9 :: nil) + (0 :: 2 :: 4 :: 6 :: 8 :: 9 :: 7 :: 5 :: 3 :: 1 :: nil). +Proof. +PermutProve. +Qed. +\end{coq_example*} +\end{minipage}} +\end{center} +\caption{Examples of {\tt PermutProve} use} +\label{permutlem} +\end{figure} + +Next, we can write naturally the tactic and the result can be seen on +Figure~\ref{permutltac}. We can notice that we use two toplevel +definitions {\tt PermutProve} and {\tt Permut}. The function to be +called is {\tt PermutProve} which computes the lengths of the two +lists and calls {\tt Permut} with the length if the two lists have the +same length. {\tt Permut} works as expected. If the two lists are +equal, it concludes. Otherwise, if the lists have identical first +elements, it applies {\tt Permut} on the tail of the lists. Finally, +if the lists have different first elements, it puts the first element +of one of the lists (here the second one which appears in the {\tt + permut} predicate) at the end if that is possible, i.e., if the new +first element has been at this place previously. To verify that all +rotations have been done for a list, we use the length of the list as +an argument for {\tt Permut} and this length is decremented for each +rotation down to, but not including, 1 because for a list of length +$n$, we can make exactly $n-1$ rotations to generate at most $n$ +distinct lists. Here, it must be noticed that we use the natural +numbers of {\Coq} for the rotation counter. On Figure~\ref{ltac}, we +can see that it is possible to use usual natural numbers but they are +only used as arguments for primitive tactics and they cannot be +handled, in particular, we cannot make computations with them. So, a +natural choice is to use {\Coq} data structures so that {\Coq} makes +the computations (reductions) by {\tt eval compute in} and we can get +the terms back by {\tt match}. + +With {\tt PermutProve}, we can now prove lemmas such those shown on +Figure~\ref{permutlem}. + + +\subsection{Deciding intuitionistic propositional logic} + +\begin{figure}[tbp] +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example} +Ltac Axioms := + match goal with + | |- True => trivial + | _:False |- _ => elimtype False; assumption + | _:?A |- ?A => auto + end. +Ltac DSimplif := + repeat + (intros; + match goal with + | id:(~ _) |- _ => red in id + | id:(_ /\ _) |- _ => + elim id; do 2 intro; clear id + | id:(_ \/ _) |- _ => + elim id; intro; clear id + | id:(?A /\ ?B -> ?C) |- _ => + cut (A -> B -> C); + [ intro | intros; apply id; split; assumption ] + | id:(?A \/ ?B -> ?C) |- _ => + cut (B -> C); + [ cut (A -> C); + [ intros; clear id + | intro; apply id; left; assumption ] + | intro; apply id; right; assumption ] + | id0:(?A -> ?B),id1:?A |- _ => + cut B; [ intro; clear id0 | apply id0; assumption ] + | |- (_ /\ _) => split + | |- (~ _) => red + end). +\end{coq_example} +\end{minipage}} +\end{center} +\caption{Deciding intuitionistic propositions (1)} +\label{tautoltaca} +\end{figure} + +\begin{figure} +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example} +Ltac TautoProp := + DSimplif; + Axioms || + match goal with + | id:((?A -> ?B) -> ?C) |- _ => + cut (B -> C); + [ intro; cut (A -> B); + [ intro; cut C; + [ intro; clear id | apply id; assumption ] + | clear id ] + | intro; apply id; intro; assumption ]; TautoProp + | id:(~ ?A -> ?B) |- _ => + cut (False -> B); + [ intro; cut (A -> False); + [ intro; cut B; + [ intro; clear id | apply id; assumption ] + | clear id ] + | intro; apply id; red; intro; assumption ]; TautoProp + | |- (_ \/ _) => (left; TautoProp) || (right; TautoProp) + end. +\end{coq_example} +\end{minipage}} +\end{center} +\caption{Deciding intuitionistic propositions (2)} +\label{tautoltacb} +\end{figure} + +The pattern matching on goals allows a complete and so a powerful +backtracking when returning tactic values. An interesting application +is the problem of deciding intuitionistic propositional logic. +Considering the contraction-free sequent calculi {\tt LJT*} of +Roy~Dyckhoff (\cite{Dyc92}), it is quite natural to code such a tactic +using the tactic language. On Figure~\ref{tautoltaca}, the tactic {\tt + Axioms} tries to conclude using usual axioms. The {\tt DSimplif} +tactic applies all the reversible rules of Dyckhoff's system. +Finally, on Figure~\ref{tautoltacb}, the {\tt TautoProp} tactic (the +main tactic to be called) simplifies with {\tt DSimplif}, tries to +conclude with {\tt Axioms} and tries several paths using the +backtracking rules (one of the four Dyckhoff's rules for the left +implication to get rid of the contraction and the right or). + +\begin{figure}[tb] +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example*} +Lemma tauto_ex1 : forall A B:Prop, A /\ B -> A \/ B. +Proof. +TautoProp. +Qed. + +Lemma tauto_ex2 : + forall A B:Prop, (~ ~ B -> B) -> (A -> B) -> ~ ~ A -> B. +Proof. +TautoProp. +Qed. +\end{coq_example*} +\end{minipage}} +\end{center} +\caption{Proofs of tautologies with {\tt TautoProp}} +\label{tautolem} +\end{figure} + +For example, with {\tt TautoProp}, we can prove tautologies like those of +Figure~\ref{tautolem}. + + +\subsection{Deciding type isomorphisms} + +A more tricky problem is to decide equalities between types and modulo +isomorphisms. Here, we choose to use the isomorphisms of the simply typed +$\lb{}$-calculus with Cartesian product and $unit$ type (see, for example, +\cite{RC95}). The axioms of this $\lb{}$-calculus are given by +Figure~\ref{isosax}. + +\begin{figure} +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_eval} +Reset Initial. +\end{coq_eval} +\begin{coq_example*} +Open Scope type_scope. +Section Iso_axioms. +Variables A B C : Set. +Axiom Com : A * B = B * A. +Axiom Ass : A * (B * C) = A * B * C. +Axiom Cur : (A * B -> C) = (A -> B -> C). +Axiom Dis : (A -> B * C) = (A -> B) * (A -> C). +Axiom P_unit : A * unit = A. +Axiom AR_unit : (A -> unit) = unit. +Axiom AL_unit : (unit -> A) = A. +Lemma Cons : B = C -> A * B = A * C. +Proof. +intro Heq; rewrite Heq; apply refl_equal. +Qed. +End Iso_axioms. +\end{coq_example*} +\end{minipage}} +\end{center} +\caption{Type isomorphism axioms} +\label{isosax} +\end{figure} + +The tactic to judge equalities modulo this axiomatization can be written as +shown on Figures~\ref{isosltac1} and~\ref{isosltac2}. The algorithm is quite +simple. Types are reduced using axioms that can be oriented (this done by {\tt +MainSimplif}). The normal forms are sequences of Cartesian +products without Cartesian product in the left component. These normal forms +are then compared modulo permutation of the components (this is done by {\tt +CompareStruct}). The main tactic to be called and realizing this algorithm is +{\tt IsoProve}. + +\begin{figure} +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example} +Ltac DSimplif trm := + match trm with + | (?A * ?B * ?C) => + rewrite <- (Ass A B C); try MainSimplif + | (?A * ?B -> ?C) => + rewrite (Cur A B C); try MainSimplif + | (?A -> ?B * ?C) => + rewrite (Dis A B C); try MainSimplif + | (?A * unit) => + rewrite (P_unit A); try MainSimplif + | (unit * ?B) => + rewrite (Com unit B); try MainSimplif + | (?A -> unit) => + rewrite (AR_unit A); try MainSimplif + | (unit -> ?B) => + rewrite (AL_unit B); try MainSimplif + | (?A * ?B) => + (DSimplif A; try MainSimplif) || (DSimplif B; try MainSimplif) + | (?A -> ?B) => + (DSimplif A; try MainSimplif) || (DSimplif B; try MainSimplif) + end + with MainSimplif := + match goal with + | |- (?A = ?B) => try DSimplif A; try DSimplif B + end. +Ltac Length trm := + match trm with + | (_ * ?B) => let succ := Length B in constr:(S succ) + | _ => constr:1 + end. +Ltac assoc := repeat rewrite <- Ass. +\end{coq_example} +\end{minipage}} +\end{center} +\caption{Type isomorphism tactic (1)} +\label{isosltac1} +\end{figure} + +\begin{figure} +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example} +Ltac DoCompare n := + match goal with + | [ |- (?A = ?A) ] => apply refl_equal + | [ |- (?A * ?B = ?A * ?C) ] => + apply Cons; let newn := Length B in DoCompare newn + | [ |- (?A * ?B = ?C) ] => + match eval compute in n with + | 1 => fail + | _ => + pattern (A * B) at 1; rewrite Com; assoc; DoCompare (pred n) + end + end. +Ltac CompareStruct := + match goal with + | [ |- (?A = ?B) ] => + let l1 := Length A + with l2 := Length B in + match eval compute in (l1 = l2) with + | (?n = ?n) => DoCompare n + end + end. +Ltac IsoProve := MainSimplif; CompareStruct. +\end{coq_example} +\end{minipage}} +\end{center} +\caption{Type isomorphism tactic (2)} +\label{isosltac2} +\end{figure} + +Figure~\ref{isoslem} gives examples of what can be solved by {\tt IsoProve}. + +\begin{figure} +\begin{center} +\fbox{\begin{minipage}{0.95\textwidth} +\begin{coq_example*} +Lemma isos_ex1 : + forall A B:Set, A * unit * B = B * (unit * A). +Proof. +intros; IsoProve. +Qed. + +Lemma isos_ex2 : + forall A B C:Set, + (A * unit -> B * (C * unit)) = + (A * unit -> (C -> unit) * C) * (unit -> A -> B). +Proof. +intros; IsoProve. +Qed. +\end{coq_example*} +\end{minipage}} +\end{center} +\caption{Type equalities solved by {\tt IsoProve}} +\label{isoslem} +\end{figure} + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-mod.tex b/doc/refman/RefMan-mod.tex new file mode 100644 index 0000000000..9f6f2abce1 --- /dev/null +++ b/doc/refman/RefMan-mod.tex @@ -0,0 +1,396 @@ +\section{Module system +\index{Modules} +\label{section:Modules}} + +The module system provides a way of packaging related elements +together, as well as a mean of massive abstraction. + +\begin{figure}[t] +\begin{centerframe} +\begin{tabular}{rcl} +{\modtype} & ::= & {\ident} \\ + & $|$ & {\modtype} \texttt{ with Definition }{\ident} := {\term} \\ + & $|$ & {\modtype} \texttt{ with Module }{\ident} := {\qualid} \\ + &&\\ + +{\onemodbinding} & ::= & {\tt ( \nelist{\ident}{} : {\modtype} )}\\ + &&\\ + +{\modbindings} & ::= & \nelist{\onemodbinding}{}\\ + &&\\ + +{\modexpr} & ::= & \nelist{\qualid}{} +\end{tabular} +\end{centerframe} +\caption{Syntax of modules} +\end{figure} + +\subsection{\tt Module {\ident} +\comindex{Module}} + +This command is used to start an interactive module named {\ident}. + +\begin{Variants} + +\item{\tt Module {\ident} {\modbindings}} + + Starts an interactive functor with parameters given by {\modbindings}. + +\item{\tt Module {\ident} \verb.:. {\modtype}} + + Starts an interactive module specifying its module type. + +\item{\tt Module {\ident} {\modbindings} \verb.:. {\modtype}} + + Starts an interactive functor with parameters given by + {\modbindings}, and output module type {\modtype}. + +\item{\tt Module {\ident} \verb.<:. {\modtype}} + + Starts an interactive module satisfying {\modtype}. + +\item{\tt Module {\ident} {\modbindings} \verb.<:. {\modtype}} + + Starts an interactive functor with parameters given by + {\modbindings}. The output module type is verified against the + module type {\modtype}. + +\end{Variants} + +\subsection{\tt End {\ident} +\comindex{End}} + +This command closes the interactive module {\ident}. If the module type +was given the content of the module is matched against it and an error +is signaled if the matching fails. If the module is basic (is not a +functor) its components (constants, inductive types, submodules etc) are +now available through the dot notation. + +\begin{ErrMsgs} +\item \errindex{No such label {\ident}} +\item \errindex{Signature components for label {\ident} do not match} +\item \errindex{This is not the last opened module} +\end{ErrMsgs} + + +\subsection{\tt Module {\ident} := {\modexpr} +\comindex{Module}} + +This command defines the module identifier {\ident} to be equal to +{\modexpr}. + +\begin{Variants} +\item{\tt Module {\ident} {\modbindings} := {\modexpr}} + + Defines a functor with parameters given by {\modbindings} and body {\modexpr}. + +% Particular cases of the next 2 items +%\item{\tt Module {\ident} \verb.:. {\modtype} := {\modexpr}} +% +% Defines a module with body {\modexpr} and interface {\modtype}. +%\item{\tt Module {\ident} \verb.<:. {\modtype} := {\modexpr}} +% +% Defines a module with body {\modexpr}, satisfying {\modtype}. + +\item{\tt Module {\ident} {\modbindings} \verb.:. {\modtype} := + {\modexpr}} + + Defines a functor with parameters given by {\modbindings} (possibly none), + and output module type {\modtype}, with body {\modexpr}. + +\item{\tt Module {\ident} {\modbindings} \verb.<:. {\modtype} := + {\modexpr}} + + Defines a functor with parameters given by {\modbindings} (possibly none) + with body {\modexpr}. The body is checked against {\modtype}. + +\end{Variants} + +\subsection{\tt Module Type {\ident} +\comindex{Module Type}} + +This command is used to start an interactive module type {\ident}. + +\begin{Variants} + +\item{\tt Module Type {\ident} {\modbindings}} + + Starts an interactive functor type with parameters given by {\modbindings}. + +\end{Variants} + +\subsection{\tt End {\ident} +\comindex{End}} + +This command closes the interactive module type {\ident}. + +\begin{ErrMsgs} +\item \errindex{This is not the last opened module type} +\end{ErrMsgs} + +\subsection{\tt Module Type {\ident} := {\modtype}} + +Defines a module type {\ident} equal to {\modtype}. + +\begin{Variants} +\item {\tt Module Type {\ident} {\modbindings} := {\modtype}} + + Defines a functor type {\ident} specifying functors taking arguments + {\modbindings} and returning {\modtype}. +\end{Variants} + +\subsection{\tt Declare Module {\ident}} + +Starts an interactive module declaration. This command is available +only in module types. + +\begin{Variants} + +\item{\tt Declare Module {\ident} {\modbindings}} + + Starts an interactive declaration of a functor with parameters given + by {\modbindings}. + +% Particular case of the next item +%\item{\tt Declare Module {\ident} \verb.<:. {\modtype}} +% +% Starts an interactive declaration of a module satisfying {\modtype}. + +\item{\tt Declare Module {\ident} {\modbindings} \verb.<:. {\modtype}} + + Starts an interactive declaration of a functor with parameters given + by {\modbindings} (possibly none). The declared output module type is + verified against the module type {\modtype}. + +\end{Variants} + +\subsection{\tt End {\ident}} + +This command closes the interactive declaration of module {\ident}. + +\subsection{\tt Declare Module {\ident} : {\modtype}} + +Declares a module of {\ident} of type {\modtype}. This command is available +only in module types. + +\begin{Variants} + +\item{\tt Declare Module {\ident} {\modbindings} \verb.:. {\modtype}} + + Declares a functor with parameters {\modbindings} and output module + type {\modtype}. + +\item{\tt Declare Module {\ident} := {\qualid}} + + Declares a module equal to the module {\qualid}. + +\item{\tt Declare Module {\ident} \verb.