5 chapter {* Isabelle/HOL \label{ch:hol} *}
7 section {* Typedef axiomatization \label{sec:hol-typedef} *}
10 \begin{matharray}{rcl}
11 @{command_def (HOL) "typedef"} & : & @{text "local_theory \<rightarrow> proof(prove)"} \\
15 'typedef' altname? abstype '=' repset
18 altname: '(' (name | 'open' | 'open' name) ')'
20 abstype: typespecsorts mixfix?
22 repset: term ('morphisms' name name)?
28 \item @{command (HOL) "typedef"}~@{text "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>n) t = A"}
29 axiomatizes a Gordon/HOL-style type definition in the background
30 theory of the current context, depending on a non-emptiness result
31 of the set @{text A} (which needs to be proven interactively).
33 The raw type may not depend on parameters or assumptions of the
34 context --- this is logically impossible in Isabelle/HOL --- but the
35 non-emptiness property can be local, potentially resulting in
36 multiple interpretations in target contexts. Thus the established
37 bijection between the representing set @{text A} and the new type
38 @{text t} may semantically depend on local assumptions.
40 By default, @{command (HOL) "typedef"} defines both a type @{text t}
41 and a set (term constant) of the same name, unless an alternative
42 base name is given in parentheses, or the ``@{text "(open)"}''
43 declaration is used to suppress a separate constant definition
44 altogether. The injection from type to set is called @{text Rep_t},
45 its inverse @{text Abs_t} --- this may be changed via an explicit
46 @{keyword (HOL) "morphisms"} declaration.
48 Theorems @{text Rep_t}, @{text Rep_t_inverse}, and @{text
49 Abs_t_inverse} provide the most basic characterization as a
50 corresponding injection/surjection pair (in both directions). Rules
51 @{text Rep_t_inject} and @{text Abs_t_inject} provide a slightly
52 more convenient view on the injectivity part, suitable for automated
53 proof tools (e.g.\ in @{attribute simp} or @{attribute iff}
54 declarations). Rules @{text Rep_t_cases}/@{text Rep_t_induct}, and
55 @{text Abs_t_cases}/@{text Abs_t_induct} provide alternative views
56 on surjectivity; these are already declared as set or type rules for
57 the generic @{method cases} and @{method induct} methods.
59 An alternative name for the set definition (and other derived
60 entities) may be specified in parentheses; the default is to use
61 @{text t} as indicated before.
67 section {* Adhoc tuples *}
70 \begin{matharray}{rcl}
71 @{attribute (HOL) split_format}@{text "\<^sup>*"} & : & @{text attribute} \\
75 'split_format' '(' 'complete' ')'
81 \item @{attribute (HOL) split_format}\ @{text "(complete)"} causes
82 arguments in function applications to be represented canonically
83 according to their tuple type structure.
85 Note that this operation tends to invent funny names for new local
86 parameters introduced.
92 section {* Records \label{sec:hol-record} *}
95 In principle, records merely generalize the concept of tuples, where
96 components may be addressed by labels instead of just position. The
97 logical infrastructure of records in Isabelle/HOL is slightly more
98 advanced, though, supporting truly extensible record schemes. This
99 admits operations that are polymorphic with respect to record
100 extension, yielding ``object-oriented'' effects like (single)
101 inheritance. See also \cite{NaraschewskiW-TPHOLs98} for more
102 details on object-oriented verification and record subtyping in HOL.
106 subsection {* Basic concepts *}
109 Isabelle/HOL supports both \emph{fixed} and \emph{schematic} records
110 at the level of terms and types. The notation is as follows:
113 \begin{tabular}{l|l|l}
114 & record terms & record types \\ \hline
115 fixed & @{text "\<lparr>x = a, y = b\<rparr>"} & @{text "\<lparr>x :: A, y :: B\<rparr>"} \\
116 schematic & @{text "\<lparr>x = a, y = b, \<dots> = m\<rparr>"} &
117 @{text "\<lparr>x :: A, y :: B, \<dots> :: M\<rparr>"} \\
121 \noindent The ASCII representation of @{text "\<lparr>x = a\<rparr>"} is @{text
124 A fixed record @{text "\<lparr>x = a, y = b\<rparr>"} has field @{text x} of value
125 @{text a} and field @{text y} of value @{text b}. The corresponding
126 type is @{text "\<lparr>x :: A, y :: B\<rparr>"}, assuming that @{text "a :: A"}
127 and @{text "b :: B"}.
129 A record scheme like @{text "\<lparr>x = a, y = b, \<dots> = m\<rparr>"} contains fields
130 @{text x} and @{text y} as before, but also possibly further fields
131 as indicated by the ``@{text "\<dots>"}'' notation (which is actually part
132 of the syntax). The improper field ``@{text "\<dots>"}'' of a record
133 scheme is called the \emph{more part}. Logically it is just a free
134 variable, which is occasionally referred to as ``row variable'' in
135 the literature. The more part of a record scheme may be
136 instantiated by zero or more further components. For example, the
137 previous scheme may get instantiated to @{text "\<lparr>x = a, y = b, z =
138 c, \<dots> = m'\<rparr>"}, where @{text m'} refers to a different more part.
139 Fixed records are special instances of record schemes, where
140 ``@{text "\<dots>"}'' is properly terminated by the @{text "() :: unit"}
141 element. In fact, @{text "\<lparr>x = a, y = b\<rparr>"} is just an abbreviation
142 for @{text "\<lparr>x = a, y = b, \<dots> = ()\<rparr>"}.
144 \medskip Two key observations make extensible records in a simply
145 typed language like HOL work out:
149 \item the more part is internalized, as a free term or type
152 \item field names are externalized, they cannot be accessed within
153 the logic as first-class values.
157 \medskip In Isabelle/HOL record types have to be defined explicitly,
158 fixing their field names and types, and their (optional) parent
159 record. Afterwards, records may be formed using above syntax, while
160 obeying the canonical order of fields as given by their declaration.
161 The record package provides several standard operations like
162 selectors and updates. The common setup for various generic proof
163 tools enable succinct reasoning patterns. See also the Isabelle/HOL
164 tutorial \cite{isabelle-hol-book} for further instructions on using
169 subsection {* Record specifications *}
172 \begin{matharray}{rcl}
173 @{command_def (HOL) "record"} & : & @{text "theory \<rightarrow> theory"} \\
177 'record' typespecsorts '=' (type '+')? (constdecl +)
183 \item @{command (HOL) "record"}~@{text "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>m) t = \<tau> + c\<^sub>1 :: \<sigma>\<^sub>1
184 \<dots> c\<^sub>n :: \<sigma>\<^sub>n"} defines extensible record type @{text "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>m) t"},
185 derived from the optional parent record @{text "\<tau>"} by adding new
186 field components @{text "c\<^sub>i :: \<sigma>\<^sub>i"} etc.
