NEWS
author wenzelm
Wed Aug 01 16:55:37 2007 +0200 (2007-08-01)
changeset 24110 4ab3084e311c
parent 24086 21900a460ba4
child 24172 06e42cf7df4e
permissions -rw-r--r--
tuned config options: eliminated separate attribute "option";
     1 Isabelle NEWS -- history user-relevant changes
     2 ==============================================
     3 
     4 New in this Isabelle version
     5 ----------------------------
     6 
     7 *** General ***
     8 
     9 * More uniform information about legacy features, notably a
    10 warning/error of "Legacy feature: ...", depending on the state of the
    11 tolerate_legacy_features flag (default true). FUTURE INCOMPATIBILITY:
    12 legacy features will disappear eventually.
    13 
    14 * Theory syntax: the header format ``theory A = B + C:'' has been
    15 discontinued in favour of ``theory A imports B C begin''.  Use isatool
    16 fixheaders to convert existing theory files.  INCOMPATIBILITY.
    17 
    18 * Theory syntax: the old non-Isar theory file format has been
    19 discontinued altogether.  Note that ML proof scripts may still be used
    20 with Isar theories; migration is usually quite simple with the ML
    21 function use_legacy_bindings.  INCOMPATIBILITY.
    22 
    23 * Theory syntax: some popular names (e.g. 'class', 'declaration',
    24 'fun', 'help', 'if') are now keywords.  INCOMPATIBILITY, use double
    25 quotes.
    26 
    27 * Theory loader: be more serious about observing the static theory
    28 header specifications (including optional directories), but not the
    29 accidental file locations of previously successful loads.  Potential
    30 INCOMPATIBILITY, may need to refine theory headers.
    31 
    32 * Theory loader: optional support for content-based file
    33 identification, instead of the traditional scheme of full physical
    34 path plus date stamp; configured by the ISABELLE_FILE_IDENT setting
    35 (cf. the system manual).  The new scheme allows to work with
    36 non-finished theories in persistent session images, such that source
    37 files may be moved later on without requiring reloads.
    38 
    39 * Legacy goal package: reduced interface to the bare minimum required
    40 to keep existing proof scripts running.  Most other user-level
    41 functions are now part of the OldGoals structure, which is *not* open
    42 by default (consider isatool expandshort before open OldGoals).
    43 Removed top_sg, prin, printyp, pprint_term/typ altogether, because
    44 these tend to cause confusion about the actual goal (!) context being
    45 used here, which is not necessarily the same as the_context().
    46 
    47 * Command 'find_theorems': supports "*" wild-card in "name:"
    48 criterion; "with_dups" option.  Certain ProofGeneral versions might
    49 support a specific search form (see ProofGeneral/CHANGES).
    50 
    51 * The ``prems limit'' option (cf. ProofContext.prems_limit) is now -1
    52 by default, which means that "prems" (and also "fixed variables") are
    53 suppressed from proof state output.  Note that the ProofGeneral
    54 settings mechanism allows to change and save options persistently, but
    55 older versions of Isabelle will fail to start up if a negative prems
    56 limit is imposed.
    57 
    58 * Local theory targets may be specified by non-nested blocks of
    59 ``context/locale/class ... begin'' followed by ``end''.  The body may
    60 contain definitions, theorems etc., including any derived mechanism
    61 that has been implemented on top of these primitives.  This concept
    62 generalizes the existing ``theorem (in ...)'' towards more versatility
    63 and scalability.
    64 
    65 * Proof General interface: proper undo of final 'end' command;
    66 discontinued Isabelle/classic mode (ML proof scripts).
    67 
    68 
    69 *** Document preparation ***
    70 
    71 * Added antiquotation @{theory name} which prints the given name,
    72 after checking that it refers to a valid ancestor theory in the
    73 current context.
    74 
    75 * Added antiquotations @{ML_type text} and @{ML_struct text} which
    76 check the given source text as ML type/structure, printing verbatim.
    77 
    78 * Added antiquotation @{abbrev "c args"} which prints the abbreviation
    79 "c args == rhs" given in the current context.  (Any number of
    80 arguments may be given on the LHS.)
    81 
    82 
    83 
    84 *** Pure ***
    85 
    86 * code generator: consts in 'consts_code' Isar commands are now referred
    87   to by usual term syntax (including optional type annotations).
    88 
    89 * code generator: 
    90   - Isar 'definition's, 'constdef's and primitive instance definitions are added
    91     explicitly to the table of defining equations
    92   - primitive definitions are not used as defining equations by default any longer
    93   - defining equations are now definitly restricted to meta "==" and object
    94         equality "="
    95   - HOL theories have been adopted accordingly
    96 
    97 * class_package.ML offers a combination of axclasses and locales to
    98 achieve Haskell-like type classes in Isabelle.  See
    99 HOL/ex/Classpackage.thy for examples.
   100 
   101 * Yet another code generator framework allows to generate executable
   102 code for ML and Haskell (including "class"es).  A short usage sketch:
   103 
   104     internal compilation:
   105         code_gen <list of constants (term syntax)> in SML
   106     writing SML code to a file:
   107         code_gen <list of constants (term syntax)> in SML <filename>
   108     writing OCaml code to a file:
   109         code_gen <list of constants (term syntax)> in OCaml <filename>
   110     writing Haskell code to a bunch of files:
   111         code_gen <list of constants (term syntax)> in Haskell <filename>
   112 
   113 Reasonable default setup of framework in HOL/Main.
   114 
   115 Theorem attributs for selecting and transforming function equations theorems:
   116 
   117     [code fun]:        select a theorem as function equation for a specific constant
   118     [code fun del]:    deselect a theorem as function equation for a specific constant
   119     [code inline]:     select an equation theorem for unfolding (inlining) in place
   120     [code inline del]: deselect an equation theorem for unfolding (inlining) in place
   121 
   122 User-defined serializations (target in {SML, OCaml, Haskell}):
   123 
   124     code_const <and-list of constants (term syntax)>
   125       {(target) <and-list of const target syntax>}+
   126 
   127     code_type <and-list of type constructors>
   128       {(target) <and-list of type target syntax>}+
   129 
   130     code_instance <and-list of instances>
   131       {(target)}+
   132         where instance ::= <type constructor> :: <class>
   133 
   134     code_class <and_list of classes>
   135       {(target) <and-list of class target syntax>}+
   136         where class target syntax ::= <class name> {where {<classop> == <target syntax>}+}?
   137 
   138 code_instance and code_class only apply to target Haskell.
   139 
   140 See HOL theories and HOL/ex/Codegenerator*.thy for usage examples.
   141 Doc/Isar/Advanced/Codegen/ provides a tutorial.
   142 
   143 * Command 'no_translations' removes translation rules from theory
   144 syntax.
   145 
   146 * Overloaded definitions are now actually checked for acyclic
   147 dependencies.  The overloading scheme is slightly more general than
   148 that of Haskell98, although Isabelle does not demand an exact
   149 correspondence to type class and instance declarations.
   150 INCOMPATIBILITY, use ``defs (unchecked overloaded)'' to admit more
   151 exotic versions of overloading -- at the discretion of the user!
   152 
   153 Polymorphic constants are represented via type arguments, i.e. the
   154 instantiation that matches an instance against the most general
   155 declaration given in the signature.  For example, with the declaration
   156 c :: 'a => 'a => 'a, an instance c :: nat => nat => nat is represented
   157 as c(nat).  Overloading is essentially simultaneous structural
   158 recursion over such type arguments.  Incomplete specification patterns
   159 impose global constraints on all occurrences, e.g. c('a * 'a) on the
   160 LHS means that more general c('a * 'b) will be disallowed on any RHS.
   161 Command 'print_theory' outputs the normalized system of recursive
   162 equations, see section "definitions".
   163 
   164 * Configuration options are maintained within the theory or proof
   165 context (with name and type bool/int/string), providing a very simple
   166 interface to a poor-man's version of general context data.  Tools may
   167 declare options in ML (e.g. using Attrib.config_int) and then refer to
   168 these values using Config.get etc.  Users may change options via an
   169 associated attribute of the same name.  This form of context
   170 declaration works particularly well with commands 'declare' or
   171 'using', for example ``declare [[foo = 42]]''.  Thus it has become
   172 very easy to avoid global references, which would not observe Isar
   173 toplevel undo/redo and fail to work with multithreading.
   174 
   175 * Named collections of theorems may be easily installed as context
   176 data using the functor NamedThmsFun (see
   177 src/Pure/Tools/named_thms.ML).  The user may add or delete facts via
   178 attributes; there is also a toplevel print command.  This facility is
   179 just a common case of general context data, which is the preferred way
   180 for anything more complex than just a list of facts in canonical
   181 order.
   182 
   183 * Isar: command 'declaration' augments a local theory by generic
   184 declaration functions written in ML.  This enables arbitrary content
   185 being added to the context, depending on a morphism that tells the
   186 difference of the original declaration context wrt. the application
   187 context encountered later on.
   188 
   189 * Isar: proper interfaces for simplification procedures.  Command
   190 'simproc_setup' declares named simprocs (with match patterns, and body
   191 text in ML).  Attribute "simproc" adds/deletes simprocs in the current
   192 context.  ML antiquotation @{simproc name} retrieves named simprocs.
   193 
   194 * Isar: an extra pair of brackets around attribute declarations
   195 abbreviates a theorem reference involving an internal dummy fact,
   196 which will be ignored later --- only the effect of the attribute on
   197 the background context will persist.  This form of in-place
   198 declarations is particularly useful with commands like 'declare' and
   199 'using', for example ``have A using [[simproc a]] by simp''.