<:. {\modtype} := {\qualid}} + + Declares a module equal to the module {\qualid}, verifying that the + module type of the latter is a subtype of {\modtype}. + +\end{Variants} + + +\subsubsection{Example} + +Let us define a simple module. +\begin{coq_example} +Module M. + Definition T := nat. + Definition x := 0. + Definition y : bool. + exact true. + Defined. +End M. +\end{coq_example} +Inside a module one can define constants, prove theorems and do any +other things that can be done in the toplevel. Components of a closed +module can be accessed using the dot notation: +\begin{coq_example} +Print M.x. +\end{coq_example} +A simple module type: +\begin{coq_example} +Module Type SIG. + Parameter T : Set. + Parameter x : T. +End SIG. +\end{coq_example} +Inside a module type the proof editing mode is not available. +Consequently commands like \texttt{Definition}\ without body, +\texttt{Lemma}, \texttt{Theorem} are not allowed. In order to declare +constants, use \texttt{Axiom} and \texttt{Parameter}. + +Now we can create a new module from \texttt{M}, giving it a less +precise specification: the \texttt{y} component is dropped as well +as the body of \texttt{x}. + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is not correct and should produce **********) +(***************** Error: N.y not a defined object *******************) +\end{coq_eval} +\begin{coq_example} +Module N : SIG with Definition T := nat := M. +Print N.T. +Print N.x. +Print N.y. +\end{coq_example} +\begin{coq_eval} +Reset N. +\end{coq_eval} + +\noindent +The definition of \texttt{N} using the module type expression +\texttt{SIG with Definition T:=nat} is equivalent to the following +one: + +\begin{coq_example*} +Module Type SIG'. + Definition T : Set := nat. + Parameter x : T. +End SIG'. +Module N : SIG' := M. +\end{coq_example*} +If we just want to be sure that the our implementation satisfies a +given module type without restricting the interface, we can use a +transparent constraint +\begin{coq_example} +Module P <: SIG := M. +Print P.y. +\end{coq_example} +Now let us create a functor, i.e. a parametric module +\begin{coq_example} +Module Two (X Y: SIG). +\end{coq_example} +\begin{coq_example*} + Definition T := (X.T * Y.T)%type. + Definition x := (X.x, Y.x). +\end{coq_example*} +\begin{coq_example} +End Two. +\end{coq_example} +and apply it to our modules and do some computations +\begin{coq_example} +Module Q := Two M N. +Eval compute in (fst Q.x + snd Q.x). +\end{coq_example} +In the end, let us define a module type with two sub-modules, sharing +some of the fields and give one of its possible implementations: +\begin{coq_example} +Module Type SIG2. + Declare Module M1 : SIG. + Declare Module M2 <: SIG. + Definition T := M1.T. + Parameter x : T. + End M2. +End SIG2. +\end{coq_example} +\begin{coq_example*} +Module Mod <: SIG2. + Module M1. + Definition T := nat. + Definition x := 1. + End M1. + Module M2 := M. +\end{coq_example*} +\begin{coq_example} +End Mod. +\end{coq_example} +Notice that \texttt{M} is a correct body for the component \texttt{M2} +since its \texttt{T} component is equal \texttt{nat} and hence +\texttt{M1.T} as specified. +\begin{coq_eval} +Reset Initial. +\end{coq_eval} + +\begin{Remarks} +\item Modules and module types can be nested components of each other. +\item When a module declaration is started inside a module type, + the proof editing mode is still unavailable. +\item One can have sections inside a module or a module type, but + not a module or a module type inside a section. +\item Commands like \texttt{Hint} or \texttt{Notation} can + also appear inside modules and module types. Note that in case of a + module definition like: + + \medskip + \noindent + {\tt Module N : SIG := M.} + \medskip + + or + + \medskip + {\tt Module N : SIG.\\ + \ \ \dots\\ + End N.} + \medskip + + hints and the like valid for \texttt{N} are not those defined in + \texttt{M} (or the module body) but the ones defined in + \texttt{SIG}. + +\end{Remarks} + +\subsection{Import {\qualid} +\comindex{Import} +\label{Import}} + +If {\qualid} denotes a valid basic module (i.e. its module type is a +signature), makes its components available by their short names. + +Example: + +\begin{coq_example} +Module Mod. +\end{coq_example} +\begin{coq_example} + Definition T:=nat. + Check T. +\end{coq_example} +\begin{coq_example} +End Mod. +Check Mod.T. +Check T. (* Incorrect ! *) +Import Mod. +Check T. (* Now correct *) +\end{coq_example} +\begin{coq_eval} +Reset Mod. +\end{coq_eval} + + +\begin{Variants} +\item{\tt Export {\qualid}}\comindex{Export} + + When the module containing the command {\tt Export {\qualid}} is + imported, {\qualid} is imported as well. +\end{Variants} + +\begin{ErrMsgs} + \item \errindexbis{{\qualid} is not a module}{is not a module} +% this error is impossible in the import command +% \item \errindex{Cannot mask the absolute name {\qualid} !} +\end{ErrMsgs} + +\begin{Warnings} + \item Warning: Trying to mask the absolute name {\qualid} ! +\end{Warnings} + +\subsection{\tt Print Module {\ident} +\comindex{Print Module}} + +Prints the module type and (optionally) the body of the module {\ident}. + +\subsection{\tt Print Module Type {\ident} +\comindex{Print Module Type}} + +Prints the module type corresponding to {\ident}. + + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-modr.tex b/doc/refman/RefMan-modr.tex new file mode 100644 index 0000000000..e08c7c401a --- /dev/null +++ b/doc/refman/RefMan-modr.tex @@ -0,0 +1,586 @@ +\chapter{The Module System} +\label{chapter:Modules} + +The module system extends the Calculus of Inductive Constructions +providing a convenient way to structure large developments as well as +a mean of massive abstraction. +%It is described in details in Judicael's thesis and Jacek's thesis + +\section{Modules and module types} + +\paragraph{Access path.} It is denoted by $p$, it can be either a module +variable $X$ or, if $p'$ is an access path and $id$ an identifier, then +$p'.id$ is an access path. + +\paragraph{Structure element.} It is denoted by \Impl\ and is either a +definition of a constant, an assumption, a definition of an inductive +or a definition of a module or a module type abbreviation. + +\paragraph{Module expression.} It is denoted by $M$ and can be: +\begin{itemize} +\item an access path $p$ +\item a structure $\struct{\nelist{\Impl}{;}}$ +\item a functor $\functor{X}{T}{M'}$, where $X$ is a module variable, + $T$ is a module type and $M'$ is a module expression +\item an application of access paths $p' p''$ +\end{itemize} + +\paragraph{Signature element.} It is denoted by \Spec, it is a +specification of a constant, an assumption, an inductive, a module or +a module type abbreviation. + +\paragraph{Module type,} denoted by $T$ can be: +\begin{itemize} +\item a module type name +\item an access path $p$ +\item a signature $\sig{\nelist{\Spec}{;}}$ +\item a functor type $\funsig{X}{T'}{T''}$, where $T'$ and $T''$ are + module types +\end{itemize} + +\paragraph{Module definition,} written $\Mod{X}{T}{M}$ can be a +structure element. It consists of a module variable $X$, a module type +$T$ and a module expression $M$. + +\paragraph{Module specification,} written $\ModS{X}{T}$ or +$\ModSEq{X}{T}{p}$ can be a signature element or a part of an +environment. It consists of a module variable $X$, a module type $T$ +and, optionally, a module path $p$. + +\paragraph{Module type abbreviation,} written $\ModType{S}{T}$, where +$S$ is a module type name and $T$ is a module type. + + +\section{Typing Modules} + +In order to introduce the typing system we first slightly extend +the syntactic class of terms and environments given in +section~\ref{Terms}. The environments, apart from definitions of +constants and inductive types now also hold any other signature elements. +Terms, apart from variables, constants and complex terms, +include also access paths. + +We also need additional typing judgments: +\begin{itemize} +\item \WFT{E}{T}, denoting that a module type $T$ is well-formed, + +\item \WTM{E}{M}{T}, denoting that a module expression $M$ has type $T$ in +environment $E$. + +\item \WTM{E}{\Impl}{\Spec}, denoting that an implementation $\Impl$ + verifies a specification $\Spec$ + +\item \WS{E}{T_1}{T_2}, denoting that a module type $T_1$ is a subtype of a +module type $T_2$. + +\item \WS{E}{\Spec_1}{\Spec_2}, denoting that a specification + $\Spec_1$ is more precise that a specification $\Spec_2$. +\end{itemize} +The rules for forming module types are the following: +\begin{description} +\item[WF-SIG] +\inference{% + \frac{ + \WF{E;E'}{} + }{%%%%%%%%%%%%%%%%%%%%% + \WFT{E}{\sig{E'}} + } +} +\item[WF-FUN] +\inference{% + \frac{ + \WFT{E;\ModS{X}{T}}{T'} + }{%%%%%%%%%%%%%%%%%%%%%%%%%% + \WFT{E}{\funsig{X}{T}{T'}} + } +} +\end{description} +Rules for typing module expressions: +\begin{description} +\item[MT-STRUCT] +\inference{% + \frac{ + \begin{array}{c} + \WFT{E}{\sig{\Spec_1;\dots;\Spec_n}}\\ + \WTM{E;\Spec_1;\dots;\Spec_{i-1}}{\Impl_i}{\Spec_i} + \textrm{ \ \ for } i=1\dots n + \end{array} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WTM{E}{\struct{\Impl_1;\dots;\Impl_n}}{\sig{\Spec_1;\dots;\Spec_n}} + } +} +\item[MT-FUN] +\inference{% + \frac{ + \WTM{E;\ModS{X}{T}}{M}{T'} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WTM{E}{\functor{X}{T}{M}}{\funsig{X}{T}{T'}} + } +} +\item[MT-APP] +\inference{% + \frac{ + \begin{array}{c} + \WTM{E}{p}{\funsig{X_1}{T_1}{\!\dots\funsig{X_n}{T_n}{T'}}}\\ + \WTM{E}{p_i}{T_i\{X_1/p_1\dots X_{i-1}/p_{i-1}\}} + \textrm{ \ \ for } i=1\dots n + \end{array} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WTM{E}{p\; p_1 \dots p_n}{T'\{X_1/p_1\dots X_n/p_n\}} + } +} +\item[MT-SUB] +\inference{% + \frac{ + \WTM{E}{M}{T}~~~~~~~~~~~~\WS{E}{T}{T'} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WTM{E}{M}{T'} + } +} +\item[MT-STR] +\inference{% + \frac{ + \WTM{E}{p}{T} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WTM{E}{p}{T/p} + } +} +\end{description} +The last rule, called strengthening is used to make all module fields +manifestly equal to themselves. The notation $T/p$ has the following +meaning: +\begin{itemize} +\item if $T=\sig{\Spec_1;\dots;\Spec_n}$ then + $T/p=\sig{\Spec_1/p;\dots;\Spec_n/p}$ where $\Spec/p$ is defined as + follows: + \begin{itemize} + \item $\Def{}{c}{U}{t}/p ~=~ \Def{}{c}{U}{t}$ + \item $\Assum{}{c}{U}/p ~=~ \Def{}{c}{p.c}{U}$ + \item $\ModS{X}{T}/p ~=~ \ModSEq{X}{T/p.X}{p.X}$ + \item $\ModSEq{X}{T}{p'}/p ~=~ \ModSEq{X}{T/p}{p'}$ + \item $\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I}/p ~=~ \Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}$ + \item $\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p'}/p ~=~ \Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p'}$ + \end{itemize} +\item if $T=\funsig{X}{T'}{T''}$ then $T/p=T$ +\item if $T$ is an access path or a module type name, then we have to + unfold its definition and proceed according to the rules above. +\end{itemize} +The notation $\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}$ denotes an +inductive definition that is definitionally equal to the inductive +definition in the module denoted by the path $p$. All rules which have +$\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I}$ as premises are also valid for +$\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}$. We give the formation rule +for $\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}$ below as well as +the equality rules on inductive types and constructors. + +The module subtyping rules: +\begin{description} +\item[MSUB-SIG] +\inference{% + \frac{ + \begin{array}{c} + \WS{E;\Spec_1;\dots;\Spec_n}{\Spec_{\sigma(i)}}{\Spec'_i} + \textrm{ \ for } i=1..m \\ + \sigma : \{1\dots m\} \ra \{1\dots n\} \textrm{ \ injective} + \end{array} + }{ + \WS{E}{\sig{\Spec_1;\dots;\Spec_n}}{\sig{\Spec'_1;\dots;\Spec'_m}} + } +} +\item[MSUB-FUN] +\inference{% T_1 -> T_2 <: T_1' -> T_2' + \frac{ + \WS{E}{T_1'}{T_1}~~~~~~~~~~\WS{E;\ModS{X}{T_1'}}{T_2}{T_2'} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WS{E}{\funsig{X}{T_1}{T_2}}{\funsig{X}{T_1'}{T_2'}} + } +} +% these are derived rules +% \item[MSUB-EQ] +% \inference{% +% \frac{ +% \WS{E}{T_1}{T_2}~~~~~~~~~~\WTERED{}{T_1}{=}{T_1'}~~~~~~~~~~\WTERED{}{T_2}{=}{T_2'} +% }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +% \WS{E}{T_1'}{T_2'} +% } +% } +% \item[MSUB-REFL] +% \inference{% +% \frac{ +% \WFT{E}{T} +% }{ +% \WS{E}{T}{T} +% } +% } +\end{description} +Specification subtyping rules: +\begin{description} +\item[ASSUM-ASSUM] +\inference{% + \frac{ + \WTELECONV{}{U_1}{U_2} + }{ + \WSE{\Assum{}{c}{U_1}}{\Assum{}{c}{U_2}} + } +} +\item[DEF-ASSUM] +\inference{% + \frac{ + \WTELECONV{}{U_1}{U_2} + }{ + \WSE{\Def{}{c}{t}{U_1}}{\Assum{}{c}{U_2}} + } +} +\item[ASSUM-DEF] +\inference{% + \frac{ + \WTELECONV{}{U_1}{U_2}~~~~~~~~\WTECONV{}{c}{t_2} + }{ + \WSE{\Assum{}{c}{U_1}}{\Def{}{c}{t_2}{U_2}} + } +} +\item[DEF-DEF] +\inference{% + \frac{ + \WTELECONV{}{U_1}{U_2}~~~~~~~~\WTECONV{}{t_1}{t_2} + }{ + \WSE{\Def{}{c}{t_1}{U_1}}{\Def{}{c}{t_2}{U_2}} + } +} +\item[IND-IND] +\inference{% + \frac{ + \WTECONV{}{\Gamma_P}{\Gamma_P'}% + ~~~~~~~~\WTECONV{\Gamma_P}{\Gamma_C}{\Gamma_C'}% + ~~~~~~~~\WTECONV{\Gamma_P;\Gamma_C}{\Gamma_I}{\Gamma_I'}% + }{ + \WSE{\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I}}% + {\Ind{}{\Gamma_P'}{\Gamma_C'}{\Gamma_I'}} + } +} +\item[INDP-IND] +\inference{% + \frac{ + \WTECONV{}{\Gamma_P}{\Gamma_P'}% + ~~~~~~~~\WTECONV{\Gamma_P}{\Gamma_C}{\Gamma_C'}% + ~~~~~~~~\WTECONV{\Gamma_P;\Gamma_C}{\Gamma_I}{\Gamma_I'}% + }{ + \WSE{\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}% + {\Ind{}{\Gamma_P'}{\Gamma_C'}{\Gamma_I'}} + } +} +\item[INDP-INDP] +\inference{% + \frac{ + \WTECONV{}{\Gamma_P}{\Gamma_P'}% + ~~~~~~\WTECONV{\Gamma_P}{\Gamma_C}{\Gamma_C'}% + ~~~~~~\WTECONV{\Gamma_P;\Gamma_C}{\Gamma_I}{\Gamma_I'}% + ~~~~~~\WTECONV{}{p}{p'} + }{ + \WSE{\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}% + {\Indp{}{\Gamma_P'}{\Gamma_C'}{\Gamma_I'}{p'}} + } +} +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +\item[MODS-MODS] +\inference{% + \frac{ + \WSE{T_1}{T_2} + }{ + \WSE{\ModS{m}{T_1}}{\ModS{m}{T_2}} + } +} +\item[MODEQ-MODS] +\inference{% + \frac{ + \WSE{T_1}{T_2} + }{ + \WSE{\ModSEq{m}{T_1}{p}}{\ModS{m}{T_2}} + } +} +\item[MODS-MODEQ] +\inference{% + \frac{ + \WSE{T_1}{T_2}~~~~~~~~\WTECONV{}{m}{p_2} + }{ + \WSE{\ModS{m}{T_1}}{\ModSEq{m}{T_2}{p_2}} + } +} +\item[MODEQ-MODEQ] +\inference{% + \frac{ + \WSE{T_1}{T_2}~~~~~~~~\WTECONV{}{p_1}{p_2} + }{ + \WSE{\ModSEq{m}{T_1}{p_1}}{\ModSEq{m}{T_2}{p_2}} + } +} +\item[MODTYPE-MODTYPE] +\inference{% + \frac{ + \WSE{T_1}{T_2}~~~~~~~~\WSE{T_2}{T_1} + }{ + \WSE{\ModType{S}{T_1}}{\ModType{S}{T_2}} + } +} +\end{description} +Verification of the specification +\begin{description} +\item[IMPL-SPEC] +\inference{% + \frac{ + \begin{array}{c} + \WF{E;\Spec}{}\\ + \Spec \textrm{\ is one of } {\sf Def, Assum, Ind, ModType} + \end{array} + }{ + \WTE{}{\Spec}{\Spec} + } +} +\item[MOD-MODS] +\inference{% + \frac{ + \WF{E;\ModS{m}{T}}{}~~~~~~~~\WTE{}{M}{T} + }{ + \WTE{}{\Mod{m}{T}{M}}{\ModS{m}{T}} + } +} +\item[MOD-MODEQ] +\inference{% + \frac{ + \WF{E;\ModSEq{m}{T}{p}}{}~~~~~~~~~~~\WTECONV{}{p}{p'} + }{ + \WTE{}{\Mod{m}{T}{p'}}{\ModSEq{m}{T}{p'}} + } +} +\end{description} +New environment formation rules +\begin{description} +\item[WF-MODS] +\inference{% + \frac{ + \WF{E}{}~~~~~~~~\WFT{E}{T} + }{ + \WF{E;\ModS{m}{T}}{} + } +} +\item[WF-MODEQ] +\inference{% + \frac{ + \WF{E}{}~~~~~~~~~~~\WTE{}{p}{T} + }{ + \WF{E,\ModSEq{m}{T}{p}}{} + } +} +\item[WF-MODTYPE] +\inference{% + \frac{ + \WF{E}{}~~~~~~~~~~~\WFT{E}{T} + }{ + \WF{E,\ModType{S}{T}}{} + } +} +\item[WF-IND] +\inference{% + \frac{ + \begin{array}{c} + \WF{E;\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I}}{}\\ + \WT{E}{}{p:\sig{\Spec_1;\dots;\Spec_n;\Ind{}{\Gamma_P'}{\Gamma_C'}{\Gamma_I'};\dots}}\\ + \WS{E}{\subst{\Ind{}{\Gamma_P'}{\Gamma_C'}{\Gamma_I'}}{p.l}{l}_{l + \in \lab{Spec_1;\dots;Spec_n}}}{\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I}} + \end{array} + }{%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + \WF{E;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}{} + } +} +\end{description} +Component