188 The type variables of @{text "\<tau>"} and @{text "\<sigma>\<^sub>i"} need to be
189 covered by the (distinct) parameters @{text "\<alpha>\<^sub>1, \<dots>,
190 \<alpha>\<^sub>m"}. Type constructor @{text t} has to be new, while @{text
191 \<tau>} needs to specify an instance of an existing record type. At
192 least one new field @{text "c\<^sub>i"} has to be specified.
193 Basically, field names need to belong to a unique record. This is
194 not a real restriction in practice, since fields are qualified by
195 the record name internally.
197 The parent record specification @{text \<tau>} is optional; if omitted
198 @{text t} becomes a root record. The hierarchy of all records
199 declared within a theory context forms a forest structure, i.e.\ a
200 set of trees starting with a root record each. There is no way to
201 merge multiple parent records!
203 For convenience, @{text "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>m) t"} is made a
204 type abbreviation for the fixed record type @{text "\<lparr>c\<^sub>1 ::
205 \<sigma>\<^sub>1, \<dots>, c\<^sub>n :: \<sigma>\<^sub>n\<rparr>"}, likewise is @{text
206 "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>m, \<zeta>) t_scheme"} made an abbreviation for
207 @{text "\<lparr>c\<^sub>1 :: \<sigma>\<^sub>1, \<dots>, c\<^sub>n :: \<sigma>\<^sub>n, \<dots> ::
214 subsection {* Record operations *}
217 Any record definition of the form presented above produces certain
218 standard operations. Selectors and updates are provided for any
219 field, including the improper one ``@{text more}''. There are also
220 cumulative record constructor functions. To simplify the
221 presentation below, we assume for now that @{text "(\<alpha>\<^sub>1, \<dots>,
222 \<alpha>\<^sub>m) t"} is a root record with fields @{text "c\<^sub>1 ::
223 \<sigma>\<^sub>1, \<dots>, c\<^sub>n :: \<sigma>\<^sub>n"}.
225 \medskip \textbf{Selectors} and \textbf{updates} are available for
226 any field (including ``@{text more}''):
228 \begin{matharray}{lll}
229 @{text "c\<^sub>i"} & @{text "::"} & @{text "\<lparr>\<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr> \<Rightarrow> \<sigma>\<^sub>i"} \\
230 @{text "c\<^sub>i_update"} & @{text "::"} & @{text "\<sigma>\<^sub>i \<Rightarrow> \<lparr>\<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr> \<Rightarrow> \<lparr>\<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr>"} \\
233 There is special syntax for application of updates: @{text "r\<lparr>x :=
234 a\<rparr>"} abbreviates term @{text "x_update a r"}. Further notation for
235 repeated updates is also available: @{text "r\<lparr>x := a\<rparr>\<lparr>y := b\<rparr>\<lparr>z :=
236 c\<rparr>"} may be written @{text "r\<lparr>x := a, y := b, z := c\<rparr>"}. Note that
237 because of postfix notation the order of fields shown here is
238 reverse than in the actual term. Since repeated updates are just
239 function applications, fields may be freely permuted in @{text "\<lparr>x
240 := a, y := b, z := c\<rparr>"}, as far as logical equality is concerned.
241 Thus commutativity of independent updates can be proven within the
242 logic for any two fields, but not as a general theorem.
244 \medskip The \textbf{make} operation provides a cumulative record
245 constructor function:
247 \begin{matharray}{lll}
248 @{text "t.make"} & @{text "::"} & @{text "\<sigma>\<^sub>1 \<Rightarrow> \<dots> \<sigma>\<^sub>n \<Rightarrow> \<lparr>\<^vec>c :: \<^vec>\<sigma>\<rparr>"} \\
251 \medskip We now reconsider the case of non-root records, which are
252 derived of some parent. In general, the latter may depend on
253 another parent as well, resulting in a list of \emph{ancestor
254 records}. Appending the lists of fields of all ancestors results in
255 a certain field prefix. The record package automatically takes care
256 of this by lifting operations over this context of ancestor fields.
257 Assuming that @{text "(\<alpha>\<^sub>1, \<dots>, \<alpha>\<^sub>m) t"} has ancestor
258 fields @{text "b\<^sub>1 :: \<rho>\<^sub>1, \<dots>, b\<^sub>k :: \<rho>\<^sub>k"},
259 the above record operations will get the following types:
263 @{text "c\<^sub>i"} & @{text "::"} & @{text "\<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr> \<Rightarrow> \<sigma>\<^sub>i"} \\
264 @{text "c\<^sub>i_update"} & @{text "::"} & @{text "\<sigma>\<^sub>i \<Rightarrow>
265 \<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr> \<Rightarrow>
266 \<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr>"} \\
267 @{text "t.make"} & @{text "::"} & @{text "\<rho>\<^sub>1 \<Rightarrow> \<dots> \<rho>\<^sub>k \<Rightarrow> \<sigma>\<^sub>1 \<Rightarrow> \<dots> \<sigma>\<^sub>n \<Rightarrow>
268 \<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>\<rparr>"} \\
272 \noindent Some further operations address the extension aspect of a
273 derived record scheme specifically: @{text "t.fields"} produces a
274 record fragment consisting of exactly the new fields introduced here
275 (the result may serve as a more part elsewhere); @{text "t.extend"}
276 takes a fixed record and adds a given more part; @{text
277 "t.truncate"} restricts a record scheme to a fixed record.
281 @{text "t.fields"} & @{text "::"} & @{text "\<sigma>\<^sub>1 \<Rightarrow> \<dots> \<sigma>\<^sub>n \<Rightarrow> \<lparr>\<^vec>c :: \<^vec>\<sigma>\<rparr>"} \\
282 @{text "t.extend"} & @{text "::"} & @{text "\<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>\<rparr> \<Rightarrow>
283 \<zeta> \<Rightarrow> \<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr>"} \\
284 @{text "t.truncate"} & @{text "::"} & @{text "\<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>, \<dots> :: \<zeta>\<rparr> \<Rightarrow> \<lparr>\<^vec>b :: \<^vec>\<rho>, \<^vec>c :: \<^vec>\<sigma>\<rparr>"} \\
288 \noindent Note that @{text "t.make"} and @{text "t.fields"} coincide
293 subsection {* Derived rules and proof tools *}
296 The record package proves several results internally, declaring
297 these facts to appropriate proof tools. This enables users to
298 reason about record structures quite conveniently. Assume that
299 @{text t} is a record type as specified above.
303 \item Standard conversions for selectors or updates applied to
304 record constructor terms are made part of the default Simplifier
305 context; thus proofs by reduction of basic operations merely require
306 the @{method simp} method without further arguments. These rules
307 are available as @{text "t.simps"}, too.