   200 
   201 * Isar: method "assumption" (and implicit closing of subproofs) now
   202 takes simple non-atomic goal assumptions into account: after applying
   203 an assumption as a rule the resulting subgoals are solved by atomic
   204 assumption steps.  This is particularly useful to finish 'obtain'
   205 goals, such as "!!x. (!!x. P x ==> thesis) ==> P x ==> thesis",
   206 without referring to the original premise "!!x. P x ==> thesis" in the
   207 Isar proof context.  POTENTIAL INCOMPATIBILITY: method "assumption" is
   208 more permissive.
   209 
   210 * Isar: implicit use of prems from the Isar proof context is
   211 considered a legacy feature.  Common applications like ``have A .''
   212 may be replaced by ``have A by fact'' or ``note `A`''.  In general,
   213 referencing facts explicitly here improves readability and
   214 maintainability of proof texts.
   215 
   216 * Isar: improper proof element 'guess' is like 'obtain', but derives
   217 the obtained context from the course of reasoning!  For example:
   218 
   219   assume "EX x y. A x & B y"   -- "any previous fact"
   220   then guess x and y by clarify
   221 
   222 This technique is potentially adventurous, depending on the facts and
   223 proof tools being involved here.
   224 
   225 * Isar: known facts from the proof context may be specified as literal
   226 propositions, using ASCII back-quote syntax.  This works wherever
   227 named facts used to be allowed so far, in proof commands, proof
   228 methods, attributes etc.  Literal facts are retrieved from the context
   229 according to unification of type and term parameters.  For example,
   230 provided that "A" and "A ==> B" and "!!x. P x ==> Q x" are known
   231 theorems in the current context, then these are valid literal facts:
   232 `A` and `A ==> B` and `!!x. P x ==> Q x" as well as `P a ==> Q a` etc.
   233 
   234 There is also a proof method "fact" which does the same composition
   235 for explicit goal states, e.g. the following proof texts coincide with
   236 certain special cases of literal facts:
   237 
   238   have "A" by fact                 ==  note `A`
   239   have "A ==> B" by fact           ==  note `A ==> B`
   240   have "!!x. P x ==> Q x" by fact  ==  note `!!x. P x ==> Q x`
   241   have "P a ==> Q a" by fact       ==  note `P a ==> Q a`
   242 
   243 * Isar: ":" (colon) is no longer a symbolic identifier character in
   244 outer syntax.  Thus symbolic identifiers may be used without
   245 additional white space in declarations like this: ``assume *: A''.
   246 
   247 * Isar: 'print_facts' prints all local facts of the current context,
   248 both named and unnamed ones.
   249 
   250 * Isar: 'def' now admits simultaneous definitions, e.g.:
   251 
   252   def x == "t" and y == "u"
   253 
   254 * Isar: added command 'unfolding', which is structurally similar to
   255 'using', but affects both the goal state and facts by unfolding given
   256 rewrite rules.  Thus many occurrences of the 'unfold' method or
   257 'unfolded' attribute may be replaced by first-class proof text.
   258 
   259 * Isar: methods 'unfold' / 'fold', attributes 'unfolded' / 'folded',
   260 and command 'unfolding' now all support object-level equalities
   261 (potentially conditional).  The underlying notion of rewrite rule is
   262 analogous to the 'rule_format' attribute, but *not* that of the
   263 Simplifier (which is usually more generous).
   264 
   265 * Isar: the goal restriction operator [N] (default N = 1) evaluates a
   266 method expression within a sandbox consisting of the first N
   267 sub-goals, which need to exist.  For example, ``simp_all [3]''
   268 simplifies the first three sub-goals, while (rule foo, simp_all)[]
   269 simplifies all new goals that emerge from applying rule foo to the
   270 originally first one.
   271 
   272 * Isar: schematic goals are no longer restricted to higher-order
   273 patterns; e.g. ``lemma "?P(?x)" by (rule TrueI)'' now works as
   274 expected.
   275 
   276 * Isar: the conclusion of a long theorem statement is now either
   277 'shows' (a simultaneous conjunction, as before), or 'obtains'
   278 (essentially a disjunction of cases with local parameters and
   279 assumptions).  The latter allows to express general elimination rules
   280 adequately; in this notation common elimination rules look like this:
   281 
   282   lemma exE:    -- "EX x. P x ==> (!!x. P x ==> thesis) ==> thesis"
   283     assumes "EX x. P x"
   284     obtains x where "P x"
   285 
   286   lemma conjE:  -- "A & B ==> (A ==> B ==> thesis) ==> thesis"
   287     assumes "A & B"
   288     obtains A and B
   289 
   290   lemma disjE:  -- "A | B ==> (A ==> thesis) ==> (B ==> thesis) ==> thesis"
   291     assumes "A | B"
   292     obtains
   293       A
   294     | B
   295 
   296 The subsequent classical rules even refer to the formal "thesis"
   297 explicitly:
   298 
   299   lemma classical:     -- "(~ thesis ==> thesis) ==> thesis"
   300     obtains "~ thesis"
   301 
   302   lemma Peirce's_Law:  -- "((thesis ==> something) ==> thesis) ==> thesis"
   303     obtains "thesis ==> something"
   304 
   305 The actual proof of an 'obtains' statement is analogous to that of the
   306 Isar proof element 'obtain', only that there may be several cases.
   307 Optional case names may be specified in parentheses; these will be
   308 available both in the present proof and as annotations in the
   309 resulting rule, for later use with the 'cases' method (cf. attribute
   310 case_names).
   311 
   312 * Isar: the assumptions of a long theorem statement are available as
   313 "assms" fact in the proof context.  This is more appropriate than the
   314 (historical) "prems", which refers to all assumptions of the current
   315 context, including those from the target locale, proof body etc.
   316 
   317 * Isar: 'print_statement' prints theorems from the current theory or
   318 proof context in long statement form, according to the syntax of a
   319 top-level lemma.
   320 
   321 * Isar: 'obtain' takes an optional case name for the local context
   322 introduction rule (default "that").
   323 
   324 * Isar: removed obsolete 'concl is' patterns.  INCOMPATIBILITY, use
   325 explicit (is "_ ==> ?foo") in the rare cases where this still happens
   326 to occur.
   327 
   328 * Pure: syntax "CONST name" produces a fully internalized constant
   329 according to the current context.  This is particularly useful for
   330 syntax translations that should refer to internal constant
   331 representations independently of name spaces.
   332 
   333 * Pure: syntax constant for foo (binder "FOO ") is called "foo_binder"
   334 instead of "FOO ". This allows multiple binder declarations to coexist
   335 in the same context.  INCOMPATIBILITY.
   336 
   337 * Isar/locales: 'notation' provides a robust interface to the 'syntax'
   338 primitive that also works in a locale context (both for constants and
   339 fixed variables).  Type declaration and internal syntactic
   340 representation of given constants retrieved from the context.
   341 
   342 * Isar/locales: new derived specification elements 'axiomatization',
   343 'definition', 'abbreviation', which support type-inference, admit
   344 object-level specifications (equality, equivalence).  See also the
   345 isar-ref manual.  Examples:
   346 
   347   axiomatization
   348     eq  (infix "===" 50) where
   349     eq_refl: "x === x" and eq_subst: "x === y ==> P x ==> P y"
   350 
   351   definition "f x y = x + y + 1"
   352   definition g where "g x = f x x"
   353 
   354   abbreviation
   355     neq  (infix "=!=" 50) where
   356     "x =!= y == ~ (x === y)"
   357 
   358 These specifications may be also used in a locale context.  Then the
   359 constants being introduced depend on certain fixed parameters, and the
   360 constant name is qualified by the locale base name.  An internal
   361 abbreviation takes care for convenient input and output, making the
   362 parameters implicit and using the original short name.  See also
   363 HOL/ex/Abstract_NAT.thy for an example of deriving polymorphic
   364 entities from a monomorphic theory.
   365 
   366 Presently, abbreviations are only available 'in' a target locale, but
   367 not inherited by general import expressions.  Also note that
   368 'abbreviation' may be used as a type-safe replacement for 'syntax' +
   369 'translations' in common applications.
   370 
   371 Concrete syntax is attached to specified constants in internal form,
   372 independently of name spaces.  The parse tree representation is
   373 slightly different -- use 'notation' instead of raw 'syntax', and
   374 'translations' with explicit "CONST" markup to accommodate this.
   375 
   376 * Pure: command 'print_abbrevs' prints all constant abbreviations of
   377 the current context.  Print mode "no_abbrevs" prevents inversion of
   378 abbreviations on output.
   379 
   380 * Isar/locales: improved parameter handling:
   381 - use of locales "var" and "struct" no longer necessary;
   382 - parameter renamings are no longer required to be injective.
   383   This enables, for example, to define a locale for endomorphisms thus:
   384   locale endom = homom mult mult h.
   385 
   386 * Isar/locales: changed the way locales with predicates are defined.
   387 Instead of accumulating the specification, the imported expression is
   388 now an interpretation.  INCOMPATIBILITY: different normal form of
   389 locale expressions.  In particular, in interpretations of locales with
   390 predicates, goals repesenting already interpreted fragments are not
   391 removed automatically.  Use methods `intro_locales' and
   392 `unfold_locales'; see below.
   393 
   394 * Isar/locales: new methods `intro_locales' and `unfold_locales'
   395 provide backward reasoning on locales predicates.  The methods are
   396 aware of interpretations and discharge corresponding goals.