access rules +\begin{description} +\item[ACC-TYPE] +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\Assum{}{c}{U};\dots}} + }{ + \WTEG{p.c}{\subst{U}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\\ +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\Def{}{c}{t}{U};\dots}} + }{ + \WTEG{p.c}{\subst{U}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\item[ACC-DELTA] +Notice that the following rule extends the delta rule defined in +section~\ref{delta} +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\Def{}{c}{t}{U};\dots}} + }{ + \WTEGRED{p.c}{\triangleright_\delta}{\subst{t}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\\ +In the rules below we assume $\Gamma_P$ is $[p_1:P_1;\ldots;p_r:P_r]$, + $\Gamma_I$ is $[I_1:A_1;\ldots;I_k:A_k]$, and $\Gamma_C$ is + $[c_1:C_1;\ldots;c_n:C_n]$ +\item[ACC-IND] +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;\Spec_i;\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I};\dots}} + }{ + \WTEG{p.I_j}{\subst{(p_1:P_1)\ldots(p_r:P_r)A_j}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;\Spec_i;\Ind{}{\Gamma_P}{\Gamma_C}{\Gamma_I};\dots}} + }{ + \WTEG{p.c_m}{\subst{(p_1:P_1)\ldots(p_r:P_r)\subst{C_m}{I_j}{(I_j~p_1\ldots + p_r)}_{j=1\ldots k}}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\item[ACC-INDP] +\inference{% + \frac{ + \WT{E}{}{p}{\sig{\Spec_1;\dots;\Spec_n;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p'};\dots}} + }{ + \WTRED{E}{}{p.I_i}{\triangleright_\delta}{p'.I_i} + } +} +\inference{% + \frac{ + \WT{E}{}{p}{\sig{\Spec_1;\dots;\Spec_n;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p'};\dots}} + }{ + \WTRED{E}{}{p.c_i}{\triangleright_\delta}{p'.c_i} + } +} +%%%%%%%%%%%%%%%%%%%%%%%%%%% MODULES +\item[ACC-MOD] +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\ModS{m}{T};\dots}} + }{ + \WTEG{p.m}{\subst{T}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\\ +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\ModSEq{m}{T}{p'};\dots}} + }{ + \WTEG{p.m}{\subst{T}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\item[ACC-MODEQ] +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\ModSEq{m}{T}{p'};\dots}} + }{ + \WTEGRED{p.m}{\triangleright_\delta}{\subst{p'}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\item[ACC-MODTYPE] +\inference{% + \frac{ + \WTEG{p}{\sig{\Spec_1;\dots;Spec_i;\ModType{S}{T};\dots}} + }{ + \WTEGRED{p.S}{\triangleright_\delta}{\subst{T}{p.l}{l}_{l \in \lab{Spec_1;\dots;Spec_i}}} + } +} +\end{description} +The function $\lab{}$ is used to calculate the set of label of +the set of specifications. It is defined by +$\lab{\Spec_1;\dots;\Spec_n}=\lab{\Spec_1}\cup\dots;\cup\lab{\Spec_n}$ +where $\lab{\Spec}$ is defined as follows: +\begin{itemize} +\item $\lab{\Assum{\Gamma}{c}{U}}=\{c\}$, +\item $\lab{\Def{\Gamma}{c}{t}{U}}=\{c\}$, +\item + $\lab{\Ind{\Gamma}{\Gamma_P}{\Gamma_C}{\Gamma_I}}=\dom{\Gamma_C}\cup\dom{\Gamma_I}$, +\item $\lab{\ModS{m}{T}}=\{m\}$, +\item $\lab{\ModSEq{m}{T}{M}}=\{m\}$, +\item $\lab{\ModType{S}{T}}=\{S\}$ +\end{itemize} +Environment access for modules and module types +\begin{description} +\item[ENV-MOD] +\inference{% + \frac{ + \WF{E;\ModS{m}{T};E'}{\Gamma} + }{ + \WT{E;\ModS{m}{T};E'}{\Gamma}{m}{T} + } +} +\item[] +\inference{% + \frac{ + \WF{E;\ModSEq{m}{T}{p};E'}{\Gamma} + }{ + \WT{E;\ModSEq{m}{T}{p};E'}{\Gamma}{m}{T} + } +} +\item[ENV-MODEQ] +\inference{% + \frac{ + \WF{E;\ModSEq{m}{T}{p};E'}{\Gamma} + }{ + \WTRED{E;\ModSEq{m}{T}{p};E'}{\Gamma}{m}{\triangleright_\delta}{p} + } +} +\item[ENV-MODTYPE] +\inference{% + \frac{ + \WF{E;\ModType{S}{T};E'}{\Gamma} + }{ + \WTRED{E;\ModType{S}{T};E'}{\Gamma}{S}{\triangleright_\delta}{T} + } +} +\item[ENV-INDP] +\inference{% + \frac{ + \WF{E;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}{} + }{ + \WTRED{E;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}{}{I_i}{\triangleright_\delta}{p.I_i} + } +} +\inference{% + \frac{ + \WF{E;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}{} + }{ + \WTRED{E;\Indp{}{\Gamma_P}{\Gamma_C}{\Gamma_I}{p}}{}{c_i}{\triangleright_\delta}{p.c_i} + } +} +\end{description} +% %%% replaced by \triangle_\delta +% Module path equality is a transitive and reflexive closure of the +% relation generated by ACC-MODEQ and ENV-MODEQ. +% \begin{itemize} +% \item []MP-EQ-REFL +% \inference{% +% \frac{ +% \WTEG{p}{T} +% }{ +% \WTEG{p}{p} +% } +% } +% \item []MP-EQ-TRANS +% \inference{% +% \frac{ +% \WTEGRED{p}{=}{p'}~~~~~~\WTEGRED{p'}{=}{p''} +% }{ +% \WTEGRED{p'}{=}{p''} +% } +% } + +% \end{itemize} + + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: + diff --git a/doc/refman/RefMan-oth.tex b/doc/refman/RefMan-oth.tex new file mode 100644 index 0000000000..35c55e07b4 --- /dev/null +++ b/doc/refman/RefMan-oth.tex @@ -0,0 +1,773 @@ +\chapter{Vernacular commands} +\label{Vernacular-commands} +\label{Other-commands} + +\section{Displaying} + +\subsection{\tt Print {\qualid}.}\comindex{Print} +This command displays on the screen informations about the declared or +defined object referred by {\qualid}. + +\begin{ErrMsgs} +\item {\qualid} \errindex{not a defined object} +\end{ErrMsgs} + +\begin{Variants} +\item {\tt Print Term {\qualid}.} +\comindex{Print Term}\\ +This is a synonym to {\tt Print {\qualid}} when {\qualid} denotes a +global constant. + +\item {\tt About {\qualid}.} +\label{About} +\comindex{About}\\ +This displays various informations about the object denoted by {\qualid}: +its kind (module, constant, assumption, inductive, +constructor, abbreviation\ldots), long name, type, implicit +arguments and argument scopes. + +%\item {\tt Print Proof {\qualid}.}\comindex{Print Proof}\\ +%In case \qualid\ denotes an opaque theorem defined in a section, +%it is stored on a special unprintable form and displayed as +%{\tt }. {\tt Print Proof} forces the printable form of \qualid\ +%to be computed and displays it. +\end{Variants} + +\subsection{\tt Print All.}\comindex{Print All} +This command displays informations about the current state of the +environment, including sections and modules. + +\begin{Variants} +\item {\tt Inspect \num.}\comindex{Inspect}\\ +This command displays the {\num} last objects of the current +environment, including sections and modules. +\item {\tt Print Section {\ident}.}\comindex{Print Section}\\ +should correspond to a currently open section, this command +displays the objects defined since the beginning of this section. +% Discontinued +%% \item {\tt Print.}\comindex{Print}\\ +%% This command displays the axioms and variables declarations in the +%% environment as well as the constants defined since the last variable +%% was introduced. +\end{Variants} + +\section{Requests to the environment} + +\subsection{\tt Check {\term}.} +\label{Check} +\comindex{Check} +This command displays the type of {\term}. When called in proof mode, +the term is checked in the local context of the current subgoal. + +\subsection{\tt Eval {\rm\sl convtactic} in {\term}.} +\comindex{Eval} + +This command performs the specified reduction on {\term}, and displays +the resulting term with its type. The term to be reduced may depend on +hypothesis introduced in the first subgoal (if a proof is in +progress). + +\SeeAlso section~\ref{Conversion-tactics}. + +\subsection{\tt Extraction \term.} +\label{ExtractionTerm} +\comindex{Extraction} +This command displays the extracted term from +{\term}. The extraction is processed according to the distinction +between {\Set} and {\Prop}; that is to say, between logical and +computational content (see section \ref{Sorts}). The extracted term is +displayed in Objective Caml syntax, where global identifiers are still +displayed as in \Coq\ terms. + +\begin{Variants} +\item \texttt{Recursive Extraction {\qualid$_1$} \ldots{} {\qualid$_n$}.}\\ + Recursively extracts all the material needed for the extraction of + globals {\qualid$_1$} \ldots{} {\qualid$_n$}. +\end{Variants} + +\SeeAlso chapter~\ref{Extraction}. + +\subsection{\tt Opaque \qualid$_1$ \dots \qualid$_n$.} +\comindex{Opaque}\label{Opaque} This command tells not to unfold the +the constants {\qualid$_1$} \dots {\qualid$_n$} in tactics using +$\delta$-conversion. Unfolding a constant is replacing it by its +definition. {\tt Opaque} can only apply on constants originally +defined as {\tt Transparent}. + +Constants defined by a proof ended by {\tt Qed} are automatically +stamped as {\tt Opaque} and can no longer be considered as {\tt +Transparent}. This is to keep with the usual mathematical practice of +{\em proof irrelevance}: what matters in a mathematical development is +the sequence of lemma statements, not their actual proofs. This +distinguishes lemmas from the usual defined constants, whose actual +values are of course relevant in general. + +\SeeAlso sections \ref{Conversion-tactics}, \ref{Automatizing}, +\ref{Theorem} + +\begin{ErrMsgs} +\item \errindex{The reference \qualid\ was not found in the current +environment}\\ + There is no constant referred by {\qualid} in the environment. + Nevertheless, if you asked \texttt{Opaque foo bar} + and if \texttt{bar} does not exist, \texttt{foo} is set opaque. +\end{ErrMsgs} + +\subsection{\tt Transparent \qualid$_1$ \dots \qualid$_n$.} +\comindex{Transparent}\label{Transparent} +This command is the converse of {\tt Opaque} and can only apply on constants originally defined as {\tt Transparent} to restore their initial behaviour after an {\tt Opaque} command. + +The constants automatically declared transparent are the ones defined by a proof ended by {\tt Defined}, or by a {\tt + Definition} or {\tt Local} with an explicit body. + +\Warning {\tt Transparent} and \texttt{Opaque} are not synchronous +with the reset mechanism. If a constant was transparent at point A, if +you set it opaque at point B and reset to point A, you return to state +of point A with the difference that the constant is still opaque. This +can cause changes in tactic scripts behaviour. + +At section or module closing, a constant recovers the status it got at +the time of its definition. + +%TODO: expliquer le rapport avec les sections + +\begin{ErrMsgs} +% \item \errindex{Can not set transparent.}\\ +% It is a constant from a required module or a parameter. +\item \errindex{The reference \qualid\ was not found in the current +environment}\\ + There is no constant referred by {\qualid} in the environment. +\end{ErrMsgs} + +\SeeAlso sections \ref{Conversion-tactics}, \ref{Automatizing}, +\ref{Theorem} + +\subsection{\tt Search {\qualid}.}\comindex{Search} +This command displays the name and type of all theorems of the current +context whose statement's conclusion has the form {\tt ({\qualid} t1 .. + tn)}. This command is useful to remind the user of the name of +library lemmas. +\begin{ErrMsgs} +\item \errindex{The reference \qualid\ was not found in the current +environment}\\ + There is no constant in the environment named \qualid. +\end{ErrMsgs} + +\begin{Variants} +\item +{\tt Search {\qualid} inside {\module$_1$} \ldots{} {\module$_n$}.} + +This restricts the search to constructions defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\item {\tt Search {\qualid} outside {\module$_1$} \ldots{} {\module$_n$}.} + +This restricts the search to constructions not defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\begin{ErrMsgs} +\item \errindex{Module/section \module{} not found} +No module \module{} has been required (see section~\ref{Require}). +\end{ErrMsgs} + +\end{Variants} + +\subsection{\tt SearchAbout {\qualid}.}\comindex{SearchAbout} +This command displays the name and type of all objects (theorems, +axioms, etc) of the current context whose statement contains \qualid. +This command is useful to remind the user of the name of library +lemmas. + +\begin{ErrMsgs} +\item \errindex{The reference \qualid\ was not found in the current +environment}\\ + There is no constant in the environment named \qualid. +\end{ErrMsgs} + +\begin{Variants} +\item {\tt SearchAbout [ \nelist{\textrm{\textsl{qualid-or-string}}}{} +].}\\ +\noindent where {\textrm{\textsl{qualid-or-string}}} is a {\qualid} or +a {\str}. + +This extension of {\tt SearchAbout} searches for all objects whose +statement mentions all of {\qualid} of the list and whose name +contains all {\str} of the list. + +\Example + +\begin{coq_example} +Require Import ZArith. +SearchAbout [ Zmult Zplus "distr" ]. +\end{coq_example} + +\item +\begin{tabular}[t]{@{}l} + {\tt SearchAbout {\term} inside {\module$_1$} \ldots{} {\module$_n$}.} \\ + {\tt SearchAbout [ \nelist{\textrm{\textsl{qualid-or-string}}}{} ] + inside {\module$_1$} \ldots{} {\module$_n$}.} +\end{tabular} + +This restricts the search to constructions defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\item +\begin{tabular}[t]{@{}l} + {\tt SearchAbout {\term} outside {\module$_1$}...{\module$_n$}.} \\ + {\tt SearchAbout [ \nelist{\textrm{\textsl{qualid-or-string}}}{} ] + outside {\module$_1$}...{\module$_n$}.} +\end{tabular} + +This restricts the search to constructions not defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\end{Variants} + +\subsection{\tt SearchPattern {\term}.}\comindex{SearchPattern} + +This command displays the name and type of all theorems of the current +context whose statement's conclusion matches the expression {\term} +where holes in the latter are denoted by ``{\texttt \_}''. + +\begin{coq_example} +Require Import Arith. +SearchPattern (_ + _ = _ + _). +\end{coq_example} + +Patterns need not be linear: you can express that the same expression +must occur in two places by using pattern variables `{\texttt +?{\ident}}''. + +\begin{coq_example} +Require Import Arith. +SearchPattern (?X1 + _ = _ + ?X1). +\end{coq_example} + +\begin{Variants} +\item {\tt SearchPattern {\term} inside +{\module$_1$} \ldots{} {\module$_n$}.}\comindex{SearchPattern \ldots{} inside +\ldots{}} + +This restricts the search to constructions defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\item {\tt SearchPattern {\term} outside {\module$_1$} \ldots{} {\module$_n$}.}\comindex{SearchPattern \ldots{} outside \ldots{}} + +This restricts the search to constructions not defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\end{Variants} + +\subsection{\tt SearchRewrite {\term}.}\comindex{SearchRewrite} + +This command displays the name and type of all theorems of the current +context whose statement's conclusion is an equality of which one side matches +the expression {\term =}. Holes in {\term} are denoted by ``{\texttt \_}''. + +\begin{coq_example} +Require Import Arith. +SearchRewrite (_ + _ + _). +\end{coq_example} + +\begin{Variants} +\item {\tt SearchRewrite {\term} inside +{\module$_1$} \ldots{} {\module$_n$}.} + +This restricts the search to constructions defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\item {\tt SearchRewrite {\term} outside {\module$_1$} \ldots{} {\module$_n$}.} + +This restricts the search to constructions not defined in modules +{\module$_1$} \ldots{} {\module$_n$}. + +\end{Variants} + +% \subsection{\tt SearchIsos {\term}.}\comindex{SearchIsos} +% \label{searchisos} +% \texttt{SearchIsos} searches terms by their type modulo isomorphism. +% This command displays the full name of all constants, variables, +% inductive types, and inductive constructors of the current +% context whose type is isomorphic to {\term} modulo the contextual part of the +% following axiomatization (the mutual inductive types with one constructor, +% without implicit arguments, and for which projections exist, are regarded as a +% sequence of $\sa{}$): + + +% \begin{tabbing} +% \ \ \ \ \=11.