309 \item Selectors applied to updated records are automatically reduced
310 by an internal simplification procedure, which is also part of the
311 standard Simplifier setup.
313 \item Inject equations of a form analogous to @{prop "(x, y) = (x',
314 y') \<equiv> x = x' \<and> y = y'"} are declared to the Simplifier and Classical
315 Reasoner as @{attribute iff} rules. These rules are available as
318 \item The introduction rule for record equality analogous to @{text
319 "x r = x r' \<Longrightarrow> y r = y r' \<dots> \<Longrightarrow> r = r'"} is declared to the Simplifier,
320 and as the basic rule context as ``@{attribute intro}@{text "?"}''.
321 The rule is called @{text "t.equality"}.
323 \item Representations of arbitrary record expressions as canonical
324 constructor terms are provided both in @{method cases} and @{method
325 induct} format (cf.\ the generic proof methods of the same name,
326 \secref{sec:cases-induct}). Several variations are available, for
327 fixed records, record schemes, more parts etc.
329 The generic proof methods are sufficiently smart to pick the most
330 sensible rule according to the type of the indicated record
331 expression: users just need to apply something like ``@{text "(cases
332 r)"}'' to a certain proof problem.
334 \item The derived record operations @{text "t.make"}, @{text
335 "t.fields"}, @{text "t.extend"}, @{text "t.truncate"} are \emph{not}
336 treated automatically, but usually need to be expanded by hand,
337 using the collective fact @{text "t.defs"}.
343 section {* Datatypes \label{sec:hol-datatype} *}
346 \begin{matharray}{rcl}
347 @{command_def (HOL) "datatype"} & : & @{text "theory \<rightarrow> theory"} \\
348 @{command_def (HOL) "rep_datatype"} & : & @{text "theory \<rightarrow> proof(prove)"} \\
352 'datatype' (dtspec + 'and')
354 'rep_datatype' ('(' (name +) ')')? (term +)
357 dtspec: parname? typespec mixfix? '=' (cons + '|')
359 cons: name ( type * ) mixfix?
364 \item @{command (HOL) "datatype"} defines inductive datatypes in
367 \item @{command (HOL) "rep_datatype"} represents existing types as
368 inductive ones, generating the standard infrastructure of derived
369 concepts (primitive recursion etc.).
373 The induction and exhaustion theorems generated provide case names
374 according to the constructors involved, while parameters are named
375 after the types (see also \secref{sec:cases-induct}).
377 See \cite{isabelle-HOL} for more details on datatypes, but beware of
378 the old-style theory syntax being used there! Apart from proper
379 proof methods for case-analysis and induction, there are also
380 emulations of ML tactics @{method (HOL) case_tac} and @{method (HOL)
381 induct_tac} available, see \secref{sec:hol-induct-tac}; these admit
382 to refer directly to the internal structure of subgoals (including
383 internally bound parameters).
387 section {* Functorial structure of types *}
390 \begin{matharray}{rcl}
391 @{command_def (HOL) "enriched_type"} & : & @{text "local_theory \<rightarrow> proof(prove)"}
395 'enriched_type' (prefix ':')? term
401 \item @{command (HOL) "enriched_type"} allows to prove and register
402 properties about the functorial structure of type constructors;
403 these properties then can be used by other packages to
404 deal with those type constructors in certain type constructions.
405 Characteristic theorems are noted in the current local theory; by
406 default, they are prefixed with the base name of the type constructor,
407 an explicit prefix can be given alternatively.
409 The given term @{text "m"} is considered as \emph{mapper} for the
410 corresponding type constructor and must conform to the following
413 \begin{matharray}{lll}
414 @{text "m"} & @{text "::"} &
415 @{text "\<sigma>\<^isub>1 \<Rightarrow> \<dots> \<sigma>\<^isub>k \<Rightarrow> (\<^vec>\<alpha>\<^isub>n) t \<Rightarrow> (\<^vec>\<beta>\<^isub>n) t"} \\
418 \noindent where @{text t} is the type constructor, @{text
419 "\<^vec>\<alpha>\<^isub>n"} and @{text "\<^vec>\<beta>\<^isub>n"} are distinct
420 type variables free in the local theory and @{text "\<sigma>\<^isub>1"},
421 \ldots, @{text "\<sigma>\<^isub>k"} is a subsequence of @{text "\<alpha>\<^isub>1 \<Rightarrow>
422 \<beta>\<^isub>1"}, @{text "\<beta>\<^isub>1 \<Rightarrow> \<alpha>\<^isub>1"}, \ldots,
423 @{text "\<alpha>\<^isub>n \<Rightarrow> \<beta>\<^isub>n"}, @{text "\<beta>\<^isub>n \<Rightarrow>
430 section {* Recursive functions \label{sec:recursion} *}
433 \begin{matharray}{rcl}
434 @{command_def (HOL) "primrec"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
435 @{command_def (HOL) "fun"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
436 @{command_def (HOL) "function"} & : & @{text "local_theory \<rightarrow> proof(prove)"} \\
437 @{command_def (HOL) "termination"} & : & @{text "local_theory \<rightarrow> proof(prove)"} \\
441 'primrec' target? fixes 'where' equations
443 ('fun' | 'function') target? functionopts? fixes \\ 'where' equations
445 equations: (thmdecl? prop + '|')
447 functionopts: '(' (('sequential' | 'domintros' | 'tailrec' | 'default' term) + ',') ')'
449 'termination' ( term )?
454 \item @{command (HOL) "primrec"} defines primitive recursive
455 functions over datatypes, see also \cite{isabelle-HOL}.
457 \item @{command (HOL) "function"} defines functions by general
458 wellfounded recursion. A detailed description with examples can be
459 found in \cite{isabelle-function}. The function is specified by a
460 set of (possibly conditional) recursive equations with arbitrary
461 pattern matching. The command generates proof obligations for the
462 completeness and the compatibility of patterns.
464 The defined function is considered partial, and the resulting
465 simplification rules (named @{text "f.psimps"}) and induction rule
466 (named @{text "f.pinduct"}) are guarded by a generated domain
467 predicate @{text "f_dom"}. The @{command (HOL) "termination"}
468 command can then be used to establish that the function is total.
470 \item @{command (HOL) "fun"} is a shorthand notation for ``@{command
471 (HOL) "function"}~@{text "(sequential)"}, followed by automated
472 proof attempts regarding pattern matching and termination. See
473 \cite{isabelle-function} for further details.
475 \item @{command (HOL) "termination"}~@{text f} commences a
476 termination proof for the previously defined function @{text f}. If
477 this is omitted, the command refers to the most recent function
478 definition. After the proof is closed, the recursive equations and
479 the induction principle is established.