   397 `intro_locales' is less aggressive then `unfold_locales' and does not
   398 unfold predicates to assumptions.
   399 
   400 * Isar/locales: the order in which locale fragments are accumulated
   401 has changed.  This enables to override declarations from fragments due
   402 to interpretations -- for example, unwanted simp rules.
   403 
   404 * Isar/locales: interpretation in theories and proof contexts has been
   405 extended.  One may now specify (and prove) equations, which are
   406 unfolded in interpreted theorems.  This is useful for replacing
   407 defined concepts (constants depending on locale parameters) by
   408 concepts already existing in the target context.  Example:
   409 
   410   interpretation partial_order ["op <= :: [int, int] => bool"]
   411     where "partial_order.less (op <=) (x::int) y = (x < y)"
   412 
   413 Typically, the constant `partial_order.less' is created by a definition
   414 specification element in the context of locale partial_order.
   415 
   416 * Provers/induct: improved internal context management to support
   417 local fixes and defines on-the-fly.  Thus explicit meta-level
   418 connectives !! and ==> are rarely required anymore in inductive goals
   419 (using object-logic connectives for this purpose has been long
   420 obsolete anyway).  The subsequent proof patterns illustrate advanced
   421 techniques of natural induction; general datatypes and inductive sets
   422 work analogously (see also src/HOL/Lambda for realistic examples).
   423 
   424 (1) This is how to ``strengthen'' an inductive goal wrt. certain
   425 parameters:
   426 
   427   lemma
   428     fixes n :: nat and x :: 'a
   429     assumes a: "A n x"
   430     shows "P n x"
   431     using a                     -- {* make induct insert fact a *}
   432   proof (induct n arbitrary: x) -- {* generalize goal to "!!x. A n x ==> P n x" *}
   433     case 0
   434     show ?case sorry
   435   next
   436     case (Suc n)
   437     note `!!x. A n x ==> P n x` -- {* induction hypothesis, according to induction rule *}
   438     note `A (Suc n) x`          -- {* induction premise, stemming from fact a *}
   439     show ?case sorry
   440   qed
   441 
   442 (2) This is how to perform induction over ``expressions of a certain
   443 form'', using a locally defined inductive parameter n == "a x"
   444 together with strengthening (the latter is usually required to get
   445 sufficiently flexible induction hypotheses):
   446 
   447   lemma
   448     fixes a :: "'a => nat"
   449     assumes a: "A (a x)"
   450     shows "P (a x)"
   451     using a
   452   proof (induct n == "a x" arbitrary: x)
   453     ...
   454 
   455 See also HOL/Isar_examples/Puzzle.thy for an application of the this
   456 particular technique.
   457 
   458 (3) This is how to perform existential reasoning ('obtains' or
   459 'obtain') by induction, while avoiding explicit object-logic
   460 encodings:
   461 
   462   lemma
   463     fixes n :: nat
   464     obtains x :: 'a where "P n x" and "Q n x"
   465   proof (induct n arbitrary: thesis)
   466     case 0
   467     obtain x where "P 0 x" and "Q 0 x" sorry
   468     then show thesis by (rule 0)
   469   next
   470     case (Suc n)
   471     obtain x where "P n x" and "Q n x" by (rule Suc.hyps)
   472     obtain x where "P (Suc n) x" and "Q (Suc n) x" sorry
   473     then show thesis by (rule Suc.prems)
   474   qed
   475 
   476 Here the 'arbitrary: thesis' specification essentially modifies the
   477 scope of the formal thesis parameter, in order to the get the whole
   478 existence statement through the induction as expected.
   479 
   480 * Provers/induct: mutual induction rules are now specified as a list
   481 of rule sharing the same induction cases.  HOL packages usually
   482 provide foo_bar.inducts for mutually defined items foo and bar
   483 (e.g. inductive sets or datatypes).  INCOMPATIBILITY, users need to
   484 specify mutual induction rules differently, i.e. like this:
   485 
   486   (induct rule: foo_bar.inducts)
   487   (induct set: foo bar)
   488   (induct type: foo bar)
   489 
   490 The ML function ProjectRule.projections turns old-style rules into the
   491 new format.
   492 
   493 * Provers/induct: improved handling of simultaneous goals.  Instead of
   494 introducing object-level conjunction, the statement is now split into
   495 several conclusions, while the corresponding symbolic cases are
   496 nested accordingly.  INCOMPATIBILITY, proofs need to be structured
   497 explicitly.  For example:
   498 
   499   lemma
   500     fixes n :: nat
   501     shows "P n" and "Q n"
   502   proof (induct n)
   503     case 0 case 1
   504     show "P 0" sorry
   505   next
   506     case 0 case 2
   507     show "Q 0" sorry
   508   next
   509     case (Suc n) case 1
   510     note `P n` and `Q n`
   511     show "P (Suc n)" sorry
   512   next
   513     case (Suc n) case 2
   514     note `P n` and `Q n`
   515     show "Q (Suc n)" sorry
   516   qed
   517 
   518 The split into subcases may be deferred as follows -- this is
   519 particularly relevant for goal statements with local premises.
   520 
   521   lemma
   522     fixes n :: nat
   523     shows "A n ==> P n" and "B n ==> Q n"
   524   proof (induct n)
   525     case 0
   526     {
   527       case 1
   528       note `A 0`
   529       show "P 0" sorry
   530     next
   531       case 2
   532       note `B 0`
   533       show "Q 0" sorry
   534     }
   535   next
   536     case (Suc n)
   537     note `A n ==> P n` and `B n ==> Q n`
   538     {
   539       case 1
   540       note `A (Suc n)`
   541       show "P (Suc n)" sorry
   542     next
   543       case 2
   544       note `B (Suc n)`
   545       show "Q (Suc n)" sorry
   546     }
   547   qed
   548 
   549 If simultaneous goals are to be used with mutual rules, the statement
   550 needs to be structured carefully as a two-level conjunction, using
   551 lists of propositions separated by 'and':
   552 
   553   lemma
   554     shows "a : A ==> P1 a"
   555           "a : A ==> P2 a"
   556       and "b : B ==> Q1 b"
   557           "b : B ==> Q2 b"
   558           "b : B ==> Q3 b"
   559   proof (induct set: A B)
   560 
   561 * Provers/induct: support coinduction as well.  See
   562 src/HOL/Library/Coinductive_List.thy for various examples.
   563 
   564 * Attribute "symmetric" produces result with standardized schematic
   565 variables (index 0).  Potential INCOMPATIBILITY.
   566 
   567 * Simplifier: by default the simplifier trace only shows top level
   568 rewrites now. That is, trace_simp_depth_limit is set to 1 by
   569 default. Thus there is less danger of being flooded by the trace. The
   570 trace indicates where parts have been suppressed.
   571   
   572 * Provers/classical: removed obsolete classical version of elim_format
   573 attribute; classical elim/dest rules are now treated uniformly when
   574 manipulating the claset.
   575 
   576 * Provers/classical: stricter checks to ensure that supplied intro,
   577 dest and elim rules are well-formed; dest and elim rules must have at
   578 least one premise.
   579 
   580 * Provers/classical: attributes dest/elim/intro take an optional
   581 weight argument for the rule (just as the Pure versions).  Weights are
   582 ignored by automated tools, but determine the search order of single
   583 rule steps.
   584 
   585 * Syntax: input syntax now supports dummy variable binding "%_. b",
   586 where the body does not mention the bound variable.  Note that dummy
   587 patterns implicitly depend on their context of bounds, which makes
   588 "{_. _}" match any set comprehension as expected.  Potential
   589 INCOMPATIBILITY -- parse translations need to cope with syntactic
   590 constant "_idtdummy" in the binding position.
   591 
   592 * Syntax: removed obsolete syntactic constant "_K" and its associated
   593 parse translation.  INCOMPATIBILITY -- use dummy abstraction instead,
   594 for example "A -> B" => "Pi A (%_. B)".
   595 
   596 * Pure: 'class_deps' command visualizes the subclass relation, using
   597 the graph browser tool.
   598 
   599 * Pure: 'print_theory' now suppresses entities with internal name
   600 (trailing "_") by default; use '!' option for full details.
   601 
   602 
   603 *** HOL ***
   604 
   605 * Code generator library theories:
   606   * Pretty_Int represents HOL integers by big integer literals in target
   607     languages.
   608   * Pretty_Char represents HOL characters by character literals in target
   609     languages.
   610   * Pretty_Char_chr like Pretty_Char, but also offers treatment of character
   611     codes; includes Pretty_Int.
   612   * Executable_Set allows to generate code for finite sets using lists.
   613   * Executable_Rat implements rational numbers as triples (sign, enumerator,
   614     denominator).
   615   * Executable_Real implements a subset of real numbers, namly those
   616     representable by rational numbers.
   617   * Efficient_Nat implements natural numbers by integers, which in general will
   618     result in higher efficency; pattern matching with 0/Suc is eliminated;
   619     includes Pretty_Int.
   620   * ML_String provides an additional datatype ml_string; in the HOL default
   621     setup, strings in HOL are mapped to lists of HOL characters in SML; values
   622     of type ml_string are mapped to strings in SML.
   623   * ML_Int provides an additional datatype ml_int which is mapped to to SML
   624     built-in integers.
   625 
   626 * New package for inductive predicates
   627 
   628   An n-ary predicate p with m parameters z_1, ..., z_m can now be defined via
   629 
   630     inductive
   631       p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   632       for z_1 :: U_1 and ... and z_n :: U_m
   633     where
   634       rule_1: "... ==> p z_1 ... z_m t_1_1 ... t_1_n"
   635     | ...