\ \=\kill +% \>1.\>$A=B\mx{ if }A\stackrel{\bt{}\io{}}{\lra{}}B$\\ +% \>2.\>$\sa{}x:A.B=\sa{}y:A.B[x\la{}y]\mx{ if }y\not\in{}FV(\sa{}x:A.B)$\\ +% \>3.\>$\Pi{}x:A.B=\Pi{}y:A.B[x\la{}y]\mx{ if }y\not\in{}FV(\Pi{}x:A.B)$\\ +% \>4.\>$\sa{}x:A.B=\sa{}x:B.A\mx{ if }x\not\in{}FV(A,B)$\\ +% \>5.\>$\sa{}x:(\sa{}y:A.B).C=\sa{}x:A.\sa{}y:B[y\la{}x].C[x\la{}(x,y)]$\\ +% \>6.\>$\Pi{}x:(\sa{}y:A.B).C=\Pi{}x:A.\Pi{}y:B[y\la{}x].C[x\la{}(x,y)]$\\ +% \>7.\>$\Pi{}x:A.\sa{}y:B.C=\sa{}y:(\Pi{}x:A.B).(\Pi{}x:A.C[y\la{}(y\sm{}x)]$\\ +% \>8.\>$\sa{}x:A.unit=A$\\ +% \>9.\>$\sa{}x:unit.A=A[x\la{}tt]$\\ +% \>10.\>$\Pi{}x:A.unit=unit$\\ +% \>11.\>$\Pi{}x:unit.A=A[x\la{}tt]$ +% \end{tabbing} + +% For more informations about the exact working of this command, see +% \cite{Del97}. + +\subsection{\tt Locate {\qualid}.}\comindex{Locate} +\label{Locate} +This command displays the full name of the qualified identifier {\qualid} +and consequently the \Coq\ module in which it is defined. + +\begin{coq_eval} +(*************** The last line should produce **************************) +(*********** Error: I.Dont.Exist not a defined object ******************) +\end{coq_eval} +\begin{coq_eval} +Set Printing Depth 50. +\end{coq_eval} +\begin{coq_example} +Locate nat. +Locate Datatypes.O. +Locate Init.Datatypes.O. +Locate Coq.Init.Datatypes.O. +Locate I.Dont.Exist. +\end{coq_example} + +\SeeAlso Section \ref{LocateSymbol} + +\section{Loading files} + +\Coq\ offers the possibility of loading different +parts of a whole development stored in separate files. Their contents +will be loaded as if they were entered from the keyboard. This means +that the loaded files are ASCII files containing sequences of commands +for \Coq's toplevel. This kind of file is called a {\em script} for +\Coq\index{Script file}. The standard (and default) extension of +\Coq's script files is {\tt .v}. + +\subsection{\tt Load {\ident}.} +\comindex{Load}\label{Load} +This command loads the file named {\ident}{\tt .v}, searching +successively in each of the directories specified in the {\em + loadpath}. (see section \ref{loadpath}) + +\begin{Variants} +\item {\tt Load {\str}.}\label{Load-str}\\ + Loads the file denoted by the string {\str}, where {\str} is any + complete filename. Then the \verb.~. and {\tt ..} + abbreviations are allowed as well as shell variables. If no + extension is specified, \Coq\ will use the default extension {\tt + .v} +\item {\tt Load Verbose {\ident}.}, + {\tt Load Verbose {\str}}\\ + \comindex{Load Verbose} + Display, while loading, the answers of \Coq\ to each command + (including tactics) contained in the loaded file + \SeeAlso section \ref{Begin-Silent} +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{Can't find file {\ident} on loadpath} +\end{ErrMsgs} + +\section{Compiled files}\label{compiled}\index{Compiled files} + +This feature allows to build files for a quick loading. When loaded, +the commands contained in a compiled file will not be {\em replayed}. +In particular, proofs will not be replayed. This avoids a useless +waste of time. + +\Rem A module containing an opened section cannot be compiled. + +% \subsection{\tt Compile Module {\ident}.} +% \index{Modules} +% \comindex{Compile Module} +% \index{.vo files} +% This command loads the file +% {\ident}{\tt .v} and plays the script it contains. Declarations, +% definitions and proofs it contains are {\em "packaged"} in a compiled +% form: the {\em module} named {\ident}. +% A file {\ident}{\tt .vo} is then created. +% The file {\ident}{\tt .v} is searched according to the +% current loadpath. +% The {\ident}{\tt .vo} is then written in the directory where +% {\ident}{\tt .v} was found. + +% \begin{Variants} +% \item \texttt{Compile Module {\ident} {\str}.}\\ +% Uses the file {\str}{\tt .v} or {\str} if the previous one does not +% exist to build the module {\ident}. In this case, {\str} is any +% string giving a filename in the UNIX sense (see section +% \ref{Load-str}). +% \Warning The given filename can not contain other caracters than +% the caracters of \Coq's identifiers : letters or digits or the +% underscore symbol ``\_''. + +% \item \texttt{Compile Module Specification {\ident}.}\\ +% \comindex{Compile Module Specification} +% Builds a specification module: only the types of terms are stored +% in the module. The bodies (the proofs) are {\em not} written +% in the module. In that case, the file created is {\ident}{\tt .vi}. +% This is only useful when proof terms take too much place in memory +% and are not necessary. + +% \item \texttt{Compile Verbose Module {\ident}.}\\ +% \comindex{Compile Verbose Module} +% Verbose version of Compile: shows the contents of the file being +% compiled. +% \end{Variants} + +% These different variants can be combined. + + +% \begin{ErrMsgs} +% \item \texttt{You cannot open a module when there are things other than}\\ +% \texttt{Modules and Imports in the context.}\\ +% The only commands allowed before a {Compile Module} command are {\tt +% Require},\\ +% {\tt Read Module} and {\tt Import}. Actually, The normal way to +% compile modules is by the {\tt coqc} command (see chapter +% \ref{Addoc-coqc}). +% \end{ErrMsgs} + +% \SeeAlso sections \ref{Opaque}, \ref{loadpath}, chapter +% \ref{Addoc-coqc} + +%\subsection{\tt Import {\qualid}.}\comindex{Import} +%\label{Import} + +%%%%%%%%%%%% +% Import and Export described in RefMan-mod.tex +% the minor difference (to avoid multiple Exporting of libraries) in +% the treatment of normal modules and libraries by Export omitted + + +\subsection{\tt Require {\dirpath}.} +\label{Require} +\comindex{Require} + +This command looks in the loadpath for a file containing module +{\dirpath}, then loads and opens (imports) its contents. +More precisely, if {\dirpath} splits into a library dirpath {\dirpath'} and a module name {\textsl{ident}}, then the file {\ident}{\tt .vo} is searched in a physical path mapped to the logical path {\dirpath'}. + +TODO: effect on the name table. + +% The implementation file ({\ident}{\tt .vo}) is searched first, +% then the specification file ({\ident}{\tt .vi}) in case of failure. +If the module required has already been loaded, \Coq\ +simply opens it (as {\tt Import {\dirpath}} would do it). +%If the module required is already loaded and open, \Coq\ +%displays the following warning: {\tt {\ident} already imported}. + +If a module {\it A} contains a command {\tt Require} {\it B} then the +command {\tt Require} {\it A} loads the module {\it B} but does not +open it (See the {\tt Require Export} variant below). + +\begin{Variants} +\item {\tt Require Export {\qualid}.}\\ + \comindex{Require Export} + This command acts as {\tt Require} {\qualid}. But if a module {\it + A} contains a command {\tt Require Export} {\it B}, then the + command {\tt Require} {\it A} opens the module {\it B} as if the + user would have typed {\tt Require}{\it B}. +% \item {\tt Require $[$ Implementation $|$ Specification $]$ {\qualid}.}\\ +% \comindex{Require Implementation} +% \comindex{Require Specification} +% Is the same as {\tt Require}, but specifying explicitly the +% implementation ({\tt.vo} file) or the specification ({\tt.vi} +% file). + +% Redundant ? +% \item {\tt Require {\qualid} {\str}.}\\ +% Specifies the file to load as being {\str} but containing module +% {\qualid}. +% The opened module is still {\ident} and therefore must have been loaded. +\item {\tt Require {\qualid} {\str}.}\\ + Specifies the file to load as being {\str} but containing module + {\qualid} which is then opened. +\end{Variants} + +These different variants can be combined. + +\begin{ErrMsgs} + +\item \errindex{Cannot load {\ident}: no physical path bound to {\dirpath}} + +\item \errindex{Can't find module toto on loadpath} + + The command did not find the file {\tt toto.vo}. Either {\tt + toto.v} exists but is not compiled or {\tt toto.vo} is in a directory + which is not in your {\tt LoadPath} (see section \ref{loadpath}). + +\item \errindex{Bad magic number} + + \index{Bad-magic-number@{\tt Bad Magic Number}} + The file {\tt{\ident}.vo} was found but either it is not a \Coq\ + compiled module, or it was compiled with an older and incompatible + version of \Coq. +\end{ErrMsgs} + +\SeeAlso chapter \ref{Addoc-coqc} + +\subsection{\tt Print Modules.} +\comindex{Print Modules} +This command shows the currently loaded and currently opened +(imported) modules. + +\subsection{\tt Declare ML Module {\str$_1$} .. {\str$_n$}.} +\comindex{Declare ML Module} +This commands loads the Objective Caml compiled files {\str$_1$} \dots +{\str$_n$} (dynamic link). It is mainly used to load tactics +dynamically. +% (see chapter \ref{WritingTactics}). + The files are +searched into the current Objective Caml loadpath (see the command {\tt +Add ML Path} in the section \ref{loadpath}). Loading of Objective Caml +files is only possible under the bytecode version of {\tt coqtop} +(i.e. {\tt coqtop} called with options {\tt -byte}, see chapter +\ref{Addoc-coqc}). + +\begin{ErrMsgs} +\item \errindex{File not found on loadpath : \str} +\item \errindex{Loading of ML object file forbidden in a native Coq} +\end{ErrMsgs} + +\subsection{\tt Print ML Modules.}\comindex{Print ML Modules} +This print the name of all \ocaml{} modules loaded with \texttt{Declare + ML Module}. To know from where these module were loaded, the user +should use the command \texttt{Locate File} (see page \pageref{Locate File}) + +\section{Loadpath} +\label{loadpath}\index{Loadpath} + +There are currently two loadpaths in \Coq. A loadpath where seeking +{\Coq} files (extensions {\tt .v} or {\tt .vo} or {\tt .vi}) and one where +seeking Objective Caml files. The default loadpath contains the +directory ``\texttt{.}'' denoting the current directory and mapped to the empty logical path (see section \ref{LongNames}). + +\subsection{\tt Pwd.}\comindex{Pwd}\label{Pwd} +This command displays the current working directory. + +\subsection{\tt Cd {\str}.}\comindex{Cd} +This command changes the current directory according to {\str} +which can be any valid path. + +\begin{Variants} +\item {\tt Cd.}\\ + Is equivalent to {\tt Pwd.} +\end{Variants} + +\subsection{\tt Add LoadPath {\str} as {\dirpath}.} +\comindex{Add LoadPath}\label{AddLoadPath} + +This command adds the path {\str} to the current {\Coq} loadpath and +maps it to the logical directory {\dirpath}, which means that every +file {\tt M.v} physically lying in directory {\str} becomes accessible +through logical name ``{\dirpath}{\tt{.M}}''. + +\Rem {\tt Add LoadPath} also adds {\str} to the current ML loadpath. + +\begin{Variants} +\item {\tt Add LoadPath {\str}.}\\ +Performs as {\tt Add LoadPath {\str} as {\dirpath}} but for the empty directory path. +\end{Variants} + +\subsection{\tt Add Rec LoadPath {\str} as {\dirpath}.}\comindex{Add Rec LoadPath}\label{AddRecLoadPath} +This command adds the directory {\str} and all its subdirectories +to the current \Coq\ loadpath. The top directory {\str} is mapped to the logical directory {\dirpath} while any subdirectory {\textsl{pdir}} is mapped to logical directory {\dirpath}{\tt{.pdir}} and so on. + +\Rem {\tt Add Rec LoadPath} also recursively adds {\str} to the current ML loadpath. + +\begin{Variants} +\item {\tt Add Rec LoadPath {\str}.}\\ +Works as {\tt Add Rec LoadPath {\str} as {\dirpath}} but for the empty logical directory path. +\end{Variants} + +\subsection{\tt Remove LoadPath {\str}.}\comindex{Remove LoadPath} +This command removes the path {\str} from the current \Coq\ loadpath. + +\subsection{\tt Print LoadPath.}\comindex{Print LoadPath} +This command displays the current \Coq\ loadpath. + +\subsection{\tt Add ML Path {\str}.}\comindex{Add ML Path} +This command adds the path {\str} to the current Objective Caml loadpath (see +the command {\tt Declare ML Module} in the section \ref{compiled}). + +\Rem This command is implied by {\tt Add LoadPath {\str} as {\dirpath}}. + +\subsection{\tt Add Rec ML Path {\str}.}\comindex{Add Rec ML Path} +This command adds the directory {\str} and all its subdirectories +to the current Objective Caml loadpath (see +the command {\tt Declare ML Module} in the section \ref{compiled}). + +\Rem This command is implied by {\tt Add Rec LoadPath {\str} as {\dirpath}}. + +\subsection{\tt Print ML Path {\str}.}\comindex{Print ML Path} +This command displays the current Objective Caml loadpath. +This command makes sense only under the bytecode version of {\tt +coqtop}, i.e. using option {\tt -byte} (see the +command {\tt Declare ML Module} in the section +\ref{compiled}). + +\subsection{\tt Locate File {\str}.}\comindex{Locate + File}\label{Locate File} +This command displays the location of file {\str} in the current loadpath. +Typically, {\str} is a \texttt{.cmo} or \texttt{.vo} or \texttt{.v} file. + +\subsection{\tt Locate Library {\dirpath}.} +\comindex{Locate Library} +This command gives the status of the \Coq\ module {\dirpath}. It tells if the +module is loaded and if not searches in the load path for a module +of logical name {\dirpath}. + +\section{States and Reset} + +\subsection{\tt Reset \ident.} +\comindex{Reset} +This command removes all the objects in the environment since \ident\ +was introduced, including \ident. \ident\ may be the name of a defined +or declared object as well as the name of a section. One cannot reset +over the name of a module or of an object inside a module. + +\begin{ErrMsgs} +\item \ident: \errindex{no such entry} +\end{ErrMsgs} + +\subsection{\tt Back.} +\comindex{Back} + +This commands undoes all the effects of the last vernacular +command. This does not include commands that only access to the +environment like those described in the previous sections of this +chapter (for instance {\tt Require} and {\tt Load} can be undone, but +not {\tt Check} and {\tt Locate}). Commands read from a vernacular +file are considered as a single command. + +\begin{Variants} +\item {\tt Back $n$} \\ + Undoes $n$ vernacular commands. +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{Reached begin of command history} \\ + Happens when there is vernacular command to undo. +\end{ErrMsgs} + +\subsection{\tt Restore State \str.} +\comindex{Restore State} + Restores the state contained in the file \str. + +\begin{Variants} +\item {\tt Restore State \ident}\\ + Equivalent to {\tt Restore State "}{\ident}{\tt .coq"}. +\item {\tt Reset Initial.}\comindex{Reset Initial}\\ + Goes back to the initial state (like after the command {\tt coqtop}, + when the interactive session began). This command is only available + interactively. +\end{Variants} + +\subsection{\tt Write State \str.} +\comindex{Write State} +Writes the current state into a file \str{} for +use in a further session. This file can be given as the {\tt + inputstate} argument of the commands {\tt coqtop} and {\tt coqc}. + +\begin{Variants} +\item {\tt Write State \ident}\\ + Equivalent to {\tt Write State "}{\ident}{\tt .coq"}. + The state is saved in the current directory (see \pageref{Pwd}). +\end{Variants} + +\section{Quitting and debugging} + +\subsection{\tt Quit.}\comindex{Quit} +This command permits to quit \Coq. + +\subsection{\tt Drop.