483 Recursive definitions introduced by the @{command (HOL) "function"}
485 reasoning by induction (cf.\ \secref{sec:cases-induct}): rule @{text
486 "c.induct"} (where @{text c} is the name of the function definition)
487 refers to a specific induction rule, with parameters named according
488 to the user-specified equations. Cases are numbered (starting from 1).
490 For @{command (HOL) "primrec"}, the induction principle coincides
491 with structural recursion on the datatype the recursion is carried
494 The equations provided by these packages may be referred later as
495 theorem list @{text "f.simps"}, where @{text f} is the (collective)
496 name of the functions defined. Individual equations may be named
499 The @{command (HOL) "function"} command accepts the following
504 \item @{text sequential} enables a preprocessor which disambiguates
505 overlapping patterns by making them mutually disjoint. Earlier
506 equations take precedence over later ones. This allows to give the
507 specification in a format very similar to functional programming.
508 Note that the resulting simplification and induction rules
509 correspond to the transformed specification, not the one given
510 originally. This usually means that each equation given by the user
511 may result in several theorems. Also note that this automatic
512 transformation only works for ML-style datatype patterns.
514 \item @{text domintros} enables the automated generation of
515 introduction rules for the domain predicate. While mostly not
516 needed, they can be helpful in some proofs about partial functions.
518 \item @{text tailrec} generates the unconstrained recursive
519 equations even without a termination proof, provided that the
520 function is tail-recursive. This currently only works
522 \item @{text "default d"} allows to specify a default value for a
523 (partial) function, which will ensure that @{text "f x = d x"}
524 whenever @{text "x \<notin> f_dom"}.
530 subsection {* Proof methods related to recursive definitions *}
533 \begin{matharray}{rcl}
534 @{method_def (HOL) pat_completeness} & : & @{text method} \\
535 @{method_def (HOL) relation} & : & @{text method} \\
536 @{method_def (HOL) lexicographic_order} & : & @{text method} \\
537 @{method_def (HOL) size_change} & : & @{text method} \\
543 'lexicographic_order' ( clasimpmod * )
545 'size_change' ( orders ( clasimpmod * ) )
547 orders: ( 'max' | 'min' | 'ms' ) *
552 \item @{method (HOL) pat_completeness} is a specialized method to
553 solve goals regarding the completeness of pattern matching, as
554 required by the @{command (HOL) "function"} package (cf.\
555 \cite{isabelle-function}).
557 \item @{method (HOL) relation}~@{text R} introduces a termination
558 proof using the relation @{text R}. The resulting proof state will
559 contain goals expressing that @{text R} is wellfounded, and that the
560 arguments of recursive calls decrease with respect to @{text R}.
561 Usually, this method is used as the initial proof step of manual
564 \item @{method (HOL) "lexicographic_order"} attempts a fully
565 automated termination proof by searching for a lexicographic
566 combination of size measures on the arguments of the function. The
567 method accepts the same arguments as the @{method auto} method,
568 which it uses internally to prove local descents. The same context
569 modifiers as for @{method auto} are accepted, see
570 \secref{sec:clasimp}.
572 In case of failure, extensive information is printed, which can help
573 to analyse the situation (cf.\ \cite{isabelle-function}).
575 \item @{method (HOL) "size_change"} also works on termination goals,
576 using a variation of the size-change principle, together with a
577 graph decomposition technique (see \cite{krauss_phd} for details).
578 Three kinds of orders are used internally: @{text max}, @{text min},
579 and @{text ms} (multiset), which is only available when the theory
580 @{text Multiset} is loaded. When no order kinds are given, they are
581 tried in order. The search for a termination proof uses SAT solving
584 For local descent proofs, the same context modifiers as for @{method
585 auto} are accepted, see \secref{sec:clasimp}.
590 subsection {* Functions with explicit partiality *}
593 \begin{matharray}{rcl}
594 @{command_def (HOL) "partial_function"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
595 @{attribute_def (HOL) "partial_function_mono"} & : & @{text attribute} \\
599 'partial_function' target? '(' mode ')' fixes \\ 'where' thmdecl? prop
604 \item @{command (HOL) "partial_function"} defines recursive
605 functions based on fixpoints in complete partial orders. No
606 termination proof is required from the user or constructed
607 internally. Instead, the possibility of non-termination is modelled
608 explicitly in the result type, which contains an explicit bottom
611 Pattern matching and mutual recursion are currently not supported.
612 Thus, the specification consists of a single function described by a
613 single recursive equation.
615 There are no fixed syntactic restrictions on the body of the
616 function, but the induced functional must be provably monotonic
617 wrt.\ the underlying order. The monotonicitity proof is performed
618 internally, and the definition is rejected when it fails. The proof
619 can be influenced by declaring hints using the
620 @{attribute (HOL) partial_function_mono} attribute.
622 The mandatory @{text mode} argument specifies the mode of operation
623 of the command, which directly corresponds to a complete partial
624 order on the result type. By default, the following modes are
628 \item @{text option} defines functions that map into the @{type
629 option} type. Here, the value @{term None} is used to model a
630 non-terminating computation. Monotonicity requires that if @{term
631 None} is returned by a recursive call, then the overall result
632 must also be @{term None}. This is best achieved through the use of
633 the monadic operator @{const "Option.bind"}.
635 \item @{text tailrec} defines functions with an arbitrary result
636 type and uses the slightly degenerated partial order where @{term
637 "undefined"} is the bottom element. Now, monotonicity requires that
638 if @{term undefined} is returned by a recursive call, then the
639 overall result must also be @{term undefined}. In practice, this is
640 only satisfied when each recursive call is a tail call, whose result
641 is directly returned. Thus, this mode of operation allows the
642 definition of arbitrary tail-recursive functions.
645 Experienced users may define new modes by instantiating the locale
646 @{const "partial_function_definitions"} appropriately.
648 \item @{attribute (HOL) partial_function_mono} declares rules for
649 use in the internal monononicity proofs of partial function
656 subsection {* Old-style recursive function definitions (TFL) *}
659 The old TFL commands @{command (HOL) "recdef"} and @{command (HOL)
660 "recdef_tc"} for defining recursive are mostly obsolete; @{command
661 (HOL) "function"} or @{command (HOL) "fun"} should be used instead.
663 \begin{matharray}{rcl}
664 @{command_def (HOL) "recdef"} & : & @{text "theory \<rightarrow> theory)"} \\
665 @{command_def (HOL) "recdef_tc"}@{text "\<^sup>*"} & : & @{text "theory \<rightarrow> proof(prove)"} \\
669 'recdef' ('(' 'permissive' ')')? \\ name term (prop +) hints?
673 hints: '(' 'hints' ( recdefmod * ) ')'
675 recdefmod: (('recdef_simp' | 'recdef_cong' | 'recdef_wf') (() | 'add' | 'del') ':' thmrefs) | clasimpmod
677 tc: nameref ('(' nat ')')?