   636 
   637   rather than
   638 
   639     consts s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   640 
   641     abbreviation p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   642     where "p z_1 ... z_m x_1 ... x_n == (x_1, ..., x_n) : s z_1 ... z_m"
   643 
   644     inductive "s z_1 ... z_m"
   645     intros
   646       rule_1: "... ==> (t_1_1, ..., t_1_n) : s z_1 ... z_m"
   647       ...
   648 
   649   For backward compatibility, there is a wrapper allowing inductive
   650   sets to be defined with the new package via
   651 
   652     inductive_set
   653       s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   654       for z_1 :: U_1 and ... and z_n :: U_m
   655     where
   656       rule_1: "... ==> (t_1_1, ..., t_1_n) : s z_1 ... z_m"
   657     | ...
   658 
   659   or
   660 
   661     inductive_set
   662       s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   663       and p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   664       for z_1 :: U_1 and ... and z_n :: U_m
   665     where
   666       "p z_1 ... z_m x_1 ... x_n == (x_1, ..., x_n) : s z_1 ... z_m"
   667     | rule_1: "... ==> p z_1 ... z_m t_1_1 ... t_1_n"
   668     | ...
   669 
   670   if the additional syntax "p ..." is required.
   671 
   672   Many examples can be found in the subdirectories Auth, Bali, Induct,
   673   or MicroJava.
   674 
   675   INCOMPATIBILITIES:
   676 
   677   - Since declaration and definition of inductive sets or predicates
   678     is no longer separated, abbreviations involving the newly introduced
   679     sets or predicates must be specified together with the introduction
   680     rules after the "where" keyword (see example above), rather than before
   681     the actual inductive definition.
   682 
   683   - The variables in induction and elimination rules are now quantified
   684     in the order of their occurrence in the introduction rules, rather than
   685     in alphabetical order. Since this may break some proofs, these proofs
   686     either have to be repaired, e.g. by reordering the variables
   687     a_i_1 ... a_i_{k_i} in Isar "case" statements of the form
   688 
   689       case (rule_i a_i_1 ... a_i_{k_i})
   690 
   691     or the old order of quantification has to be restored by explicitly adding
   692     meta-level quantifiers in the introduction rules, i.e.
   693 
   694       | rule_i: "!!a_i_1 ... a_i_{k_i}. ... ==> p z_1 ... z_m t_i_1 ... t_i_n"
   695 
   696   - The format of the elimination rules is now
   697 
   698       p z_1 ... z_m x_1 ... x_n ==>
   699         (!!a_1_1 ... a_1_{k_1}. x_1 = t_1_1 ==> ... ==> x_n = t_1_n ==> ... ==> P)
   700         ==> ... ==> P
   701 
   702     for predicates and
   703 
   704       (x_1, ..., x_n) : s z_1 ... z_m ==>
   705         (!!a_1_1 ... a_1_{k_1}. x_1 = t_1_1 ==> ... ==> x_n = t_1_n ==> ... ==> P)
   706         ==> ... ==> P
   707 
   708     for sets rather than
   709 
   710       x : s z_1 ... z_m ==>
   711         (!!a_1_1 ... a_1_{k_1}. x = (t_1_1, ..., t_1_n) ==> ... ==> P)
   712         ==> ... ==> P
   713 
   714     This may require terms in goals to be expanded to n-tuples (e.g. using case_tac
   715     or simplification with the split_paired_all rule) before the above elimination
   716     rule is applicable.
   717 
   718   - The elimination or case analysis rules for (mutually) inductive sets or
   719     predicates are now called "p_1.cases" ... "p_k.cases". The list of rules
   720     "p_1_..._p_k.elims" is no longer available.
   721 
   722 * Method "metis" proves goals by applying the Metis general-purpose
   723 resolution prover.  Examples are in the directory MetisExamples.  See
   724 also http://gilith.com/software/metis/
   725   
   726 * Command 'sledgehammer' invokes external automatic theorem provers as
   727 background processes.  It generates calls to the "metis" method if
   728 successful. These can be pasted into the proof.  Users do not have to
   729 wait for the automatic provers to return.
   730 
   731 * Case-expressions allow arbitrary constructor-patterns (including "_") and
   732   take their order into account, like in functional programming.
   733   Internally, this is translated into nested case-expressions; missing cases
   734   are added and mapped to the predefined constant "undefined". In complicated
   735   cases printing may no longer show the original input but the internal
   736   form. Lambda-abstractions allow the same form of pattern matching:
   737   "% pat1 => e1 | ..." is an abbreviation for
   738   "%x. case x of pat1 => e1 | ..." where x is a new variable.
   739 
   740 * IntDef: The constant "int :: nat => int" has been removed; now "int"
   741   is an abbreviation for "of_nat :: nat => int". The simplification rules
   742   for "of_nat" have been changed to work like "int" did previously.
   743   (potential INCOMPATIBILITY)
   744   - "of_nat (Suc m)" simplifies to "1 + of_nat m" instead of "of_nat m + 1"
   745   - of_nat_diff and of_nat_mult are no longer default simp rules
   746 
   747 * Method "algebra" solves polynomial equations over (semi)rings using
   748   Groebner bases. The (semi)ring structure is defined by locales and
   749   the tool setup depends on that generic context. Installing the
   750   method for a specific type involves instantiating the locale and
   751   possibly adding declarations for computation on the coefficients.
   752   The method is already instantiated for natural numbers and for the
   753   axiomatic class of idoms with numerals.  See also the paper by
   754   Chaieb and Wenzel at CALCULEMUS 2007 for the general principles
   755   underlying this architecture of context-aware proof-tools.
   756 
   757 * constant "List.op @" now named "List.append".  Use ML antiquotations
   758 @{const_name List.append} or @{term " ... @ ... "} to circumvent
   759 possible incompatibilities when working on ML level.
   760 
   761 * Constant renames due to introduction of canonical name prefixing for
   762   class package:
   763 
   764     HOL.abs ~> HOL.minus_class.abs
   765     HOL.divide ~> HOL.divide_class.divide
   766     Nat.power ~> Nat.power_class.power
   767     Nat.size ~> Nat.size_class.size
   768     Numeral.number_of ~> Numeral.number_class.number_of
   769     FixedPoint.Inf ~> FixedPoint.complete_lattice_class.Inf
   770 
   771 * Rudimentary class target mechanism involves constant renames:
   772 
   773     Orderings.min ~> Orderings.ord_class.min
   774     Orderings.max ~> Orderings.ord_class.max
   775     FixedPoint.Sup ~> FixedPoint.complete_lattice_class.Sup
   776 
   777 * primrec: missing cases mapped to "undefined" instead of "arbitrary"
   778 
   779 * new constant "undefined" with axiom "undefined x = undefined"
   780 
   781 * new class "default" with associated constant "default"
   782 
   783 * new function listsum :: 'a list => 'a for arbitrary monoids.
   784   Special syntax: "SUM x <- xs. f x" (and latex variants)
   785 
   786 * new (input only) syntax for Haskell-like list comprehension, eg
   787   [(x,y). x <- xs, y <- ys, x ~= y]
   788   For details see List.thy.
   789 
   790 * The special syntax for function "filter" has changed from [x : xs. P] to
   791   [x <- xs. P] to avoid an ambiguity caused by list comprehension syntax,
   792   and for uniformity. INCOMPATIBILITY
   793 
   794 * Lemma "set_take_whileD" renamed to "set_takeWhileD"
   795 
   796 * New lemma collection field_simps (an extension of ring_simps)
   797   for manipulating (in)equations involving division. Multiplies
   798   with all denominators that can be proved to be non-zero (in equations)
   799   or positive/negative (in inequations).
   800 
   801 * Lemma collections ring_eq_simps, group_eq_simps and ring_distrib
   802   have been improved and renamed to ring_simps, group_simps and ring_distribs.
   803   Removed lemmas field_xyz in Ring_and_Field
   804   because they were subsumed by lemmas xyz.
   805 INCOMPATIBILITY.
   806 
   807 * Library/Pretty_Int.thy: maps HOL numerals on target language integer literals
   808   when generating code.
   809 
   810 * Library/Pretty_Char.thy: maps HOL characters on target language character literals
   811   when generating code.
   812 
   813 * Library/Commutative_Ring.thy: switched from recdef to function package;
   814   constants add, mul, pow now curried.  Infix syntax for algebraic operations.
   815 
   816 * Some steps towards more uniform lattice theory development in HOL.
   817 
   818     constants "meet" and "join" now named "inf" and "sup"
   819     constant "Meet" now named "Inf"
   820 
   821     classes "meet_semilorder" and "join_semilorder" now named
   822       "lower_semilattice" and "upper_semilattice"
   823     class "lorder" now named "lattice"
   824     class "comp_lat" now named "complete_lattice"
   825 
   826     Instantiation of lattice classes allows explicit definitions
   827     for "inf" and "sup" operations.