}\comindex{Drop}\label{Drop} + +This is used mostly as a debug facility by \Coq's implementors +and does not concern the casual user. +This command permits to leave {\Coq} temporarily and enter the +Objective Caml toplevel. The Objective Caml command: + +\begin{flushleft} +\begin{verbatim} +#use "include";; +\end{verbatim} +\end{flushleft} + +\noindent add the right loadpaths and loads some toplevel printers for +all abstract types of \Coq - section\_path, identfifiers, terms, judgements, +\dots. You can also use the file \texttt{base\_include} instead, +that loads only the pretty-printers for section\_paths and +identifiers. +% See section \ref{test-and-debug} more information on the +% usage of the toplevel. +You can return back to \Coq{} with the command: + +\begin{flushleft} +\begin{verbatim} +go();; +\end{verbatim} +\end{flushleft} + +\begin{Warnings} +\item It only works with the bytecode version of {\Coq} (i.e. {\tt coqtop} called with option {\tt -byte}, see page \pageref{binary-images}). +\item You must have compiled {\Coq} from the source package and set the + environment variable \texttt{COQTOP} to the root of your copy of the sources (see section \ref{EnvVariables}). +\end{Warnings} + +\subsection{\tt Time \textrm{\textsl{command}}.}\comindex{Time} +\label{time} +This command executes the vernac command \textrm{\textsl{command}} +and display the time needed to execute it. + +\section{Controlling display} + +\subsection{\tt Set Silent.} +\comindex{Begin Silent} +\label{Begin-Silent} +\index{Silent mode} +This command turns off the normal displaying. + +\subsection{\tt Unset Silent.}\comindex{End Silent} +This command turns the normal display on. + +\subsection{\tt Set Printing Width {\integer}.}\comindex{Set Printing Width} +This command sets which left-aligned part of the width of the screen +is used for display. + +\subsection{\tt Unset Printing Width.}\comindex{Unset Printing Width} +This command resets the width of the screen used for display to its +default value (which is 78 at the time of writing this documentation). + +\subsection{\tt Test Printing Width.}\comindex{Test Printing Width} +This command displays the current screen width used for display. + +\subsection{\tt Set Printing Depth {\integer}.}\comindex{Set Printing Depth} +This command sets the nesting depth of the formatter used for +pretty-printing. Beyond this depth, display of subterms is replaced by +dots. + +\subsection{\tt Unset Printing Depth.}\comindex{Unset Printing Depth} +This command resets the nesting depth of the formatter used for +pretty-printing to its default value (at the +time of writing this documentation, the default value is 50). + +\subsection{\tt Test Printing Depth.}\comindex{Test Printing Depth} +This command displays the current nesting depth used for display. + +%\subsection{\tt Explain ...} +%Not yet documented. + +%\subsection{\tt Go ...} +%Not yet documented. + +%\subsection{\tt Abstraction ...} +%Not yet documented. + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-pre.tex b/doc/refman/RefMan-pre.tex new file mode 100755 index 0000000000..4559aaf471 --- /dev/null +++ b/doc/refman/RefMan-pre.tex @@ -0,0 +1,519 @@ +\setheaders{Credits} +\chapter*{Credits} +%\addcontentsline{toc}{section}{Credits} + +\Coq{}~ is a proof assistant for higher-order logic, allowing the +development of computer programs consistent with their formal +specification. It is the result of about ten years of research of the +Coq project. We shall briefly survey here three main aspects: the +\emph{logical language} in which we write our axiomatizations and +specifications, the \emph{proof assistant} which allows the development +of verified mathematical proofs, and the \emph{program extractor} which +synthesizes computer programs obeying their formal specifications, +written as logical assertions in the language. + +The logical language used by {\Coq} is a variety of type theory, +called the \emph{Calculus of Inductive Constructions}. Without going +back to Leibniz and Boole, we can date the creation of what is now +called mathematical logic to the work of Frege and Peano at the turn +of the century. The discovery of antinomies in the free use of +predicates or comprehension principles prompted Russell to restrict +predicate calculus with a stratification of \emph{types}. This effort +culminated with \emph{Principia Mathematica}, the first systematic +attempt at a formal foundation of mathematics. A simplification of +this system along the lines of simply typed $\lambda$-calculus +occurred with Church's \emph{Simple Theory of Types}. The +$\lambda$-calculus notation, originally used for expressing +functionality, could also be used as an encoding of natural deduction +proofs. This Curry-Howard isomorphism was used by N. de Bruijn in the +\emph{Automath} project, the first full-scale attempt to develop and +mechanically verify mathematical proofs. This effort culminated with +Jutting's verification of Landau's \emph{Grundlagen} in the 1970's. +Exploiting this Curry-Howard isomorphism, notable achievements in +proof theory saw the emergence of two type-theoretic frameworks; the +first one, Martin-L\"of's \emph{Intuitionistic Theory of Types}, +attempts a new foundation of mathematics on constructive principles. +The second one, Girard's polymorphic $\lambda$-calculus $F_\omega$, is +a very strong functional system in which we may represent higher-order +logic proof structures. Combining both systems in a higher-order +extension of the Automath languages, T. Coquand presented in 1985 the +first version of the \emph{Calculus of Constructions}, CoC. This strong +logical system allowed powerful axiomatizations, but direct inductive +definitions were not possible, and inductive notions had to be defined +indirectly through functional encodings, which introduced +inefficiencies and awkwardness. The formalism was extended in 1989 by +T. Coquand and C. Paulin with primitive inductive definitions, leading +to the current \emph{Calculus of Inductive Constructions}. This +extended formalism is not rigorously defined here. Rather, numerous +concrete examples are discussed. We refer the interested reader to +relevant research papers for more information about the formalism, its +meta-theoretic properties, and semantics. However, it should not be +necessary to understand this theoretical material in order to write +specifications. It is possible to understand the Calculus of Inductive +Constructions at a higher level, as a mixture of predicate calculus, +inductive predicate definitions presented as typed PROLOG, and +recursive function definitions close to the language ML. + +Automated theorem-proving was pioneered in the 1960's by Davis and +Putnam in propositional calculus. A complete mechanization (in the +sense of a semi-decision procedure) of classical first-order logic was +proposed in 1965 by J.A. Robinson, with a single uniform inference +rule called \emph{resolution}. Resolution relies on solving equations +in free algebras (i.e. term structures), using the \emph{unification + algorithm}. Many refinements of resolution were studied in the +1970's, but few convincing implementations were realized, except of +course that PROLOG is in some sense issued from this effort. A less +ambitious approach to proof development is computer-aided +proof-checking. The most notable proof-checkers developed in the +1970's were LCF, designed by R. Milner and his colleagues at U. +Edinburgh, specialized in proving properties about denotational +semantics recursion equations, and the Boyer and Moore theorem-prover, +an automation of primitive recursion over inductive data types. While +the Boyer-Moore theorem-prover attempted to synthesize proofs by a +combination of automated methods, LCF constructed its proofs through +the programming of \emph{tactics}, written in a high-level functional +meta-language, ML. + +The salient feature which clearly distinguishes our proof assistant +from say LCF or Boyer and Moore's, is its possibility to extract +programs from the constructive contents of proofs. This computational +interpretation of proof objects, in the tradition of Bishop's +constructive mathematics, is based on a realizability interpretation, +in the sense of Kleene, due to C. Paulin. The user must just mark his +intention by separating in the logical statements the assertions +stating the existence of a computational object from the logical +assertions which specify its properties, but which may be considered +as just comments in the corresponding program. Given this information, +the system automatically extracts a functional term from a consistency +proof of its specifications. This functional term may be in turn +compiled into an actual computer program. This methodology of +extracting programs from proofs is a revolutionary paradigm for +software engineering. Program synthesis has long been a theme of +research in artificial intelligence, pioneered by R. Waldinger. The +Tablog system of Z. Manna and R. Waldinger allows the deductive +synthesis of functional programs from proofs in tableau form of their +specifications, written in a variety of first-order logic. Development +of a systematic \emph{programming logic}, based on extensions of +Martin-L\"of's type theory, was undertaken at Cornell U. by the Nuprl +team, headed by R. Constable. The first actual program extractor, PX, +was designed and implemented around 1985 by S. Hayashi from Kyoto +University. It allows the extraction of a LISP program from a proof +in a logical system inspired by the logical formalisms of S. Feferman. +Interest in this methodology is growing in the theoretical computer +science community. We can foresee the day when actual computer systems +used in applications will contain certified modules, automatically +generated from a consistency proof of their formal specifications. We +are however still far from being able to use this methodology in a +smooth interaction with the standard tools from software engineering, +i.e. compilers, linkers, run-time systems taking advantage of special +hardware, debuggers, and the like. We hope that {\Coq} can be of use +to researchers interested in experimenting with this new methodology. + +A first implementation of CoC was started in 1984 by G. Huet and T. +Coquand. Its implementation language was CAML, a functional +programming language from the ML family designed at INRIA in +Rocquencourt. The core of this system was a proof-checker for CoC seen +as a typed $\lambda$-calculus, called the \emph{Constructive Engine}. +This engine was operated through a high-level notation permitting the +declaration of axioms and parameters, the definition of mathematical +types and objects, and the explicit construction of proof objects +encoded as $\lambda$-terms. A section mechanism, designed and +implemented by G. Dowek, allowed hierarchical developments of +mathematical theories. This high-level language was called the +\emph{Mathematical Vernacular}. Furthermore, an interactive +\emph{Theorem Prover} permitted the incremental construction of proof +trees in a top-down manner, subgoaling recursively and backtracking +from dead-alleys. The theorem prover executed tactics written in CAML, +in the LCF fashion. A basic set of tactics was predefined, which the +user could extend by his own specific tactics. This system (Version +4.10) was released in 1989. Then, the system was extended to deal +with the new calculus with inductive types by C. Paulin, with +corresponding new tactics for proofs by induction. A new standard set +of tactics was streamlined, and the vernacular extended for tactics +execution. A package to compile programs extracted from proofs to +actual computer programs in CAML or some other functional language was +designed and implemented by B. Werner. A new user-interface, relying +on a CAML-X interface by D. de Rauglaudre, was designed and +implemented by A. Felty. It allowed operation of the theorem-prover +through the manipulation of windows, menus, mouse-sensitive buttons, +and other widgets. This system (Version 5.6) was released in 1991. + +\Coq{} was ported to the new implementation Caml-light of X. Leroy and +D. Doligez by D. de Rauglaudre (Version 5.7) in 1992. A new version +of \Coq{} was then coordinated by C. Murthy, with new tools designed +by C. Parent to prove properties of ML programs (this methodology is +dual to program extraction) and a new user-interaction loop. This +system (Version 5.8) was released in May 1993. A Centaur interface +\textsc{CTCoq} was then developed by Y. Bertot from the Croap project +from INRIA-Sophia-Antipolis. + +In parallel, G. Dowek and H. Herbelin developed a new proof engine, +allowing the general manipulation of existential variables +consistently with dependent types in an experimental version of \Coq{} +(V5.9). + +The version V5.10 of \Coq{} is based on a generic system for +manipulating terms with binding operators due to Chet Murthy. A new +proof engine allows the parallel development of partial proofs for +independent subgoals. The structure of these proof trees is a mixed +representation of derivation trees for the Calculus of Inductive +Constructions with abstract syntax trees for the tactics scripts, +allowing the navigation in a proof at various levels of details. The +proof engine allows generic environment items managed in an +object-oriented way. This new architecture, due to C. Murthy, +supports several new facilities which make the system easier to extend +and to scale up: + +\begin{itemize} +\item User-programmable tactics are allowed +\item It is possible to separately verify development modules, and to + load their compiled images without verifying them again - a quick + relocation process allows their fast loading +\item A generic parsing scheme allows user-definable notations, with a + symmetric table-driven pretty-printer +\item Syntactic definitions allow convenient abbreviations +\item A limited facility of meta-variables allows the automatic + synthesis of certain type expressions, allowing generic notations + for e.g. equality, pairing, and existential quantification. +\end{itemize} + +In the Fall of 1994, C. Paulin-Mohring replaced the structure of +inductively defined types and families by a new structure, allowing +the mutually recursive definitions. P. Manoury implemented a +translation of recursive definitions into the primitive recursive +style imposed by the internal recursion operators, in the style of the +ProPre system. C. Mu{\~n}oz implemented a decision procedure for +intuitionistic propositional logic, based on results of R. Dyckhoff. +J.C. Filli{\^a}tre implemented a decision procedure for first-order +logic without contraction, based on results of J. Ketonen and R. +Weyhrauch. Finally C. Murthy implemented a library of inversion +tactics, relieving the user from tedious definitions of ``inversion +predicates''. + +\begin{flushright} +Rocquencourt, Feb. 1st 1995\\ +Gérard Huet +\end{flushright} + +\section*{Credits: addendum for version 6.1} +%\addcontentsline{toc}{section}{Credits: addendum for version V6.1} + +The present version 6.1 of \Coq{} is based on the V5.10 architecture. It +was ported to the new language Objective Caml by Bruno Barras. The +underlying framework has slightly changed and allows more conversions +between sorts. + +The new version provides powerful tools for easier developments. + +Cristina Cornes designed an extension of the \Coq{} syntax to allow +definition of terms using a powerful pattern-matching analysis in the +style of ML programs. + +Amokrane Saïbi wrote a mechanism to simulate +inheritance between