683 \item @{command (HOL) "recdef"} defines general well-founded
684 recursive functions (using the TFL package), see also
685 \cite{isabelle-HOL}. The ``@{text "(permissive)"}'' option tells
686 TFL to recover from failed proof attempts, returning unfinished
687 results. The @{text recdef_simp}, @{text recdef_cong}, and @{text
688 recdef_wf} hints refer to auxiliary rules to be used in the internal
689 automated proof process of TFL. Additional @{syntax clasimpmod}
690 declarations (cf.\ \secref{sec:clasimp}) may be given to tune the
691 context of the Simplifier (cf.\ \secref{sec:simplifier}) and
692 Classical reasoner (cf.\ \secref{sec:classical}).
694 \item @{command (HOL) "recdef_tc"}~@{text "c (i)"} recommences the
695 proof for leftover termination condition number @{text i} (default
696 1) as generated by a @{command (HOL) "recdef"} definition of
699 Note that in most cases, @{command (HOL) "recdef"} is able to finish
700 its internal proofs without manual intervention.
704 \medskip Hints for @{command (HOL) "recdef"} may be also declared
705 globally, using the following attributes.
707 \begin{matharray}{rcl}
708 @{attribute_def (HOL) recdef_simp} & : & @{text attribute} \\
709 @{attribute_def (HOL) recdef_cong} & : & @{text attribute} \\
710 @{attribute_def (HOL) recdef_wf} & : & @{text attribute} \\
714 ('recdef_simp' | 'recdef_cong' | 'recdef_wf') (() | 'add' | 'del')
720 section {* Inductive and coinductive definitions \label{sec:hol-inductive} *}
723 An \textbf{inductive definition} specifies the least predicate (or
724 set) @{text R} closed under given rules: applying a rule to elements
725 of @{text R} yields a result within @{text R}. For example, a
726 structural operational semantics is an inductive definition of an
729 Dually, a \textbf{coinductive definition} specifies the greatest
730 predicate~/ set @{text R} that is consistent with given rules: every
731 element of @{text R} can be seen as arising by applying a rule to
732 elements of @{text R}. An important example is using bisimulation
733 relations to formalise equivalence of processes and infinite data
736 \medskip The HOL package is related to the ZF one, which is
737 described in a separate paper,\footnote{It appeared in CADE
738 \cite{paulson-CADE}; a longer version is distributed with Isabelle.}
739 which you should refer to in case of difficulties. The package is
740 simpler than that of ZF thanks to implicit type-checking in HOL.
741 The types of the (co)inductive predicates (or sets) determine the
742 domain of the fixedpoint definition, and the package does not have
743 to use inference rules for type-checking.
745 \begin{matharray}{rcl}
746 @{command_def (HOL) "inductive"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
747 @{command_def (HOL) "inductive_set"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
748 @{command_def (HOL) "coinductive"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
749 @{command_def (HOL) "coinductive_set"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
750 @{attribute_def (HOL) mono} & : & @{text attribute} \\
754 ('inductive' | 'inductive_set' | 'coinductive' | 'coinductive_set') target? fixes ('for' fixes)? \\
755 ('where' clauses)? ('monos' thmrefs)?
757 clauses: (thmdecl? prop + '|')
759 'mono' (() | 'add' | 'del')
765 \item @{command (HOL) "inductive"} and @{command (HOL)
766 "coinductive"} define (co)inductive predicates from the
767 introduction rules given in the @{keyword "where"} part. The
768 optional @{keyword "for"} part contains a list of parameters of the
769 (co)inductive predicates that remain fixed throughout the
770 definition. The optional @{keyword "monos"} section contains
771 \emph{monotonicity theorems}, which are required for each operator
772 applied to a recursive set in the introduction rules. There
773 \emph{must} be a theorem of the form @{text "A \<le> B \<Longrightarrow> M A \<le> M B"},
774 for each premise @{text "M R\<^sub>i t"} in an introduction rule!
776 \item @{command (HOL) "inductive_set"} and @{command (HOL)
777 "coinductive_set"} are wrappers for to the previous commands,
778 allowing the definition of (co)inductive sets.
780 \item @{attribute (HOL) mono} declares monotonicity rules. These
781 rule are involved in the automated monotonicity proof of @{command
788 subsection {* Derived rules *}
791 Each (co)inductive definition @{text R} adds definitions to the
792 theory and also proves some theorems:
796 \item @{text R.intros} is the list of introduction rules as proven
797 theorems, for the recursive predicates (or sets). The rules are
798 also available individually, using the names given them in the
801 \item @{text R.cases} is the case analysis (or elimination) rule;
803 \item @{text R.induct} or @{text R.coinduct} is the (co)induction
808 When several predicates @{text "R\<^sub>1, \<dots>, R\<^sub>n"} are
809 defined simultaneously, the list of introduction rules is called
810 @{text "R\<^sub>1_\<dots>_R\<^sub>n.intros"}, the case analysis rules are
811 called @{text "R\<^sub>1.cases, \<dots>, R\<^sub>n.cases"}, and the list
812 of mutual induction rules is called @{text
813 "R\<^sub>1_\<dots>_R\<^sub>n.inducts"}.
817 subsection {* Monotonicity theorems *}
820 Each theory contains a default set of theorems that are used in
821 monotonicity proofs. New rules can be added to this set via the
822 @{attribute (HOL) mono} attribute. The HOL theory @{text Inductive}
823 shows how this is done. In general, the following monotonicity
824 theorems may be added:
828 \item Theorems of the form @{text "A \<le> B \<Longrightarrow> M A \<le> M B"}, for proving
829 monotonicity of inductive definitions whose introduction rules have
830 premises involving terms such as @{text "M R\<^sub>i t"}.
832 \item Monotonicity theorems for logical operators, which are of the
833 general form @{text "(\<dots> \<longrightarrow> \<dots>) \<Longrightarrow> \<dots> (\<dots> \<longrightarrow> \<dots>) \<Longrightarrow> \<dots> \<longrightarrow> \<dots>"}. For example, in
834 the case of the operator @{text "\<or>"}, the corresponding theorem is
836 \infer{@{text "P\<^sub>1 \<or> P\<^sub>2 \<longrightarrow> Q\<^sub>1 \<or> Q\<^sub>2"}}{@{text "P\<^sub>1 \<longrightarrow> Q\<^sub>1"} & @{text "P\<^sub>2 \<longrightarrow> Q\<^sub>2"}}
839 \item De Morgan style equations for reasoning about the ``polarity''
842 @{prop "\<not> \<not> P \<longleftrightarrow> P"} \qquad\qquad
843 @{prop "\<not> (P \<and> Q) \<longleftrightarrow> \<not> P \<or> \<not> Q"}
846 \item Equations for reducing complex operators to more primitive
847 ones whose monotonicity can easily be proved, e.g.