   828 
   829   INCOMPATIBILITY.  Theorem renames:
   830 
   831     meet_left_le            ~> inf_le1
   832     meet_right_le           ~> inf_le2
   833     join_left_le            ~> sup_ge1
   834     join_right_le           ~> sup_ge2
   835     meet_join_le            ~> inf_sup_ord
   836     le_meetI                ~> le_infI
   837     join_leI                ~> le_supI
   838     le_meet                 ~> le_inf_iff
   839     le_join                 ~> ge_sup_conv
   840     meet_idempotent         ~> inf_idem
   841     join_idempotent         ~> sup_idem
   842     meet_comm               ~> inf_commute
   843     join_comm               ~> sup_commute
   844     meet_leI1               ~> le_infI1
   845     meet_leI2               ~> le_infI2
   846     le_joinI1               ~> le_supI1
   847     le_joinI2               ~> le_supI2
   848     meet_assoc              ~> inf_assoc
   849     join_assoc              ~> sup_assoc
   850     meet_left_comm          ~> inf_left_commute
   851     meet_left_idempotent    ~> inf_left_idem
   852     join_left_comm          ~> sup_left_commute
   853     join_left_idempotent    ~> sup_left_idem
   854     meet_aci                ~> inf_aci
   855     join_aci                ~> sup_aci
   856     le_def_meet             ~> le_iff_inf
   857     le_def_join             ~> le_iff_sup
   858     join_absorp2            ~> sup_absorb2
   859     join_absorp1            ~> sup_absorb1
   860     meet_absorp1            ~> inf_absorb1
   861     meet_absorp2            ~> inf_absorb2
   862     meet_join_absorp        ~> inf_sup_absorb
   863     join_meet_absorp        ~> sup_inf_absorb
   864     distrib_join_le         ~> distrib_sup_le
   865     distrib_meet_le         ~> distrib_inf_le
   866 
   867     add_meet_distrib_left   ~> add_inf_distrib_left
   868     add_join_distrib_left   ~> add_sup_distrib_left
   869     is_join_neg_meet        ~> is_join_neg_inf
   870     is_meet_neg_join        ~> is_meet_neg_sup
   871     add_meet_distrib_right  ~> add_inf_distrib_right
   872     add_join_distrib_right  ~> add_sup_distrib_right
   873     add_meet_join_distribs  ~> add_sup_inf_distribs
   874     join_eq_neg_meet        ~> sup_eq_neg_inf
   875     meet_eq_neg_join        ~> inf_eq_neg_sup
   876     add_eq_meet_join        ~> add_eq_inf_sup
   877     meet_0_imp_0            ~> inf_0_imp_0
   878     join_0_imp_0            ~> sup_0_imp_0
   879     meet_0_eq_0             ~> inf_0_eq_0
   880     join_0_eq_0             ~> sup_0_eq_0
   881     neg_meet_eq_join        ~> neg_inf_eq_sup
   882     neg_join_eq_meet        ~> neg_sup_eq_inf
   883     join_eq_if              ~> sup_eq_if
   884 
   885     mono_meet               ~> mono_inf
   886     mono_join               ~> mono_sup
   887     meet_bool_eq            ~> inf_bool_eq
   888     join_bool_eq            ~> sup_bool_eq
   889     meet_fun_eq             ~> inf_fun_eq
   890     join_fun_eq             ~> sup_fun_eq
   891     meet_set_eq             ~> inf_set_eq
   892     join_set_eq             ~> sup_set_eq
   893     meet1_iff               ~> inf1_iff
   894     meet2_iff               ~> inf2_iff
   895     meet1I                  ~> inf1I
   896     meet2I                  ~> inf2I
   897     meet1D1                 ~> inf1D1
   898     meet2D1                 ~> inf2D1
   899     meet1D2                 ~> inf1D2
   900     meet2D2                 ~> inf2D2
   901     meet1E                  ~> inf1E
   902     meet2E                  ~> inf2E
   903     join1_iff               ~> sup1_iff
   904     join2_iff               ~> sup2_iff
   905     join1I1                 ~> sup1I1
   906     join2I1                 ~> sup2I1
   907     join1I1                 ~> sup1I1
   908     join2I2                 ~> sup1I2
   909     join1CI                 ~> sup1CI
   910     join2CI                 ~> sup2CI
   911     join1E                  ~> sup1E
   912     join2E                  ~> sup2E
   913 
   914     is_meet_Meet            ~> is_meet_Inf
   915     Meet_bool_def           ~> Inf_bool_def
   916     Meet_fun_def            ~> Inf_fun_def
   917     Meet_greatest           ~> Inf_greatest
   918     Meet_lower              ~> Inf_lower
   919     Meet_set_def            ~> Inf_set_def
   920 
   921     listsp_meetI            ~> listsp_infI
   922     listsp_meet_eq          ~> listsp_inf_eq
   923 
   924     meet_min                ~> inf_min
   925     join_max                ~> sup_max
   926 
   927 * Classes "order" and "linorder": facts "refl", "trans" and
   928 "cases" renamed ro "order_refl", "order_trans" and "linorder_cases", to
   929 avoid clashes with HOL "refl" and "trans". INCOMPATIBILITY.
   930 
   931 * Classes "order" and "linorder": 
   932 potential INCOMPATIBILITY: order of proof goals in order/linorder instance
   933 proofs changed.
   934 
   935 * Dropped lemma duplicate def_imp_eq in favor of meta_eq_to_obj_eq.
   936 INCOMPATIBILITY.
   937 
   938 * Dropped lemma duplicate if_def2 in favor of if_bool_eq_conj.
   939 INCOMPATIBILITY.
   940 
   941 * Added syntactic class "size"; overloaded constant "size" now has
   942 type "'a::size ==> bool"
   943 
   944 * Renamed constants "Divides.op div", "Divides.op mod" and "Divides.op
   945 dvd" to "Divides.div_class.div", "Divides.div_class.mod" and "Divides.dvd". INCOMPATIBILITY.
   946 
   947 * Added method "lexicographic_order" automatically synthesizes
   948 termination relations as lexicographic combinations of size measures
   949 -- 'function' package.
   950 
   951 * HOL/records: generalised field-update to take a function on the
   952 field rather than the new value: r(|A := x|) is translated to A_update
   953 (K x) r The K-combinator that is internally used is called K_record.
   954 INCOMPATIBILITY: Usage of the plain update functions has to be
   955 adapted.
   956  
   957 * axclass "semiring_0" now contains annihilation axioms x * 0 = 0 and
   958 0 * x = 0, which are required for a semiring.  Richer structures do
   959 not inherit from semiring_0 anymore, because this property is a
   960 theorem there, not an axiom.  INCOMPATIBILITY: In instances of
   961 semiring_0, there is more to prove, but this is mostly trivial.
   962 
   963 * axclass "recpower" was generalized to arbitrary monoids, not just
   964 commutative semirings.  INCOMPATIBILITY: If you use recpower and need
   965 commutativity or a semiring property, add the corresponding classes.
   966 
   967 * Unified locale partial_order with class definition (cf. theory
   968 Orderings), added parameter ``less''.  INCOMPATIBILITY.
   969 
   970 * Constant "List.list_all2" in List.thy now uses authentic syntax.
   971 INCOMPATIBILITY: translations containing list_all2 may go wrong.  On
   972 Isar level, use abbreviations instead.
   973 
   974 * Renamed constant "List.op mem" to "List.memberl" INCOMPATIBILITY:
   975 rarely occuring name references (e.g. ``List.op mem.simps'') require
   976 renaming (e.g. ``List.memberl.simps'').
   977 
   978 * Renamed constants "0" to "HOL.zero_class.zero" and "1" to "HOL.one_class.one".
   979 INCOMPATIBILITY.
   980 
   981 * Added class "HOL.eq", allowing for code generation with polymorphic equality.
   982 
   983 * Numeral syntax: type 'bin' which was a mere type copy of 'int' has
   984 been abandoned in favour of plain 'int'. INCOMPATIBILITY --
   985 significant changes for setting up numeral syntax for types:
   986 
   987   - new constants Numeral.pred and Numeral.succ instead
   988       of former Numeral.bin_pred and Numeral.bin_succ.
   989   - Use integer operations instead of bin_add, bin_mult and so on.
   990   - Numeral simplification theorems named Numeral.numeral_simps instead of Bin_simps.
   991   - ML structure Bin_Simprocs now named Int_Numeral_Base_Simprocs.
   992 
   993 See HOL/Integ/IntArith.thy for an example setup.
   994 
   995 * New top level command 'normal_form' computes the normal form of a
   996 term that may contain free variables. For example ``normal_form
   997 "rev[a,b,c]"'' produces ``[b,c,a]'' (without proof).  This command is
   998 suitable for heavy-duty computations because the functions are
   999 compiled to ML first.
  1000 
  1001 * Alternative iff syntax "A <-> B" for equality on bool (with priority
  1002 25 like -->); output depends on the "iff" print_mode, the default is
  1003 "A = B" (with priority 50).
  1004 
  1005 * Renamed constants in HOL.thy and Orderings.thy:
  1006     op +   ~> HOL.plus_class.plus
  1007     op -   ~> HOL.minus_class.minus
  1008     uminus ~> HOL.minus_class.uminus
  1009     abs    ~> HOL.abs_class.abs
  1010     op *   ~> HOL.times_class.times
  1011     op <   ~> HOL.ord_class.less
  1012     op <=  ~> HOL.ord_class.less_eq
  1013 
  1014 Adaptions may be required in the following cases:
  1015 
  1016 a) User-defined constants using any of the names "plus", "minus", "times",
  1017 "less" or "less_eq". The standard syntax translations for "+", "-" and "*"
  1018 may go wrong.
  1019 INCOMPATIBILITY: use more specific names.
  1020 
  1021 b) Variables named "plus", "minus", "times", "less", "less_eq"
  1022 INCOMPATIBILITY: use more specific names.
  1023 
  1024 c) Permutative equations (e.g. "a + b = b + a")
  1025 Since the change of names also changes the order of terms, permutative
  1026 rewrite rules may get applied in a different order. Experience shows that
  1027 this is rarely the case (only two adaptions in the whole Isabelle
  1028 distribution).