types families extending a proposal by Peter +Aczel. He also developed a mechanism to automatically compute which +arguments of a constant may be inferred by the system and consequently +do not need to be explicitly written. + +Yann Coscoy designed a command which explains a proof term using +natural language. Pierre Cr{\'e}gut built a new tactic which solves +problems in quantifier-free Presburger Arithmetic. Both +functionalities have been integrated to the \Coq{} system by Hugo +Herbelin. + +Samuel Boutin designed a tactic for simplification of commutative +rings using a canonical set of rewriting rules and equality modulo +associativity and commutativity. + +Finally the organisation of the \Coq{} distribution has been supervised +by Jean-Christophe Filliâtre with the help of Judicaël Courant +and Bruno Barras. + +\begin{flushright} +Lyon, Nov. 18th 1996\\ +Christine Paulin +\end{flushright} + +\section*{Credits: addendum for version 6.2} +%\addcontentsline{toc}{section}{Credits: addendum for version V6.2} + +In version 6.2 of \Coq{}, the parsing is done using camlp4, a +preprocessor and pretty-printer for CAML designed by Daniel de +Rauglaudre at INRIA. Daniel de Rauglaudre made the first adaptation +of \Coq{} for camlp4, this work was continued by Bruno Barras who also +changed the structure of \Coq{} abstract syntax trees and the primitives +to manipulate them. The result of +these changes is a faster parsing procedure with greatly improved +syntax-error messages. The user-interface to introduce grammar or +pretty-printing rules has also changed. + +Eduardo Giménez redesigned the internal +tactic libraries, giving uniform names +to Caml functions corresponding to \Coq{} tactic names. + +Bruno Barras wrote new more efficient reductions functions. + +Hugo Herbelin introduced more uniform notations in the \Coq{} +specification language: the definitions by fixpoints and +pattern-matching have a more readable syntax. Patrick Loiseleur +introduced user-friendly notations for arithmetic expressions. + +New tactics were introduced: Eduardo Giménez improved a mechanism to +introduce macros for tactics, and designed special tactics for +(co)inductive definitions; Patrick Loiseleur designed a tactic to +simplify polynomial expressions in an arbitrary commutative ring which +generalizes the previous tactic implemented by Samuel Boutin. +Jean-Christophe Filli\^atre introduced a tactic for refining a goal, +using a proof term with holes as a proof scheme. + +David Delahaye designed the \textsf{SearchIsos} tool to search an +object in the library given its type (up to isomorphism). + +Henri Laulhère produced the \Coq{} distribution for the Windows environment. + +Finally, Hugo Herbelin was the main coordinator of the \Coq{} +documentation with principal contributions by Bruno Barras, David Delahaye, +Jean-Christophe Filli\^atre, Eduardo +Giménez, Hugo Herbelin and Patrick Loiseleur. + +\begin{flushright} +Orsay, May 4th 1998\\ +Christine Paulin +\end{flushright} + +\section*{Credits: addendum for version 6.3} +The main changes in version V6.3 was the introduction of a few new tactics +and the extension of the guard condition for fixpoint definitions. + + +B. Barras extended the unification algorithm to complete partial terms +and solved various tricky bugs related to universes.\\ +D. Delahaye developed the \texttt{AutoRewrite} tactic. He also designed the new +behavior of \texttt{Intro} and provided the tacticals \texttt{First} and +\texttt{Solve}.\\ +J.-C. Filli\^atre developed the \texttt{Correctness} tactic.\\ +E. Gim\'enez extended the guard condition in fixpoints.\\ +H. Herbelin designed the new syntax for definitions and extended the +\texttt{Induction} tactic.\\ +P. Loiseleur developed the \texttt{Quote} tactic and +the new design of the \texttt{Auto} +tactic, he also introduced the index of +errors in the documentation.\\ +C. Paulin wrote the \texttt{Focus} command and introduced +the reduction functions in definitions, this last feature +was proposed by J.-F. Monin from CNET Lannion. + +\begin{flushright} +Orsay, Dec. 1999\\ +Christine Paulin +\end{flushright} + +%\newpage + +\section*{Credits: versions 7} + +The version V7 is a new implementation started in September 1999 by +Jean-Christophe Filliâtre. This is a major revision with respect to +the internal architecture of the system. The \Coq{} version 7.0 was +distributed in March 2001, version 7.1 in September 2001, version +7.2 in January 2002, version 7.3 in May 2002 and version 7.4 in +February 2003. + +Jean-Christophe Filliâtre designed the architecture of the new system, he +introduced a new representation for environments and wrote a new kernel +for type-checking terms. His approach was to use functional +data-structures in order to get more sharing, to prepare the addition +of modules and also to get closer to a certified kernel. + +Hugo Herbelin introduced a new structure of terms with local +definitions. He introduced ``qualified'' names, wrote a new +pattern-matching compilation algorithm and designed a more compact +algorithm for checking the logical consistency of universes. He +contributed to the simplification of {\Coq} internal structures and the +optimisation of the system. He added basic tactics for forward +reasoning and coercions in patterns. + +David Delahaye introduced a new language for tactics. General tactics +using pattern-matching on goals and context can directly be written +from the {\Coq} toplevel. He also provided primitives for the design +of user-defined tactics in \textsc{Caml}. + +Micaela Mayero contributed the library on real numbers. +Olivier Desmettre extended this library with axiomatic +trigonometric functions, square, square roots, finite sums, Chasles +property and basic plane geometry. + +Jean-Christophe Filliâtre and Pierre Letouzey redesigned a new +extraction procedure from \Coq{} terms to \textsc{Caml} or +\textsc{Haskell} programs. This new +extraction procedure, unlike the one implemented in previous version +of \Coq{} is able to handle all terms in the Calculus of Inductive +Constructions, even involving universes and strong elimination. P. +Letouzey adapted user contributions to extract ML programs when it was +sensible. +Jean-Christophe Filliâtre wrote \verb=coqdoc=, a documentation +tool for {\Coq} libraries usable from version 7.2. + +Bruno Barras improved the reduction algorithms efficiency and +the confidence level in the correctness of {\Coq} critical type-checking +algorithm. + +Yves Bertot designed the \texttt{SearchPattern} and +\texttt{SearchRewrite} tools and the support for the \textsc{pcoq} interface +(\url{http://www-sop.inria.fr/lemme/pcoq/}). + +Micaela Mayero and David Delahaye introduced {\tt Field}, a decision tactic for commutative fields. + +Christine Paulin changed the elimination rules for empty and singleton +propositional inductive types. + +Loïc Pottier developed {\tt Fourier}, a tactic solving linear inequalities on real numbers. + +Pierre Crégut developed a new version based on reflexion of the {\tt Omega} +decision tactic. + +Claudio Sacerdoti Coen designed an XML output for the {\Coq} +modules to be used in the Hypertextual Electronic Library of +Mathematics (HELM cf \url{http://www.cs.unibo.it/helm}). + +A library for efficient representation of finite maps using binary trees +contributed by Jean Goubault was integrated in the basic theories. + +Jacek Chrz\k{a}szcz designed and implemented the module system of +{\Coq} whose foundations are in Judicaël Courant's PhD thesis. + +\bigskip + +The development was coordinated by C. Paulin. + +Many discussions within the Démons team and the LogiCal project +influenced significantly the design of {\Coq} especially with +%J. Chrz\k{a}szcz, +J. Courant, P. Courtieu, J. Duprat, J. Goubault, A. Miquel, +C. Marché, B. Monate and B. Werner. + +Intensive users suggested improvements of the system : +Y. Bertot, L. Pottier, L. Théry , P. Zimmerman from INRIA, +C. Alvarado, P. Crégut, J.-F. Monin from France Telecom R \& D. +\begin{flushright} +Orsay, May. 2002\\ +Hugo Herbelin \& Christine Paulin +\end{flushright} + +\section*{Credits: version 8.0} + +{\Coq} version 8 is a major revision of the {\Coq} proof assistant. +First, the underlying logic is slightly different. The so-called {\em +impredicativity} of the sort {\tt Set} has been dropped. The main +reason is that it is inconsistent with the principle of description +which is quite a useful principle for formalizing %classical +mathematics within classical logic. Moreover, even in an constructive +setting, the impredicativity of {\tt Set} does not add so much in +practice and is even subject of criticism from a large part of the +intuitionistic mathematician community. Nevertheless, the +impredicativity of {\tt Set} remains optional for users interested in +investigating mathematical developments which rely on it. + +Secondly, the concrete syntax of terms has been completely +revised. The main motivations were + +\begin{itemize} +\item a more uniform, purified style: all constructions are now lowercase, + with a functional programming perfume (e.g. abstraction is now + written {\tt fun}), and more directly accessible to the novice + (e.g. dependent product is now written {\tt forall} and allows + omission of types). Also, parentheses and are no longer mandatory + for function application. +\item extensibility: some standard notations (e.g. ``<'' and ``>'') were + incompatible with the previous syntax. Now all standard arithmetic + notations (=, +, *, /, <, <=, ... and more) are directly part of the + syntax. +\end{itemize} + +Together with the revision of the concrete syntax, a new mechanism of +{\em interpretation scopes} permits to reuse the same symbols +(typically +, -, *, /, <, <=) in various mathematical theories without +any ambiguities for {\Coq}, leading to a largely improved readability of +{\Coq} scripts. New commands to easily add new symbols are also +provided. + +Coming with the new syntax of terms, a slight reform of the tactic +language and of the language of commands has been carried out. The +purpose here is a better uniformity making the tactics and commands +easier to use and to remember. + +Thirdly, a restructuration and uniformisation of the standard library +of {\Coq} has been performed. There is now just one Leibniz' equality +usable for all the different kinds of {\Coq} objects. Also, the set of +real numbers now lies at the same level as the sets of natural and +integer numbers. Finally, the names of the standard properties of +numbers now follow a standard pattern and the symbolic +notations for the standard definitions as well. + +The fourth point is the release of \CoqIDE{}, a new graphical +gtk2-based interface fully integrated to {\Coq}. Close in style from +the Proof General Emacs interface, it is faster and its integration +with {\Coq} makes interactive developments more friendly. All +mathematical Unicode symbols are usable within \CoqIDE{}. + +Finally, the module system of {\Coq} completes the picture of {\Coq} +version 8.0. Though released with an experimental status in the previous +version 7.4, it should be considered as a salient feature of the new +version. + +Besides, {\Coq} comes with its load of novelties and improvements: new +or improved tactics (including a new tactic for solving first-order +statements), new management commands, extended libraries. + +\bigskip + +Bruno Barras and Hugo Herbelin have been the main contributors of the +reflexion and the implementation of the new syntax. The smart +automatic translator from old to new syntax released with {\Coq} is also +their work with contributions by Olivier Desmettre. + +Hugo Herbelin is the main designer and implementor of the notion of +interpretation scopes and of the commands for easily adding new notations. + +Hugo Herbelin is the main implementor of the restructuration of the +standard library. + +Pierre Corbineau is the main designer and implementor of the new +tactic for solving first-order statements in presence of inductive +types. He is also the maintainer of the non-domain specific automation +tactics. + +Benjamin Monate is the developer of the \CoqIDE{} graphical +interface with contributions by Jean-Christophe Filliâtre, Pierre +Letouzey and Claude Marché. + +Claude Marché coordinated the edition of the Reference Manual for + \Coq{} V8.0. + +Pierre Letouzey and Jacek Chrz\k{a}szcz respectively maintained the +extraction tool and module system of {\Coq}. + +Jean-Christophe Filliâtre, Pierre Letouzey, Hugo Herbelin and +contributors from Sophia-Antipolis and Nijmegen participated to the +extension of the library. + +Hugo Herbelin and Christine Paulin coordinated the development which +was under the responsability of Christine Paulin. + +\begin{flushright} +Palaiseau \& Orsay, Apr. 2004\\ +Hugo Herbelin \& Christine Paulin +\end{flushright} + + +%\newpage + +% Integration of ZArith lemmas from Sophia and Nijmegen. + + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-pro.tex b/doc/refman/RefMan-pro.tex new file mode 100755 index 0000000000..7565612d14 --- /dev/null +++ b/doc/refman/RefMan-pro.tex @@ -0,0 +1,389 @@ +\chapter{Proof handling} +\index{Proof editing} +\label{Proof-handling} + +In \Coq's proof editing mode all top-level commands documented in +chapter \ref{Vernacular-commands} remain available +and the user has access to specialized commands dealing with proof +development pragmas documented in this section. He can also use some +other specialized commands called {\em tactics}. They are the very +tools allowing the user to deal with logical reasoning. They are +documented in chapter \ref{Tactics}.\\ +When switching in editing proof mode, the prompt +\index{Prompt} +{\tt Coq <} is changed into {\tt {\ident} <} where {\ident} is the +declared name of the theorem currently edited. + +At each stage of a proof development, one has a list of goals to +prove. Initially, the list consists only in the theorem itself. After +having applied some tactics, the list of goals contains the subgoals +generated by the tactics. + +To each subgoal is associated a number of +hypotheses we call the {\em \index*{local context}} of the goal. +Initially, the local context is empty. It is enriched by the use of +certain tactics (see mainly section~\ref{intro}). + +When a proof is achieved the message {\tt Proof completed} is +displayed. One can then store this proof as a defined constant in the +environment. Because there exists a correspondence between proofs and +terms of $\lambda$-calculus, known as the {\em Curry-Howard +isomorphism} \cite{How80,Bar91,Gir89,Hue89}, \Coq~ stores proofs as +terms of {\sc Cic}. Those terms are called {\em proof + terms}\index{Proof term}. + +It is possible to edit several proofs at the same time: see section +\ref{Resume} + +\ErrMsg When one attempts to use a proof editing command out of the +proof editing mode, \Coq~ raises the error message : \errindex{No focused + proof}. + +\section{Switching on/off the proof editing mode} + +\subsection{\tt Goal {\form}.} +\comindex{Goal}\label{Goal} +This command switches \Coq~ to editing proof mode and sets {\form} as +the original goal. It associates the name {\tt Unnamed\_thm} to +that goal. + +\begin{ErrMsgs} +\item \errindex{the term \form\ has type \ldots{} which should be Set, + Prop or Type} +%\item \errindex{Proof objects can only be abstracted} +%\item \errindex{A goal should be a type} +%\item \errindex{repeated goal not permitted in refining mode} +%the command {\tt Goal} cannot be used while a proof is already being edited. +\end{ErrMsgs} + +\SeeAlso section \ref{Theorem} + +\subsection{\tt Qed.