849 @{prop "(P \<longrightarrow> Q) \<longleftrightarrow> \<not> P \<or> Q"} \qquad\qquad
850 @{prop "Ball A P \<equiv> \<forall>x. x \<in> A \<longrightarrow> P x"}
855 %FIXME: Example of an inductive definition
859 section {* Arithmetic proof support *}
862 \begin{matharray}{rcl}
863 @{method_def (HOL) arith} & : & @{text method} \\
864 @{attribute_def (HOL) arith} & : & @{text attribute} \\
865 @{attribute_def (HOL) arith_split} & : & @{text attribute} \\
868 The @{method (HOL) arith} method decides linear arithmetic problems
869 (on types @{text nat}, @{text int}, @{text real}). Any current
870 facts are inserted into the goal before running the procedure.
872 The @{attribute (HOL) arith} attribute declares facts that are
873 always supplied to the arithmetic provers implicitly.
875 The @{attribute (HOL) arith_split} attribute declares case split
876 rules to be expanded before @{method (HOL) arith} is invoked.
878 Note that a simpler (but faster) arithmetic prover is
879 already invoked by the Simplifier.
883 section {* Intuitionistic proof search *}
886 \begin{matharray}{rcl}
887 @{method_def (HOL) iprover} & : & @{text method} \\
891 'iprover' ( rulemod * )
895 The @{method (HOL) iprover} method performs intuitionistic proof
896 search, depending on specifically declared rules from the context,
897 or given as explicit arguments. Chained facts are inserted into the
898 goal before commencing proof search.
900 Rules need to be classified as @{attribute (Pure) intro},
901 @{attribute (Pure) elim}, or @{attribute (Pure) dest}; here the
902 ``@{text "!"}'' indicator refers to ``safe'' rules, which may be
903 applied aggressively (without considering back-tracking later).
904 Rules declared with ``@{text "?"}'' are ignored in proof search (the
905 single-step @{method rule} method still observes these). An
906 explicit weight annotation may be given as well; otherwise the
907 number of rule premises will be taken into account here.
911 section {* Coherent Logic *}
914 \begin{matharray}{rcl}
915 @{method_def (HOL) "coherent"} & : & @{text method} \\
923 The @{method (HOL) coherent} method solves problems of
924 \emph{Coherent Logic} \cite{Bezem-Coquand:2005}, which covers
925 applications in confluence theory, lattice theory and projective
926 geometry. See @{file "~~/src/HOL/ex/Coherent.thy"} for some
931 section {* Checking and refuting propositions *}
934 Identifying incorrect propositions usually involves evaluation of
935 particular assignments and systematic counter example search. This
936 is supported by the following commands.
938 \begin{matharray}{rcl}
939 @{command_def (HOL) "value"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
940 @{command_def (HOL) "quickcheck"}@{text "\<^sup>*"} & : & @{text "proof \<rightarrow>"} \\
941 @{command_def (HOL) "quickcheck_params"} & : & @{text "theory \<rightarrow> theory"}
945 'value' ( ( '[' name ']' ) ? ) modes? term
948 'quickcheck' ( ( '[' args ']' ) ? ) nat?
951 'quickcheck_params' ( ( '[' args ']' ) ? )
954 modes: '(' (name + ) ')'
957 args: ( name '=' value + ',' )
963 \item @{command (HOL) "value"}~@{text t} evaluates and prints a
964 term; optionally @{text modes} can be specified, which are
965 appended to the current print mode (see also \cite{isabelle-ref}).
966 Internally, the evaluation is performed by registered evaluators,
967 which are invoked sequentially until a result is returned.
968 Alternatively a specific evaluator can be selected using square
969 brackets; typical evaluators use the current set of code equations
970 to normalize and include @{text simp} for fully symbolic evaluation
971 using the simplifier, @{text nbe} for \emph{normalization by evaluation}
972 and \emph{code} for code generation in SML.
974 \item @{command (HOL) "quickcheck"} tests the current goal for
975 counter examples using a series of assignments for its
976 free variables; by default the first subgoal is tested, an other
977 can be selected explicitly using an optional goal index.
978 Assignments can be chosen exhausting the search space upto a given
979 size or using a fixed number of random assignments in the search space.
980 By default, quickcheck uses exhaustive testing.
981 A number of configuration options are supported for
982 @{command (HOL) "quickcheck"}, notably:
986 \item[@{text tester}] specifies how to explore the search space
987 (e.g. exhaustive or random).
988 An unknown configuration option is treated as an argument to tester,
989 making @{text "tester ="} optional.
990 \item[@{text size}] specifies the maximum size of the search space
991 for assignment values.
993 \item[@{text iterations}] sets how many sets of assignments are
994 generated for each particular size.
996 \item[@{text no_assms}] specifies whether assumptions in
997 structured proofs should be ignored.
999 \item[@{text timeout}] sets the time limit in seconds.
1001 \item[@{text default_type}] sets the type(s) generally used to
1002 instantiate type variables.
1004 \item[@{text report}] if set quickcheck reports how many tests
1005 fulfilled the preconditions.
1007 \item[@{text quiet}] if not set quickcheck informs about the
1008 current size for assignment values.
1010 \item[@{text expect}] can be used to check if the user's
1011 expectation was met (@{text no_expectation}, @{text
1012 no_counterexample}, or @{text counterexample}).
1016 These option can be given within square brackets.
1018 \item @{command (HOL) "quickcheck_params"} changes quickcheck
1019 configuration options persitently.
1025 section {* Unstructured case analysis and induction \label{sec:hol-induct-tac} *}
1028 The following tools of Isabelle/HOL support cases analysis and
1029 induction in unstructured tactic scripts; see also
1030 \secref{sec:cases-induct} for proper Isar versions of similar ideas.
1032 \begin{matharray}{rcl}
1033 @{method_def (HOL) case_tac}@{text "\<^sup>*"} & : & @{text method} \\
1034 @{method_def (HOL) induct_tac}@{text "\<^sup>*"} & : & @{text method} \\
1035 @{method_def (HOL) ind_cases}@{text "\<^sup>*"} & : & @{text method} \\
1036 @{command_def (HOL) "inductive_cases"}@{text "\<^sup>*"} & : & @{text "local_theory \<rightarrow> local_theory"} \\
1040 'case_tac' goalspec? term rule?
1042 'induct_tac' goalspec? (insts * 'and') rule?
1044 'ind_cases' (prop +) ('for' (name +)) ?