  1029 INCOMPATIBILITY: rewrite proofs
  1030 
  1031 d) ML code directly refering to constant names
  1032 This in general only affects hand-written proof tactics, simprocs and so on.
  1033 INCOMPATIBILITY: grep your sourcecode and replace names.  Consider use
  1034 of const_name ML antiquotations.
  1035 
  1036 * Relations less (<) and less_eq (<=) are also available on type bool.
  1037 Modified syntax to disallow nesting without explicit parentheses,
  1038 e.g. "(x < y) < z" or "x < (y < z)", but NOT "x < y < z".
  1039 
  1040 * "LEAST x:A. P" expands to "LEAST x. x:A & P" (input only).
  1041 
  1042 * Relation composition operator "op O" now has precedence 75 and binds
  1043 stronger than union and intersection. INCOMPATIBILITY.
  1044 
  1045 * The old set interval syntax "{m..n(}" (and relatives) has been
  1046 removed.  Use "{m..<n}" (and relatives) instead.
  1047 
  1048 * In the context of the assumption "~(s = t)" the Simplifier rewrites
  1049 "t = s" to False (by simproc "neq_simproc").  For backward
  1050 compatibility this can be disabled by ML "reset use_neq_simproc".
  1051 
  1052 * "m dvd n" where m and n are numbers is evaluated to True/False by
  1053 simp.
  1054 
  1055 * Theorem Cons_eq_map_conv no longer declared as ``simp''.
  1056 
  1057 * Theorem setsum_mult renamed to setsum_right_distrib.
  1058 
  1059 * Prefer ex1I over ex_ex1I in single-step reasoning, e.g. by the
  1060 ``rule'' method.
  1061 
  1062 * Reimplemented methods ``sat'' and ``satx'', with several
  1063 improvements: goals no longer need to be stated as "<prems> ==>
  1064 False", equivalences (i.e. "=" on type bool) are handled, variable
  1065 names of the form "lit_<n>" are no longer reserved, significant
  1066 speedup.
  1067 
  1068 * Methods ``sat'' and ``satx'' can now replay MiniSat proof traces.
  1069 zChaff is still supported as well.
  1070 
  1071 * 'inductive' and 'datatype': provide projections of mutual rules,
  1072 bundled as foo_bar.inducts;
  1073 
  1074 * Library: moved theories Parity, GCD, Binomial, Infinite_Set to
  1075 Library.
  1076 
  1077 * Library: moved theory Accessible_Part to main HOL.
  1078 
  1079 * Library: added theory Coinductive_List of potentially infinite lists
  1080 as greatest fixed-point.
  1081 
  1082 * Library: added theory AssocList which implements (finite) maps as
  1083 association lists.
  1084 
  1085 * Added proof method ``evaluation'' for efficiently solving a goal
  1086 (i.e. a boolean expression) by compiling it to ML. The goal is
  1087 "proved" (via an oracle) if it evaluates to True.
  1088 
  1089 * Linear arithmetic now splits certain operators (e.g. min, max, abs)
  1090 also when invoked by the simplifier.  This results in the simplifier
  1091 being more powerful on arithmetic goals.  INCOMPATIBILITY.  Set
  1092 fast_arith_split_limit to 0 to obtain the old behavior.
  1093 
  1094 * Support for hex (0x20) and binary (0b1001) numerals.
  1095 
  1096 * New method: reify eqs (t), where eqs are equations for an
  1097 interpretation I :: 'a list => 'b => 'c and t::'c is an optional
  1098 parameter, computes a term s::'b and a list xs::'a list and proves the
  1099 theorem I xs s = t. This is also known as reification or quoting. The
  1100 resulting theorem is applied to the subgoal to substitute t with I xs
  1101 s.  If t is omitted, the subgoal itself is reified.
  1102 
  1103 * New method: reflection corr_thm eqs (t). The parameters eqs and (t)
  1104 are as explained above. corr_thm is a theorem for I vs (f t) = I vs t,
  1105 where f is supposed to be a computable function (in the sense of code
  1106 generattion). The method uses reify to compute s and xs as above then
  1107 applies corr_thm and uses normalization by evaluation to "prove" f s =
  1108 r and finally gets the theorem t = r, which is again applied to the
  1109 subgoal. An Example is available in HOL/ex/ReflectionEx.thy.
  1110 
  1111 * Reflection: Automatic reification now handels binding, an example
  1112 is available in HOL/ex/ReflectionEx.thy
  1113 
  1114 
  1115 *** HOL-Algebra ***
  1116 
  1117 * Formalisation of ideals and the quotient construction over rings.
  1118 
  1119 * Order and lattice theory no longer based on records.
  1120 INCOMPATIBILITY.
  1121 
  1122 * Renamed lemmas least_carrier -> least_closed and greatest_carrier ->
  1123 greatest_closed.  INCOMPATIBILITY.
  1124 
  1125 * Method algebra is now set up via an attribute.  For examples see
  1126 Ring.thy.  INCOMPATIBILITY: the method is now weaker on combinations
  1127 of algebraic structures.
  1128 
  1129 * Renamed theory CRing to Ring.
  1130 
  1131 
  1132 *** HOL-Complex ***
  1133 
  1134 * Theory Real: new method ferrack implements quantifier elimination
  1135 for linear arithmetic over the reals. The quantifier elimination
  1136 feature is used only for decision, for compatibility with arith. This
  1137 means a goal is either solved or left unchanged, no simplification.
  1138 
  1139 * Hyperreal: Functions root and sqrt are now defined on negative real
  1140 inputs so that root n (- x) = - root n x and sqrt (- x) = - sqrt x.
  1141 Nonnegativity side conditions have been removed from many lemmas, so
  1142 that more subgoals may now be solved by simplification; potential
  1143 INCOMPATIBILITY.
  1144 
  1145 * Real: New axiomatic classes formalize real normed vector spaces and
  1146 algebras, using new overloaded constants scaleR :: real => 'a => 'a
  1147 and norm :: 'a => real.
  1148 
  1149 * Real: New constant of_real :: real => 'a::real_algebra_1 injects
  1150 from reals into other types. The overloaded constant Reals :: 'a set
  1151 is now defined as range of_real; potential INCOMPATIBILITY.
  1152 
  1153 * Real: ML code generation is supported now and hence also quickcheck.
  1154 Reals are implemented as arbitrary precision rationals.
  1155 
  1156 * Hyperreal: Several constants that previously worked only for the
  1157 reals have been generalized, so they now work over arbitrary vector
  1158 spaces. Type annotations may need to be added in some cases; potential
  1159 INCOMPATIBILITY.
  1160 
  1161   Infinitesimal  :: ('a::real_normed_vector) star set
  1162   HFinite        :: ('a::real_normed_vector) star set
  1163   HInfinite      :: ('a::real_normed_vector) star set
  1164   approx         :: ('a::real_normed_vector) star => 'a star => bool
  1165   monad          :: ('a::real_normed_vector) star => 'a star set
  1166   galaxy         :: ('a::real_normed_vector) star => 'a star set
  1167   (NS)LIMSEQ     :: [nat => 'a::real_normed_vector, 'a] => bool
  1168   (NS)convergent :: (nat => 'a::real_normed_vector) => bool
  1169   (NS)Bseq       :: (nat => 'a::real_normed_vector) => bool
  1170   (NS)Cauchy     :: (nat => 'a::real_normed_vector) => bool
  1171   (NS)LIM        :: ['a::real_normed_vector => 'b::real_normed_vector, 'a, 'b] => bool
  1172   is(NS)Cont     :: ['a::real_normed_vector => 'b::real_normed_vector, 'a] => bool
  1173   deriv          :: ['a::real_normed_field => 'a, 'a, 'a] => bool
  1174   sgn            :: 'a::real_normed_vector => 'a
  1175   exp            :: 'a::{recpower,real_normed_field,banach} => 'a
  1176 
  1177 * Complex: Some complex-specific constants are now abbreviations for
  1178 overloaded ones: complex_of_real = of_real, cmod = norm, hcmod =
  1179 hnorm.  Other constants have been entirely removed in favor of the
  1180 polymorphic versions (INCOMPATIBILITY):
  1181 
  1182   approx        <-- capprox
  1183   HFinite       <-- CFinite
  1184   HInfinite     <-- CInfinite
  1185   Infinitesimal <-- CInfinitesimal
  1186   monad         <-- cmonad
  1187   galaxy        <-- cgalaxy
  1188   (NS)LIM       <-- (NS)CLIM, (NS)CRLIM
  1189   is(NS)Cont    <-- is(NS)Contc, is(NS)contCR
  1190   (ns)deriv     <-- (ns)cderiv
  1191 
  1192 
  1193 *** ML ***
  1194 
  1195 * Generic arithmetic modules: Tools/integer.ML, Tools/rat.ML, Tools/float.ML
  1196 
  1197 * Context data interfaces (Theory/Proof/GenericDataFun): removed
  1198 name/print, uninitialized data defaults to ad-hoc copy of empty value,
  1199 init only required for impure data.  INCOMPATIBILITY: empty really
  1200 need to be empty (no dependencies on theory content!)