}\comindex{Qed}\label{Qed} +This command is available in interactive editing proof mode when the +proof is completed. Then {\tt Qed} extracts a proof term from the +proof script, switches back to {\Coq} top-level and attaches the +extracted proof term to the declared name of the original goal. This +name is added to the environment as an {\tt Opaque} constant. + +\begin{ErrMsgs} +\item \errindex{Attempt to save an incomplete proof} +%\item \ident\ \errindex{already exists}\\ +% The implicit name is already defined. You have then to provide +% explicitly a new name (see variant 3 below). +\item Sometimes an error occurs when building the proof term, +because tactics do not enforce completely the term construction +constraints. + +The user should also be aware of the fact that since the proof term is +completely rechecked at this point, one may have to wait a while when +the proof is large. In some exceptional cases one may even incur a +memory overflow. +\end{ErrMsgs} + +\begin{Variants} + +\item {\tt Defined.} +\comindex{Defined} +\label{Defined} + + Defines the proved term as a transparent constant. + +\item {\tt Save.} +\comindex{Save} + + Is equivalent to {\tt Qed}. + +\item {\tt Save {\ident}.} + + Forces the name of the original goal to be {\ident}. This command + (and the following ones) can only be used if the original goal has + been opened using the {\tt Goal} command. + +\item {\tt Save Theorem {\ident}.} \\ + {\tt Save Lemma {\ident}.} \\ + {\tt Save Remark {\ident}.}\\ + {\tt Save Fact {\ident}.} + + Are equivalent to {\tt Save {\ident}.} +\end{Variants} + +\subsection{\tt Admitted.}\comindex{Admitted}\label{Admitted} +This command is available in interactive editing proof mode to give up +the current proof and declare the initial goal as an axiom. + +\subsection{\tt Theorem {\ident} : {\form}.} +\comindex{Theorem} +\label{Theorem} + +This command switches to interactive editing proof mode and declares +{\ident} as being the name of the original goal {\form}. When declared +as a {\tt Theorem}, the name {\ident} is known at all section levels: +{\tt Theorem} is a {\sl global} lemma. + +%\ErrMsg (see section \ref{Goal}) + +\begin{ErrMsgs} + +\item \errindex{the term \form\ has type \ldots{} which should be Set, + Prop or Type} + +\item \errindexbis{\ident already exists}{already exists} + + The name you provided already defined. You have then to choose + another name. + +\end{ErrMsgs} + + +\begin{Variants} + +\item {\tt Lemma {\ident} : {\form}.} +\comindex{Lemma} + + It is equivalent to {\tt Theorem {\ident} : {\form}.} + +\item {\tt Remark {\ident} : {\form}.}\comindex{Remark}\\ + {\tt Fact {\ident} : {\form}.}\comindex{Fact} + + Used to have a different meaning, but are now equivalent to {\tt + Theorem {\ident} : {\form}.} They are kept for compatibility. + +\item {\tt Definition {\ident} : {\form}.} +\comindex{Definition} + + Analogous to {\tt Theorem}, intended to be used in conjunction with + {\tt Defined} (see \ref{Defined}) in order to define a + transparent constant. + +\item {\tt Let {\ident} : {\form}.} +\comindex{Let} + + Analogous to {\tt Definition} except that the definition is turned + into a local definition on objects depending on it after closing the + current section. +\end{Variants} + +\subsection{\tt Proof {\term}.}\comindex{Proof} +This command applies in proof editing mode. It is equivalent to {\tt + exact {\term}; Save.} That is, you have to give the full proof in +one gulp, as a proof term (see section \ref{exact}). + +\begin{Variants} + +\item{\tt Proof.} + + Is a noop which is useful to delimit the sequence of tactic commands + which start a proof, after a {\tt Theorem} command. It is a good + practice to use {\tt Proof.} as an opening parenthesis, closed in + the script with a closing {\tt Qed.} + +\item{\tt Proof with {\tac}.} + + This command may be used to start a proof. It defines a default + tactic to be used each time a tactic command is ended by + ``\verb#...#''. In this case the tactic command typed by the user is + equivalent to \emph{command};{\tac}. + +\end{Variants} + +\subsection{\tt Abort.} +\comindex{Abort} + +This command cancels the current proof development, switching back to +the previous proof development, or to the \Coq\ toplevel if no other +proof was edited. + +\begin{ErrMsgs} +\item \errindex{No focused proof (No proof-editing in progress)} +\end{ErrMsgs} + +\begin{Variants} + +\item {\tt Abort {\ident}.} + + Aborts the editing of the proof named {\ident}. + +\item {\tt Abort All.} + + Aborts all current goals, switching back to the \Coq\ toplevel. + +\end{Variants} + +\subsection{\tt Suspend.} +\comindex{Suspend} + +This command applies in proof editing mode. It switches back to the +\Coq\ toplevel, but without canceling the current proofs. + +\subsection{\tt Resume.} +\comindex{Resume}\label{Resume} + +This commands switches back to the editing of the last edited proof. + +\begin{ErrMsgs} +\item \errindex{No proof-editing in progress} +\end{ErrMsgs} + +\begin{Variants} + +\item {\tt Resume {\ident}.} + + Restarts the editing of the proof named {\ident}. This can be used + to navigate between currently edited proofs. + +\end{Variants} + +\begin{ErrMsgs} +\item \errindex{No such proof} +\end{ErrMsgs} + +\section{Navigation in the proof tree} + +\subsection{\tt Undo.} +\comindex{Undo} + +This command cancels the effect of the last tactic command. Thus, it +backtracks one step. + +\begin{ErrMsgs} +\item \errindex{No focused proof (No proof-editing in progress)} +\item \errindex{Undo stack would be exhausted} +\end{ErrMsgs} + +\begin{Variants} + +\item {\tt Undo {\num}.} + + Repeats {\tt Undo} {\num} times. + +\end{Variants} + +\subsection{\tt Set Undo {\num}.} +\comindex{Set Undo} + +This command changes the maximum number of {\tt Undo}'s that will be +possible when doing a proof. It only affects proofs started after +this command, such that if you want to change the current undo limit +inside a proof, you should first restart this proof. + +\subsection{\tt Unset Undo.} +\comindex{Unset Undo} + +This command resets the default number of possible {\tt Undo} commands +(which is currently 12). + +\subsection{\tt Restart.}\comindex{Restart} +This command restores the proof editing process to the original goal. + +\begin{ErrMsgs} +\item \errindex{No focused proof to restart} +\end{ErrMsgs} + +\subsection{\tt Focus.}\comindex{Focus} +This focuses the attention on the first subgoal to prove and the printing +of the other subgoals is suspended until the focused subgoal is +solved or unfocused. This is useful when there are many current +subgoals which clutter your screen. + +\begin{Variant} +\item {\tt Focus {\num}.}\\ +This focuses the attention on the $\num^{\scriptsize th}$ subgoal to prove. + +\end{Variant} + +\subsection{\tt Unfocus.}\comindex{Unfocus} +Turns off the focus mode. + + +\section{Displaying information} + +\subsection{\tt Show.}\comindex{Show}\label{Show} +This command displays the current goals. + +\begin{Variants} +\item {\tt Show {\num}.}\\ + Displays only the {\num}-th subgoal.\\ +\begin{ErrMsgs} +\item \errindex{No such goal} +\item \errindex{No focused proof} +\end{ErrMsgs} + +\item {\tt Show Implicits.}\comindex{Show Implicits}\\ + Displays the current goals, printing the implicit arguments of + constants. + +\item {\tt Show Implicits {\num}.}\\ + Same as above, only displaying the {\num}-th subgoal. + +\item {\tt Show Script.}\comindex{Show Script}\\ + Displays the whole list of tactics applied from the beginning + of the current proof. + This tactics script may contain some holes (subgoals not yet proved). + They are printed under the form \verb!!. + +\item {\tt Show Tree.}\comindex{Show Tree}\\ +This command can be seen as a more structured way of +displaying the state of the proof than that +provided by {\tt Show Script}. Instead of just giving +the list of tactics that have been applied, it +shows the derivation tree constructed by then. +Each node of the tree contains the conclusion +of the corresponding sub-derivation (i.e. a +goal with its corresponding local context) and +the tactic that has generated all the +sub-derivations. The leaves of this tree are +the goals which still remain to be proved. + +%\item {\tt Show Node}\comindex{Show Node}\\ +% Not yet documented + +\item {\tt Show Proof.}\comindex{Show Proof}\\ +It displays the proof term generated by the +tactics that have been applied. +If the proof is not completed, this term contain holes, +which correspond to the sub-terms which are still to be +constructed. These holes appear as a question mark indexed +by an integer, and applied to the list of variables in +the context, since it may depend on them. +The types obtained by abstracting away the context from the +type of each hole-placer are also printed. + +\item {\tt Show Conjectures.}\comindex{Show Conjectures}\\ +It prints the list of the names of all the theorems that +are currently being proved. +As it is possible to start proving a previous lemma during +the proof of a theorem, this list may contain several +names. + +\item{\tt Show Intro.}\comindex{Show Intro}\\ +If the current goal begins by at least one product, this command +prints the name of the first product, as it would be generated by +an anonymous {\tt Intro}. The aim of this command is to ease the +writing of more robust scripts. For example, with an appropriate +Proof General macro, it is possible to transform any anonymous {\tt + Intro} into a qualified one such as {\tt Intro y13}. +In the case of a non-product goal, it prints nothing. + +\item{\tt Show Intros.}\comindex{Show Intros}\\ +This command is similar to the previous one, it simulates the naming +process of an {\tt Intros}. + +\end{Variants} + +\subsection{\tt Set Hyps Limit {\num}.} +\comindex{Set Hyps Limit} +This command sets the maximum number of hypotheses displayed in +goals after the application of a tactic. +All the hypotheses remains usable in the proof development. + +\subsection{\tt Unset Hyps Limit.} +\comindex{Unset Hyps Limit} +This command goes back to the default mode which is to print all +available hypotheses. + +% $Id$ + +%%% Local Variables: +%%% mode: latex +%%% TeX-master: "Reference-Manual" +%%% End: diff --git a/doc/refman/RefMan-syn.tex b/doc/refman/RefMan-syn.tex new file mode 100755 index 0000000000..95f3d806b0 --- /dev/null +++ b/doc/refman/RefMan-syn.tex @@ -0,0 +1,1016 @@ +\chapter{Syntax extensions and interpretation scopes} +\label{Addoc-syntax} + +In this chapter, we introduce advanced commands to modify the way +{\Coq} parses and prints objects, i.e. the translations between the +concrete and internal representations of terms and commands. The main +commands are {\tt Notation} and {\tt Infix} which are described in +section \ref{Notation}. It also happens that the same symbolic +notation is expected in different contexts. To achieve this form of +overloading, {\Coq} offers a notion of interpretation scope. This is +described in section \ref{scopes}. + +\Rem The commands {\tt Grammar}, {\tt Syntax} and {\tt Distfix} which +were present for a while in {\Coq} are no longer available from {\Coq} +version 8.0. The underlying AST structure is also no longer available. +The functionalities of the command {\tt Syntactic Definition} are +still available, see section \ref{Abbreviations}. + +\section{Notations} +\label{Notation} +\comindex{Notation} + +\subsection{Basic notations} + +A {\em notation} is a symbolic abbreviation denoting some term +or term pattern. + +A typical notation is the use of the infix symbol \verb=/\= to denote +the logical conjunction (\texttt{and}). Such a notation is declared +by + +\begin{coq_example*} +Notation "A /\ B" := (and A B). +\end{coq_example*} + +The expression \texttt{(and A B)} is the abbreviated term and the +string \verb="A /\ B"= (called a {\em notation}) tells how it is +symbolically written. + +A notation is always surrounded by double quotes (excepted when the +abbreviation is a single ident, see \ref{Abbreviations}). The +notation is composed of {\em tokens} separated by spaces. Identifiers +in the string (such as \texttt{A} and \texttt{B}) are the {\em +parameters} of the notation. They must occur at least once each in the +denoted term. The other elements of the string (such as \verb=/\=) are +the {\em symbols}. + +An identifier can be used as a symbol but it must be surrounded by +simple quotes to avoid the confusion with a parameter. Similarly, +every symbol of at least 3 characters and starting with a simple quote +must be quoted (then it starts by two single quotes). Here is an example. + +\begin{coq_example*} +Notation "'IF' c1 'then' c2 'else' c3" := (IF_then_else c1 c2 c3). +\end{coq_example*} + +%TODO quote the identifier when not in front, not a keyword, as in "x 'U' y" ? + +A notation binds a syntactic expression to a term. Unless the parser +and pretty-printer of {\Coq} already know how to deal with the +syntactic expression (see \ref{ReservedNotation}), explicit precedences and +associativity rules have to be given. + +\subsection{Precedences and associativity} +\index{Precedences} +\index{Associativity} + +Mixing different symbolic notations in a same text may cause serious +parsing ambiguity. To deal with the ambiguity of notations, {\Coq} +uses precedence levels ranging from 0 to 100 (plus one extra level +numbered 200) and associativity rules. + +Consider for example the new notation + +\begin{coq_example*} +Notation "A \/ B" := (or A B). +\end{coq_example*} + +Clearly, an expression such as {\tt (A:Prop)True \verb=/\= A \verb=\/= +A \verb=\/= False} is ambiguous. To tell the {\Coq} parser how to +interpret the expression, a priority between the symbols \verb=/\= and +\verb=\/= has to be given. Assume for instance that we want conjunction +to bind more than disjunction. This is expressed by assigning a +precedence level to each notation, knowing that a lower level binds +more than a higher level. Hence the level for disjunction must be +higher than the level for conjunction. + +Since connectives are the less tight articulation points of a text, it +is reasonable to choose levels not so far from the higher level which +is 100, for example 85 for disjunction and 80 for +conjunction\footnote{which are the levels effectively chosen in the +current implementation of {\Coq}}. + +Similarly, an associativity is needed to decide whether {\tt True \verb=/\= +False \verb=/\= False} defaults to {\tt True \verb=/\= (False +\verb=/\= False)} (right associativity) or to {\tt (True +\verb=/\= False) \verb=/\= False} (left associativity). We may +even consider that the expression is not well-formed and that +parentheses are mandatory (this is a ``no associativity'')\footnote{ +{\Coq} accepts notations declared as no associative but the parser on +which {\Coq} is built, namely {\camlpppp}, currently does not implement the +no-associativity and replace it by a left associativity; hence it is +the same for {\Coq}: no-associativity is in fact left associativity}. +We don't know of a special convention of the associativity of +disjunction and conjunction, let's apply for instance a right +associativity (which is the choice of {\Coq}). + +Precedence levels and associativity rules of notations have