1046 'inductive_cases' (thmdecl? (prop +) + 'and')
1049 rule: ('rule' ':' thmref)
1055 \item @{method (HOL) case_tac} and @{method (HOL) induct_tac} admit
1056 to reason about inductive types. Rules are selected according to
1057 the declarations by the @{attribute cases} and @{attribute induct}
1058 attributes, cf.\ \secref{sec:cases-induct}. The @{command (HOL)
1059 datatype} package already takes care of this.
1061 These unstructured tactics feature both goal addressing and dynamic
1062 instantiation. Note that named rule cases are \emph{not} provided
1063 as would be by the proper @{method cases} and @{method induct} proof
1064 methods (see \secref{sec:cases-induct}). Unlike the @{method
1065 induct} method, @{method induct_tac} does not handle structured rule
1066 statements, only the compact object-logic conclusion of the subgoal
1069 \item @{method (HOL) ind_cases} and @{command (HOL)
1070 "inductive_cases"} provide an interface to the internal @{ML_text
1071 mk_cases} operation. Rules are simplified in an unrestricted
1074 While @{method (HOL) ind_cases} is a proof method to apply the
1075 result immediately as elimination rules, @{command (HOL)
1076 "inductive_cases"} provides case split theorems at the theory level
1077 for later use. The @{keyword "for"} argument of the @{method (HOL)
1078 ind_cases} method allows to specify a list of variables that should
1079 be generalized before applying the resulting rule.
1085 section {* Executable code *}
1088 Isabelle/Pure provides two generic frameworks to support code
1089 generation from executable specifications. Isabelle/HOL
1090 instantiates these mechanisms in a way that is amenable to end-user
1093 \medskip One framework generates code from functional programs
1094 (including overloading using type classes) to SML \cite{SML}, OCaml
1095 \cite{OCaml}, Haskell \cite{haskell-revised-report} and Scala
1096 \cite{scala-overview-tech-report}.
1097 Conceptually, code generation is split up in three steps:
1098 \emph{selection} of code theorems, \emph{translation} into an
1099 abstract executable view and \emph{serialization} to a specific
1100 \emph{target language}. Inductive specifications can be executed
1101 using the predicate compiler which operates within HOL.
1102 See \cite{isabelle-codegen} for an introduction.
1104 \begin{matharray}{rcl}
1105 @{command_def (HOL) "export_code"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
1106 @{attribute_def (HOL) code} & : & @{text attribute} \\
1107 @{command_def (HOL) "code_abort"} & : & @{text "theory \<rightarrow> theory"} \\
1108 @{command_def (HOL) "code_datatype"} & : & @{text "theory \<rightarrow> theory"} \\
1109 @{command_def (HOL) "print_codesetup"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
1110 @{attribute_def (HOL) code_inline} & : & @{text attribute} \\
1111 @{attribute_def (HOL) code_post} & : & @{text attribute} \\
1112 @{command_def (HOL) "print_codeproc"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
1113 @{command_def (HOL) "code_thms"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
1114 @{command_def (HOL) "code_deps"}@{text "\<^sup>*"} & : & @{text "context \<rightarrow>"} \\
1115 @{command_def (HOL) "code_const"} & : & @{text "theory \<rightarrow> theory"} \\
1116 @{command_def (HOL) "code_type"} & : & @{text "theory \<rightarrow> theory"} \\
1117 @{command_def (HOL) "code_class"} & : & @{text "theory \<rightarrow> theory"} \\
1118 @{command_def (HOL) "code_instance"} & : & @{text "theory \<rightarrow> theory"} \\
1119 @{command_def (HOL) "code_reserved"} & : & @{text "theory \<rightarrow> theory"} \\
1120 @{command_def (HOL) "code_monad"} & : & @{text "theory \<rightarrow> theory"} \\
1121 @{command_def (HOL) "code_include"} & : & @{text "theory \<rightarrow> theory"} \\
1122 @{command_def (HOL) "code_modulename"} & : & @{text "theory \<rightarrow> theory"} \\
1123 @{command_def (HOL) "code_reflect"} & : & @{text "theory \<rightarrow> theory"}
1127 'export_code' ( constexpr + ) \\
1128 ( ( 'in' target ( 'module_name' string ) ? \\
1129 ( 'file' ( string | '-' ) ) ? ( '(' args ')' ) ?) + ) ?
1135 constexpr: ( const | 'name._' | '_' )
1138 typeconstructor: nameref
1144 target: 'SML' | 'OCaml' | 'Haskell' | 'Scala'
1147 'code' ( 'del' | 'abstype' | 'abstract' ) ?
1150 'code_abort' ( const + )
1153 'code_datatype' ( const + )
1156 'code_inline' ( 'del' ) ?
1159 'code_post' ( 'del' ) ?
1162 'code_thms' ( constexpr + ) ?
1165 'code_deps' ( constexpr + ) ?
1168 'code_const' (const + 'and') \\
1169 ( ( '(' target ( syntax ? + 'and' ) ')' ) + )
1172 'code_type' (typeconstructor + 'and') \\
1173 ( ( '(' target ( syntax ? + 'and' ) ')' ) + )
1176 'code_class' (class + 'and') \\
1177 ( ( '(' target \\ ( string ? + 'and' ) ')' ) + )
1180 'code_instance' (( typeconstructor '::' class ) + 'and') \\
1181 ( ( '(' target ( '-' ? + 'and' ) ')' ) + )
1184 'code_reserved' target ( string + )
1187 'code_monad' const const target
1190 'code_include' target ( string ( string | '-') )
1193 'code_modulename' target ( ( string string ) + )
1196 'code_reflect' string \\
1197 ( 'datatypes' ( string '=' ( '_' | ( string + '|' ) + 'and' ) ) ) ? \\
1198 ( 'functions' ( string + ) ) ? ( 'file' string ) ?
1201 syntax: string | ( 'infix' | 'infixl' | 'infixr' ) nat string
1208 \item @{command (HOL) "export_code"} generates code for a given list
1209 of constants in the specified target language(s). If no
1210 serialization instruction is given, only abstract code is generated
1213 Constants may be specified by giving them literally, referring to
1214 all executable contants within a certain theory by giving @{text
1215 "name.*"}, or referring to \emph{all} executable constants currently
1216 available by giving @{text "*"}.
1218 By default, for each involved theory one corresponding name space
1219 module is generated. Alternativly, a module name may be specified
1220 after the @{keyword "module_name"} keyword; then \emph{all} code is
1221 placed in this module.
1223 For \emph{SML}, \emph{OCaml} and \emph{Scala} the file specification
1224 refers to a single file; for \emph{Haskell}, it refers to a whole
1225 directory, where code is generated in multiple files reflecting the
1226 module hierarchy. Omitting the file specification denotes standard
1229 Serializers take an optional list of arguments in parentheses. For
1230 \emph{SML} and \emph{OCaml}, ``@{text no_signatures}`` omits
1231 explicit module signatures.