  1201 
  1202 * ML within Isar: antiquotations allow to embed statically-checked
  1203 formal entities in the source, referring to the context available at
  1204 compile-time.  For example:
  1205 
  1206 ML {* @{typ "'a => 'b"} *}
  1207 ML {* @{term "%x. x"} *}
  1208 ML {* @{prop "x == y"} *}
  1209 ML {* @{ctyp "'a => 'b"} *}
  1210 ML {* @{cterm "%x. x"} *}
  1211 ML {* @{cprop "x == y"} *}
  1212 ML {* @{thm asm_rl} *}
  1213 ML {* @{thms asm_rl} *}
  1214 ML {* @{const_name c} *}
  1215 ML {* @{const_syntax c} *}
  1216 ML {* @{context} *}
  1217 ML {* @{theory} *}
  1218 ML {* @{theory Pure} *}
  1219 ML {* @{simpset} *}
  1220 ML {* @{claset} *}
  1221 ML {* @{clasimpset} *}
  1222 
  1223 The same works for sources being ``used'' within an Isar context.
  1224 
  1225 * ML in Isar: improved error reporting; extra verbosity with
  1226 Toplevel.debug enabled.
  1227 
  1228 * Pure/library:
  1229 
  1230   val burrow: ('a list -> 'b list) -> 'a list list -> 'b list list
  1231   val fold_burrow: ('a list -> 'c -> 'b list * 'd) -> 'a list list -> 'c -> 'b list list * 'd
  1232 
  1233 The semantics of "burrow" is: "take a function with *simulatanously*
  1234 transforms a list of value, and apply it *simulatanously* to a list of
  1235 list of values of the appropriate type". Compare this with "map" which
  1236 would *not* apply its argument function simulatanously but in
  1237 sequence; "fold_burrow" has an additional context.
  1238 
  1239 * Pure/library: functions map2 and fold2 with curried syntax for
  1240 simultanous mapping and folding:
  1241 
  1242     val map2: ('a -> 'b -> 'c) -> 'a list -> 'b list -> 'c list
  1243     val fold2: ('a -> 'b -> 'c -> 'c) -> 'a list -> 'b list -> 'c -> 'c
  1244 
  1245 * Pure/library: indexed lists - some functions in the Isabelle library
  1246 treating lists over 'a as finite mappings from [0...n] to 'a have been
  1247 given more convenient names and signatures reminiscent of similar
  1248 functions for alists, tables, etc:
  1249 
  1250   val nth: 'a list -> int -> 'a 
  1251   val nth_map: int -> ('a -> 'a) -> 'a list -> 'a list
  1252   val fold_index: (int * 'a -> 'b -> 'b) -> 'a list -> 'b -> 'b
  1253 
  1254 Note that fold_index starts counting at index 0, not 1 like foldln
  1255 used to.
  1256 
  1257 * Pure/library: added general ``divide_and_conquer'' combinator on
  1258 lists.
  1259 
  1260 * Pure/General/table.ML: the join operations now works via exceptions
  1261 DUP/SAME instead of type option.  This is simpler in simple cases, and
  1262 admits slightly more efficient complex applications.
  1263 
  1264 * Pure: datatype Context.generic joins theory/Proof.context and
  1265 provides some facilities for code that works in either kind of
  1266 context, notably GenericDataFun for uniform theory and proof data.
  1267 
  1268 * Pure: 'advanced' translation functions (parse_translation etc.) now
  1269 use Context.generic instead of just theory.
  1270 
  1271 * Pure: simplified internal attribute type, which is now always
  1272 Context.generic * thm -> Context.generic * thm.  Global (theory)
  1273 vs. local (Proof.context) attributes have been discontinued, while
  1274 minimizing code duplication.  Thm.rule_attribute and
  1275 Thm.declaration_attribute build canonical attributes; see also
  1276 structure Context for further operations on Context.generic, notably
  1277 GenericDataFun.  INCOMPATIBILITY, need to adapt attribute type
  1278 declarations and definitions.
  1279 
  1280 * Pure/kernel: consts certification ignores sort constraints given in
  1281 signature declarations.  (This information is not relevant to the
  1282 logic, but only for type inference.)  IMPORTANT INTERNAL CHANGE,
  1283 potential INCOMPATIBILITY.
  1284 
  1285 * Pure: axiomatic type classes are now purely definitional, with
  1286 explicit proofs of class axioms and super class relations performed
  1287 internally.  See Pure/axclass.ML for the main internal interfaces --
  1288 notably AxClass.define_class supercedes AxClass.add_axclass, and
  1289 AxClass.axiomatize_class/classrel/arity supercede
  1290 Sign.add_classes/classrel/arities.
  1291 
  1292 * Pure/Isar: Args/Attrib parsers operate on Context.generic --
  1293 global/local versions on theory vs. Proof.context have been
  1294 discontinued; Attrib.syntax and Method.syntax have been adapted
  1295 accordingly.  INCOMPATIBILITY, need to adapt parser expressions for
  1296 attributes, methods, etc.
  1297 
  1298 * Pure: several functions of signature "... -> theory -> theory * ..."
  1299 have been reoriented to "... -> theory -> ... * theory" in order to
  1300 allow natural usage in combination with the ||>, ||>>, |-> and
  1301 fold_map combinators.
  1302 
  1303 * Pure: official theorem names (closed derivations) and additional
  1304 comments (tags) are now strictly separate.  Name hints -- which are
  1305 maintained as tags -- may be attached any time without affecting the
  1306 derivation.
  1307 
  1308 * Pure: primitive rule lift_rule now takes goal cterm instead of an
  1309 actual goal state (thm).  Use Thm.lift_rule (Thm.cprem_of st i) to
  1310 achieve the old behaviour.
  1311 
  1312 * Pure: the "Goal" constant is now called "prop", supporting a
  1313 slightly more general idea of ``protecting'' meta-level rule
  1314 statements.
  1315 
  1316 * Pure: Logic.(un)varify only works in a global context, which is now
  1317 enforced instead of silently assumed.  INCOMPATIBILITY, may use
  1318 Logic.legacy_(un)varify as temporary workaround.
  1319 
  1320 * Pure: structure Name provides scalable operations for generating
  1321 internal variable names, notably Name.variants etc.  This replaces
  1322 some popular functions from term.ML:
  1323 
  1324   Term.variant		->  Name.variant
  1325   Term.variantlist	->  Name.variant_list  (*canonical argument order*)
  1326   Term.invent_names	->  Name.invent_list
  1327 
  1328 Note that low-level renaming rarely occurs in new code -- operations
  1329 from structure Variable are used instead (see below).
  1330 
  1331 * Pure: structure Variable provides fundamental operations for proper
  1332 treatment of fixed/schematic variables in a context.  For example,
  1333 Variable.import introduces fixes for schematics of given facts and
  1334 Variable.export reverses the effect (up to renaming) -- this replaces
  1335 various freeze_thaw operations.
  1336 
  1337 * Pure: structure Goal provides simple interfaces for
  1338 init/conclude/finish and tactical prove operations (replacing former
  1339 Tactic.prove).  Goal.prove is the canonical way to prove results
  1340 within a given context; Goal.prove_global is a degraded version for
  1341 theory level goals, including a global Drule.standard.  Note that
  1342 OldGoals.prove_goalw_cterm has long been obsolete, since it is
  1343 ill-behaved in a local proof context (e.g. with local fixes/assumes or
  1344 in a locale context).
  1345 
  1346 * Isar: simplified treatment of user-level errors, using exception
  1347 ERROR of string uniformly.  Function error now merely raises ERROR,
  1348 without any side effect on output channels.  The Isar toplevel takes
  1349 care of proper display of ERROR exceptions.  ML code may use plain
  1350 handle/can/try; cat_error may be used to concatenate errors like this:
  1351 
  1352   ... handle ERROR msg => cat_error msg "..."
  1353 
  1354 Toplevel ML code (run directly or through the Isar toplevel) may be
  1355 embedded into the Isar toplevel with exception display/debug like
  1356 this:
  1357 
  1358   Isar.toplevel (fn () => ...)
  1359 
  1360 INCOMPATIBILITY, removed special transform_error facilities, removed
  1361 obsolete variants of user-level exceptions (ERROR_MESSAGE,
  1362 Context.PROOF, ProofContext.CONTEXT, Proof.STATE, ProofHistory.FAIL)
  1363 -- use plain ERROR instead.
  1364 
  1365 * Isar: theory setup now has type (theory -> theory), instead of a
  1366 list.  INCOMPATIBILITY, may use #> to compose setup functions.
  1367 
  1368 * Isar: installed ML toplevel pretty printer for type Proof.context,
  1369 subject to ProofContext.debug/verbose flags.
  1370 
  1371 * Isar: Toplevel.theory_to_proof admits transactions that modify the
  1372 theory before entering a proof state.  Transactions now always see a
  1373 quasi-functional intermediate checkpoint, both in interactive and
  1374 batch mode.
  1375 
  1376 * Simplifier: the simpset of a running simplification process now
  1377 contains a proof context (cf. Simplifier.the_context), which is the
  1378 very context that the initial simpset has been retrieved from (by
  1379 simpset_of/local_simpset_of).  Consequently, all plug-in components
  1380 (solver, looper etc.) may depend on arbitrary proof data.
  1381 
  1382 * Simplifier.inherit_context inherits the proof context (plus the
  1383 local bounds) of the current simplification process; any simproc
  1384 etc. that calls the Simplifier recursively should do this!  Removed
  1385 former Simplifier.inherit_bounds, which is already included here --
  1386 INCOMPATIBILITY.  Tools based on low-level rewriting may even have to
  1387 specify an explicit context using Simplifier.context/theory_context.
  1388 
  1389 * Simplifier/Classical Reasoner: more abstract interfaces
  1390 change_simpset/claset for modifying the simpset/claset reference of a
  1391 theory; raw versions simpset/claset_ref etc. have been discontinued --
  1392 INCOMPATIBILITY.
  1393 
  1394 * Provers: more generic wrt. syntax of object-logics, avoid hardwired
  1395 "Trueprop" etc.