to be +given between parentheses in a list of modifiers that the +\texttt{Notation} command understands. Here is how the previous +examples refine. + +\begin{coq_example*} +Notation "A /\ B" := (and A B) (at level 80, right associativity). +Notation "A \/ B" := (or A B) (at level 85, right associativity). +\end{coq_example*} + +By default, a notation is considered non associative, but the +precedence level is mandatory (except for special cases whose level is +canonical). The level is either a number or the mention {\tt next +level} whose meaning is obvious. The list of levels already assigned +is on Figure~\ref{init-notations}. + +\subsection{Complex notations} + +Notations can be made from arbitraly complex symbols. One can for +instance define prefix notations. + +\begin{coq_example*} +Notation "~ x" := (not x) (at level 75, right associativity). +\end{coq_example*} + +One can also define notations for incomplete terms, with the hole +expected to be inferred at typing time. + +\begin{coq_example*} +Notation "x = y" := (@eq _ x y) (at level 70, no associativity). +\end{coq_example*} + +One can define {\em closed} notations whose both sides are symbols. In +this case, the default precedence level for inner subexpression is 200. + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is correct but produces **********) +(**** an incompatibility with the reserved notation ********) +\end{coq_eval} +\begin{coq_example*} +Notation "( x , y )" := (@pair _ _ x y) (at level 0). +\end{coq_example*} + +One can also define notations for binders. + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is correct but produces **********) +(**** an incompatibility with the reserved notation ********) +\end{coq_eval} +\begin{coq_example*} +Notation "{ x : A | P }" := (sig A (fun x => P)) (at level 0). +\end{coq_example*} + +In the last case though, there is a conflict with the notation for +type casts. This last notation, as shown by the command {\tt Print Grammar +constr} is at level 100. To avoid \verb=x : A= being parsed as a type cast, +it is necessary to put {\tt x} at a level below 100, typically 99. Hence, a +correct definition is + +\begin{coq_example*} +Notation "{ x : A | P }" := (sig A (fun x => P)) (at level 0, x at level 99). +\end{coq_example*} + +%This change has retrospectively an effect on the notation for notation +%{\tt "{ A } + { B }"}. For the sake of factorization, {\tt A} must be +%put at level 99 too, which gives +% +%\begin{coq_example*} +%Notation "{ A } + { B }" := (sumbool A B) (at level 0, A at level 99). +%\end{coq_example*} + +See the next section for more about factorization. + +\subsection{Simple factorization rules} + +{\Coq} extensible parsing is performed by Camlp4 which is essentially a +LL1 parser. Hence, some care has to be taken not to hide already +existing rules by new rules. Some simple left factorization work has +to be done. Here is an example. + +\begin{coq_eval} +(********** The next rule for notation _ < _ < _ produces **********) +(*** Error: Notation _ < _ < _ is already defined at level 70 ... ***) +\end{coq_eval} +\begin{coq_example*} +Notation "x < y" := (lt x y) (at level 70). +Notation "x < y < z" := (x < y /\ y < z) (at level 70). +\end{coq_example*} + +In order to factorize the left part of the rules, the subexpression +referred by {\tt y} has to be at the same level in both rules. However +the default behavior puts {\tt y} at the next level below 70 +in the first rule (no associativity is the default), and at the level +200 in the second rule (level 200 is the default for inner expressions). +To fix this, we need to force the parsing level of {\tt y}, +as follows. + +\begin{coq_example*} +Notation "x < y" := (lt x y) (at level 70). +Notation "x < y < z" := (x < y /\ y < z) (at level 70, y at next level). +\end{coq_example*} + +For the sake of factorization with {\Coq} predefined rules, simple +rules have to be observed for notations starting with a symbol: +e.g. rules starting with ``\{'' or ``('' should be put at level 0. The +list of {\Coq} predefined notations can be found in chapter \ref{Theories}. + +The command to display the current state of the {\Coq} term parser is +\comindex{Print Grammar constr} + +\begin{quote} +\tt Print Grammar constr. +\end{quote} + +\subsection{Displaying symbolic notations} + +The command \texttt{Notation} has an effect both on the {\Coq} parser and +on the {\Coq} printer. For example: + +\begin{coq_example} +Check (and True True). +\end{coq_example} + +However, printing, especially pretty-printing, requires +more care than parsing. We may want specific indentations, +line breaks, alignment if on several lines, etc. + +The default printing of notations is very rudimentary. For printing a +notation, a {\em formatting box} is opened in such a way that if the +notation and its arguments cannot fit on a single line, a line break +is inserted before the symbols of the notation and the arguments on +the next lines are aligned with the argument on the first line. + +A first, simple control that a user can have on the printing of a +notation is the insertion of spaces at some places of the +notation. This is performed by adding extra spaces between the symbols +and parameters: each extra space (other than the single space needed +to separate the components) is interpreted as a space to be inserted +by the printer. Here is an example showing how to add spaces around +the bar of the notation. + +\begin{coq_example} +Notation "{{ x : A | P }}" := (sig (fun x : A => P)) + (at level 0, x at level 99). +Check (sig (fun x : nat => x=x)). +\end{coq_example} + +The second, more powerful control on printing is by using the {\tt +format} modifier. Here is an example + +\begin{small} +\begin{coq_example} +Notation "'If' c1 'then' c2 'else' c3" := (IF_then_else c1 c2 c3) +(at level 200, right associativity, format +"'[v ' 'If' c1 '/' '[' 'then' c2 ']' '/' '[' 'else' c3 ']' ']'"). +\end{coq_example} +\end{small} + +A {\em format} is an extension of the string denoting the notation with +the possible following elements delimited by single quotes: + +\begin{itemize} +\item extra spaces are translated into simple spaces +\item tokens of the form \verb='/ '= are translated into breaking point, + in case a line break occurs, an indentation of the number of spaces + after the ``\verb=/='' is applied (2 spaces in the given example) +\item token of the form \verb='//'= force writing on a new line +\item well-bracketed pairs of tokens of the form \verb='[ '= and \verb=']'= + are translated into printing boxes; in case a line break occurs, + an extra indentation of the number of spaces given after the ``\verb=[='' + is applied (4 spaces in the example) +\item well-bracketed pairs of tokens of the form \verb='[hv '= and \verb=']'= + are translated into horizontal-orelse-vertical printing boxes; + if the content of the box does not fit on a single line, then every breaking + point forces a newline and an extra indentation of the number of spaces + given after the ``\verb=[='' is applied at the beginning of each newline + (3 spaces in the example) +\item well-bracketed pairs of tokens of the form \verb='[v '= and + \verb=']'= are translated into vertical printing boxes; every + breaking point forces a newline, even if the line is large enough to + display the whole content of the box, and an extra indentation of the + number of spaces given after the ``\verb=[='' is applied at the beginning + of each newline +\end{itemize} + +Thus, for the previous example, we get +%\footnote{The ``@'' is here to shunt +%the notation "'IF' A 'then' B 'else' C" which is defined in {\Coq} +%initial state}: + +Notations do not survive the end of sections. No typing of the denoted +expression is performed at definition time. Type-checking is done only +at the time of use of the notation. + +\begin{coq_example} +Check + (IF_then_else (IF_then_else True False True) + (IF_then_else True False True) + (IF_then_else True False True)). +\end{coq_example} + +\Rem +Sometimes, a notation is expected only for the parser. +%(e.g. because +%the underlying parser of {\Coq}, namely {\camlpppp}, is LL1 and some extra +%rules are needed to circumvent the absence of factorization). +To do so, the option {\em only parsing} is allowed in the list of modifiers of +\texttt{Notation}. + +\subsection{The \texttt{Infix} command +\comindex{Infix}} + +The \texttt{Infix} command is a shortening for declaring notations of +infix symbols. Its syntax is + +\begin{quote} +\noindent\texttt{Infix "{\symbolentry}" :=} {\qualid} {\tt (} \nelist{\em modifier}{,} {\tt )}. +\end{quote} + +and it is equivalent to + +\begin{quote} +\noindent\texttt{Notation "x {\symbolentry} y" := ({\qualid} x y) (} \nelist{\em modifier}{,} {\tt )}. +\end{quote} + +where {\tt x} and {\tt y} are fresh names distinct from {\qualid}. Here is an example. + +\begin{coq_example*} +Infix "/\" := and (at level 80, right associativity). +\end{coq_example*} + +\subsection{Reserving notations +\label{ReservedNotation} +\comindex{ReservedNotation}} + +A given notation may be used in different contexts. {\Coq} expects all +uses of the notation to be defined at the same precedence and with the +same associativity. To avoid giving the precedence and associativity +every time, it is possible to declare a parsing rule in advance +without giving its interpretation. Here is an example from the initial +state of {\Coq}. + +\begin{coq_example} +Reserved Notation "x = y" (at level 70, no associativity). +\end{coq_example} + +Reserving a notation is also useful for simultaneously defined an +inductive type or a recursive constant and a notation for it. + +\Rem The notations mentioned on Figure~\ref{init-notations} are +reserved. Hence their precedence and associativity cannot be changed. + +\subsection{Simultaneous definition of terms and notations +\comindex{Fixpoint {\ldots} where {\ldots}} +\comindex{CoFixpoint {\ldots} where {\ldots}} +\comindex{Inductive {\ldots} where {\ldots}}} + +Thanks to reserved notations, the inductive, coinductive, recursive +and corecursive definitions can benefit of customized notations. To do +this, insert a {\tt where} notation clause after the definition of the +(co)inductive type or (co)recursive term (or after the definition of +each of them in case of mutual definitions). The exact syntax is given +on Figure \ref{notation-syntax}. Here are examples: + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is correct but produces an error **********) +(********** because the symbol /\ is already bound **********) +(**** Error: The conclusion of A -> B -> A /\ B is not valid *****) +\end{coq_eval} + +\begin{coq_example*} +Inductive and (A B:Prop) : Prop := conj : A -> B -> A /\ B +where "A /\ B" := (and A B). +\end{coq_example*} + +\begin{coq_eval} +Set Printing Depth 50. +(********** The following is correct but produces an error **********) +(********** because the symbol + is already bound **********) +(**** Error: no recursive definition *****) +\end{coq_eval} + +\begin{coq_example*} +Fixpoint plus (n m:nat) {struct n} : nat := + match n with + | O => m + | S p => S (p+m) + end +where "n + m" := (plus n m). +\end{coq_example*} + +\subsection{Displaying informations about notations +\comindex{Set Printing Notations} +\comindex{Unset Printing Notations}} + +To deactivate the printing of all notations, use the command +\begin{quote} +\tt Unset Printing Notations. +\end{quote} +To reactivate it, use the command +\begin{quote} +\tt Set Printing Notations. +\end{quote} +The default is to use notations for printing terms wherever possible. + +\SeeAlso {\tt Set Printing All} in section \ref{SetPrintingAll}. + +\subsection{Locating notations +\comindex{Locate} +\label{LocateSymbol}} + +To know to which notations a given symbol belongs to, use the command +\begin{quote} +\tt Locate {\symbolentry} +\end{quote} +where symbol is any (composite) symbol surrounded by quotes. To locate +a particular notation, use a string where the variables of the +notation are replaced by ``\_''. + +\Example +\begin{coq_example} +Locate "exists". +Locate "'exists' _ , _". +\end{coq_example} + +\SeeAlso Section \ref{Locate}. + +\begin{figure} +\begin{centerframe} +\begin{tabular}{lcl} +{\sentence} & ::= & + \texttt{Notation} \zeroone{\tt Local} {\str} \texttt{:=} {\term} + \zeroone{\modifiers} \zeroone{:{\scope}} .\\ + & $|$ & + \texttt{Infix} \zeroone{\tt Local} {\str} \texttt{:=} {\qualid} + \zeroone{\modifiers} \zeroone{:{\scope}} .\\ + & $|$ & + \texttt{Reserved Notation} \zeroone{\tt Local} {\str} + \zeroone{\modifiers} .\\ + & $|$ & {\tt Inductive} + \nelist{{\inductivebody} \zeroone{\declnotation}}{with}{\tt .}\\ + & $|$ & {\tt CoInductive} + \nelist{{\inductivebody} \zeroone{\declnotation}}{with}{\tt .}\\ + & $|$ & {\tt Fixpoint} + \nelist{{\fixpointbody} \zeroone{\declnotation}}{with} {\tt .} \\ + & $|$ & {\tt CoFixpoint} + \nelist{{\cofixpointbody} \zeroone{\declnotation}}{with} {\tt .} \\ +\\ +{\declnotation} & ::= & + \zeroone{{\tt where} {\str} {\tt :=} {\term} \zeroone{:{\scope}}} . +\\ +\\ +{\modifiers} + & ::= & \nelist{\ident}{,} {\tt at level} {\naturalnumber} \\ + & $|$ & \nelist{\ident}{,} {\tt at next level} \\ + & $|$ & {\tt at level} {\naturalnumber} \\ + & $|$ & {\tt left associativity} \\ + & $|$ & {\tt right associativity} \\ + & $|$ & {\tt no associativity} \\ + & $|$ & {\ident} {\tt ident} \\ + & $|$ & {\ident} {\tt global} \\ + & $|$ & {\ident} {\tt bigint} \\ + & $|$ & {\tt only parsing} \\ + & $|$ & {\tt format} {\str} +\end{tabular} +\end{centerframe} +\caption{Syntax of the variants of {\tt Notation}} +\label{notation-syntax} +\end{figure} + +\subsection{Notations with recursive patterns} + +An experimental mechanism is provided for declaring elementary +notations including recursive patterns. The basic syntax is + +\begin{coq_eval} +Require Import List. +\end{coq_eval} + +\begin{coq_example*} +Notation "[ x ; .. ; y ]" := (cons x .. (cons y nil) ..). +\end{coq_example*} + +On the right-hand-side, an extra construction of the form {\tt ..} ($f$ +$t_1$ $\ldots$ $t_n$) {\tt ..} can be used. Notice that {\tt ..} is part of +the {\Coq} syntax while $\ldots$ is just a meta-notation of this +manual to denote a sequence of terms of arbitrary size. + +This extra construction enclosed within {\tt ..}, let's call it $t$, +must be one of the argument of an applicative term of the form {\tt +($f$ $u_1$ $\ldots$ $u_n$)}. The sequences $t_1$ $\ldots$ $t_n$ and +$u_1$ $\ldots$ $u_n$ must coincide everywhere but in two places. In +one place, say the terms of indice $i$, we must have $u_i = t$. In the +other place, say the terms of indice $j$, both $u_j$ and $t_j$ must be +variables, say $x$ and $y$ which are bound by the notation string on +the left-hand-side of the declaration. The variables $x$ and $y$ in +the string must occur in a substring of the form "$x$ $s$ {\tt ..} $s$ +$y$" where {\tt ..} is part of the syntax and $s$ is two times the +same sequence of terminal symbols (i.e. symbols which are not +variables). + +These invariants must be satisfied in order the notation to be +correct. The term $t_i$ is the {\em terminating} expression of +the notation and the pattern {\tt ($f$ $u_1$ $\ldots$ $u_{i-1}$ {\rm [I]} +$u_{i+1}$ $\ldots$ $u_{j-1}$ {\rm [E]} $u_{j+1}$ $\ldots$ $u_{n}$)} is the +{\em iterating pattern}. The hole [I] is the {\em iterative} place +and the hole [E] is the {\em enumerating} place. Remark that if $j