1233 For \emph{Haskell} a module name prefix may be given using the
1234 ``@{text "root:"}'' argument; ``@{text string_classes}'' adds a
1235 ``@{verbatim "deriving (Read, Show)"}'' clause to each appropriate
1236 datatype declaration.
1238 \item @{attribute (HOL) code} explicitly selects (or with option
1239 ``@{text "del"}'' deselects) a code equation for code generation.
1240 Usually packages introducing code equations provide a reasonable
1241 default setup for selection. Variants @{text "code abstype"} and
1242 @{text "code abstract"} declare abstract datatype certificates or
1243 code equations on abstract datatype representations respectively.
1245 \item @{command (HOL) "code_abort"} declares constants which are not
1246 required to have a definition by means of code equations; if needed
1247 these are implemented by program abort instead.
1249 \item @{command (HOL) "code_datatype"} specifies a constructor set
1252 \item @{command (HOL) "print_codesetup"} gives an overview on
1253 selected code equations and code generator datatypes.
1255 \item @{attribute (HOL) code_inline} declares (or with option
1256 ``@{text "del"}'' removes) inlining theorems which are applied as
1257 rewrite rules to any code equation during preprocessing.
1259 \item @{attribute (HOL) code_post} declares (or with option ``@{text
1260 "del"}'' removes) theorems which are applied as rewrite rules to any
1261 result of an evaluation.
1263 \item @{command (HOL) "print_codeproc"} prints the setup of the code
1264 generator preprocessor.
1266 \item @{command (HOL) "code_thms"} prints a list of theorems
1267 representing the corresponding program containing all given
1268 constants after preprocessing.
1270 \item @{command (HOL) "code_deps"} visualizes dependencies of
1271 theorems representing the corresponding program containing all given
1272 constants after preprocessing.
1274 \item @{command (HOL) "code_const"} associates a list of constants
1275 with target-specific serializations; omitting a serialization
1276 deletes an existing serialization.
1278 \item @{command (HOL) "code_type"} associates a list of type
1279 constructors with target-specific serializations; omitting a
1280 serialization deletes an existing serialization.
1282 \item @{command (HOL) "code_class"} associates a list of classes
1283 with target-specific class names; omitting a serialization deletes
1284 an existing serialization. This applies only to \emph{Haskell}.
1286 \item @{command (HOL) "code_instance"} declares a list of type
1287 constructor / class instance relations as ``already present'' for a
1288 given target. Omitting a ``@{text "-"}'' deletes an existing
1289 ``already present'' declaration. This applies only to
1292 \item @{command (HOL) "code_reserved"} declares a list of names as
1293 reserved for a given target, preventing it to be shadowed by any
1296 \item @{command (HOL) "code_monad"} provides an auxiliary mechanism
1297 to generate monadic code for Haskell.
1299 \item @{command (HOL) "code_include"} adds arbitrary named content
1300 (``include'') to generated code. A ``@{text "-"}'' as last argument
1301 will remove an already added ``include''.
1303 \item @{command (HOL) "code_modulename"} declares aliasings from one
1304 module name onto another.
1306 \item @{command (HOL) "code_reflect"} without a ``@{text "file"}''
1307 argument compiles code into the system runtime environment and
1308 modifies the code generator setup that future invocations of system
1309 runtime code generation referring to one of the ``@{text
1310 "datatypes"}'' or ``@{text "functions"}'' entities use these precompiled
1311 entities. With a ``@{text "file"}'' argument, the corresponding code
1312 is generated into that specified file without modifying the code
1317 The other framework generates code from both functional and
1318 relational programs to SML. See \cite{isabelle-HOL} for further
1319 information (this actually covers the new-style theory format as
1322 \begin{matharray}{rcl}
1323 @{command_def (HOL) "code_module"} & : & @{text "theory \<rightarrow> theory"} \\
1324 @{command_def (HOL) "code_library"} & : & @{text "theory \<rightarrow> theory"} \\
1325 @{command_def (HOL) "consts_code"} & : & @{text "theory \<rightarrow> theory"} \\
1326 @{command_def (HOL) "types_code"} & : & @{text "theory \<rightarrow> theory"} \\
1327 @{attribute_def (HOL) code} & : & @{text attribute} \\
1331 ( 'code_module' | 'code_library' ) modespec ? name ? \\
1332 ( 'file' name ) ? ( 'imports' ( name + ) ) ? \\
1333 'contains' ( ( name '=' term ) + | term + )
1336 modespec: '(' ( name * ) ')'
1339 'consts_code' (codespec +)
1342 codespec: const template attachment ?
1345 'types_code' (tycodespec +)
1348 tycodespec: name template attachment ?
1354 template: '(' string ')'
1357 attachment: 'attach' modespec ? verblbrace text verbrbrace
1367 section {* Definition by specification \label{sec:hol-specification} *}
1370 \begin{matharray}{rcl}
1371 @{command_def (HOL) "specification"} & : & @{text "theory \<rightarrow> proof(prove)"} \\
1372 @{command_def (HOL) "ax_specification"} & : & @{text "theory \<rightarrow> proof(prove)"} \\
1376 ('specification' | 'ax_specification') '(' (decl +) ')' \\ (thmdecl? prop +)
1378 decl: ((name ':')? term '(' 'overloaded' ')'?)
1383 \item @{command (HOL) "specification"}~@{text "decls \<phi>"} sets up a
1384 goal stating the existence of terms with the properties specified to
1385 hold for the constants given in @{text decls}. After finishing the
1386 proof, the theory will be augmented with definitions for the given
1387 constants, as well as with theorems stating the properties for these
1390 \item @{command (HOL) "ax_specification"}~@{text "decls \<phi>"} sets up
1391 a goal stating the existence of terms with the properties specified
1392 to hold for the constants given in @{text decls}. After finishing
1393 the proof, the theory will be augmented with axioms expressing the
1394 properties given in the first place.
1396 \item @{text decl} declares a constant to be defined by the
1397 specification given. The definition for the constant @{text c} is
1398 bound to the name @{text c_def} unless a theorem name is given in
1399 the declaration. Overloaded constants should be declared as such.
1403 Whether to use @{command (HOL) "specification"} or @{command (HOL)
1404 "ax_specification"} is to some extent a matter of style. @{command
1405 (HOL) "specification"} introduces no new axioms, and so by
1406 construction cannot introduce inconsistencies, whereas @{command
1407 (HOL) "ax_specification"} does introduce axioms, but only after the
1408 user has explicitly proven it to be safe. A practical issue must be
1409 considered, though: After introducing two constants with the same
1410 properties using @{command (HOL) "specification"}, one can prove
1411 that the two constants are, in fact, equal. If this might be a
1412 problem, one should use @{command (HOL) "ax_specification"}.