  1396 
  1397 
  1398 *** System ***
  1399 
  1400 * settings: ML_IDENTIFIER -- which is appended to user specific heap
  1401 locations -- now includes the Isabelle version identifier as well.
  1402 This simplifies use of multiple Isabelle installations.
  1403 
  1404 * isabelle-process: option -S (secure mode) disables some critical
  1405 operations, notably runtime compilation and evaluation of ML source
  1406 code.
  1407 
  1408 
  1409 New in Isabelle2005 (October 2005)
  1410 ----------------------------------
  1411 
  1412 *** General ***
  1413 
  1414 * Theory headers: the new header syntax for Isar theories is
  1415 
  1416   theory <name>
  1417   imports <theory1> ... <theoryN>
  1418   uses <file1> ... <fileM>
  1419   begin
  1420 
  1421 where the 'uses' part is optional.  The previous syntax
  1422 
  1423   theory <name> = <theory1> + ... + <theoryN>:
  1424 
  1425 will disappear in the next release.  Use isatool fixheaders to convert
  1426 existing theory files.  Note that there is no change in ancient
  1427 non-Isar theories now, but these will disappear soon.
  1428 
  1429 * Theory loader: parent theories can now also be referred to via
  1430 relative and absolute paths.
  1431 
  1432 * Command 'find_theorems' searches for a list of criteria instead of a
  1433 list of constants. Known criteria are: intro, elim, dest, name:string,
  1434 simp:term, and any term. Criteria can be preceded by '-' to select
  1435 theorems that do not match. Intro, elim, dest select theorems that
  1436 match the current goal, name:s selects theorems whose fully qualified
  1437 name contain s, and simp:term selects all simplification rules whose
  1438 lhs match term.  Any other term is interpreted as pattern and selects
  1439 all theorems matching the pattern. Available in ProofGeneral under
  1440 'ProofGeneral -> Find Theorems' or C-c C-f.  Example:
  1441 
  1442   C-c C-f (100) "(_::nat) + _ + _" intro -name: "HOL."
  1443 
  1444 prints the last 100 theorems matching the pattern "(_::nat) + _ + _",
  1445 matching the current goal as introduction rule and not having "HOL."
  1446 in their name (i.e. not being defined in theory HOL).
  1447 
  1448 * Command 'thms_containing' has been discontinued in favour of
  1449 'find_theorems'; INCOMPATIBILITY.
  1450 
  1451 * Communication with Proof General is now 8bit clean, which means that
  1452 Unicode text in UTF-8 encoding may be used within theory texts (both
  1453 formal and informal parts).  Cf. option -U of the Isabelle Proof
  1454 General interface.  Here are some simple examples (cf. src/HOL/ex):
  1455 
  1456   http://isabelle.in.tum.de/library/HOL/ex/Hebrew.html
  1457   http://isabelle.in.tum.de/library/HOL/ex/Chinese.html
  1458 
  1459 * Improved efficiency of the Simplifier and, to a lesser degree, the
  1460 Classical Reasoner.  Typical big applications run around 2 times
  1461 faster.
  1462 
  1463 
  1464 *** Document preparation ***
  1465 
  1466 * Commands 'display_drafts' and 'print_drafts' perform simple output
  1467 of raw sources.  Only those symbols that do not require additional
  1468 LaTeX packages (depending on comments in isabellesym.sty) are
  1469 displayed properly, everything else is left verbatim.  isatool display
  1470 and isatool print are used as front ends (these are subject to the
  1471 DVI/PDF_VIEWER and PRINT_COMMAND settings, respectively).
  1472 
  1473 * Command tags control specific markup of certain regions of text,
  1474 notably folding and hiding.  Predefined tags include "theory" (for
  1475 theory begin and end), "proof" for proof commands, and "ML" for
  1476 commands involving ML code; the additional tags "visible" and
  1477 "invisible" are unused by default.  Users may give explicit tag
  1478 specifications in the text, e.g. ''by %invisible (auto)''.  The
  1479 interpretation of tags is determined by the LaTeX job during document
  1480 preparation: see option -V of isatool usedir, or options -n and -t of
  1481 isatool document, or even the LaTeX macros \isakeeptag, \isafoldtag,
  1482 \isadroptag.
  1483 
  1484 Several document versions may be produced at the same time via isatool
  1485 usedir (the generated index.html will link all of them).  Typical
  1486 specifications include ''-V document=theory,proof,ML'' to present
  1487 theory/proof/ML parts faithfully, ''-V outline=/proof,/ML'' to fold
  1488 proof and ML commands, and ''-V mutilated=-theory,-proof,-ML'' to omit
  1489 these parts without any formal replacement text.  The Isabelle site
  1490 default settings produce ''document'' and ''outline'' versions as
  1491 specified above.
  1492 
  1493 * Several new antiquotations:
  1494 
  1495   @{term_type term} prints a term with its type annotated;
  1496 
  1497   @{typeof term} prints the type of a term;
  1498 
  1499   @{const const} is the same as @{term const}, but checks that the
  1500   argument is a known logical constant;
  1501 
  1502   @{term_style style term} and @{thm_style style thm} print a term or
  1503   theorem applying a "style" to it
  1504 
  1505   @{ML text}
  1506 
  1507 Predefined styles are 'lhs' and 'rhs' printing the lhs/rhs of
  1508 definitions, equations, inequations etc., 'concl' printing only the
  1509 conclusion of a meta-logical statement theorem, and 'prem1' .. 'prem19'
  1510 to print the specified premise.  TermStyle.add_style provides an ML
  1511 interface for introducing further styles.  See also the "LaTeX Sugar"
  1512 document practical applications.  The ML antiquotation prints
  1513 type-checked ML expressions verbatim.
  1514 
  1515 * Markup commands 'chapter', 'section', 'subsection', 'subsubsection',
  1516 and 'text' support optional locale specification '(in loc)', which
  1517 specifies the default context for interpreting antiquotations.  For
  1518 example: 'text (in lattice) {* @{thm inf_assoc}*}'.
  1519 
  1520 * Option 'locale=NAME' of antiquotations specifies an alternative
  1521 context interpreting the subsequent argument.  For example: @{thm
  1522 [locale=lattice] inf_assoc}.
  1523 
  1524 * Proper output of proof terms (@{prf ...} and @{full_prf ...}) within
  1525 a proof context.
  1526 
  1527 * Proper output of antiquotations for theory commands involving a
  1528 proof context (such as 'locale' or 'theorem (in loc) ...').
  1529 
  1530 * Delimiters of outer tokens (string etc.) now produce separate LaTeX
  1531 macros (\isachardoublequoteopen, isachardoublequoteclose etc.).
  1532 
  1533 * isatool usedir: new option -C (default true) controls whether option
  1534 -D should include a copy of the original document directory; -C false
  1535 prevents unwanted effects such as copying of administrative CVS data.
  1536 
  1537 
  1538 *** Pure ***
  1539 
  1540 * Considerably improved version of 'constdefs' command.  Now performs
  1541 automatic type-inference of declared constants; additional support for
  1542 local structure declarations (cf. locales and HOL records), see also
  1543 isar-ref manual.  Potential INCOMPATIBILITY: need to observe strictly
  1544 sequential dependencies of definitions within a single 'constdefs'
  1545 section; moreover, the declared name needs to be an identifier.  If
  1546 all fails, consider to fall back on 'consts' and 'defs' separately.
  1547 
  1548 * Improved indexed syntax and implicit structures.  First of all,
  1549 indexed syntax provides a notational device for subscripted
  1550 application, using the new syntax \<^bsub>term\<^esub> for arbitrary
  1551 expressions.  Secondly, in a local context with structure
  1552 declarations, number indexes \<^sub>n or the empty index (default
  1553 number 1) refer to a certain fixed variable implicitly; option
  1554 show_structs controls printing of implicit structures.  Typical
  1555 applications of these concepts involve record types and locales.
  1556 
  1557 * New command 'no_syntax' removes grammar declarations (and
  1558 translations) resulting from the given syntax specification, which is
  1559 interpreted in the same manner as for the 'syntax' command.
  1560 
  1561 * 'Advanced' translation functions (parse_translation etc.) may depend
  1562 on the signature of the theory context being presently used for
  1563 parsing/printing, see also isar-ref manual.
  1564 
  1565 * Improved 'oracle' command provides a type-safe interface to turn an
  1566 ML expression of type theory -> T -> term into a primitive rule of
  1567 type theory -> T -> thm (i.e. the functionality of Thm.invoke_oracle
  1568 is already included here); see also FOL/ex/IffExample.thy;
  1569 INCOMPATIBILITY.
  1570 
  1571 * axclass: name space prefix for class "c" is now "c_class" (was "c"
  1572 before); "cI" is no longer bound, use "c.intro" instead.
  1573 INCOMPATIBILITY.  This change avoids clashes of fact bindings for
  1574 axclasses vs. locales.
  1575 
  1576 * Improved internal renaming of symbolic identifiers -- attach primes
  1577 instead of base 26 numbers.
  1578 
  1579 * New flag show_question_marks controls printing of leading question
  1580 marks in schematic variable names.
  1581 
  1582 * In schematic variable names, *any* symbol following \<^isub> or
  1583 \<^isup> is now treated as part of the base name.  For example, the
  1584 following works without printing of awkward ".0" indexes:
  1585 
  1586   lemma "x\<^isub>1 = x\<^isub>2 ==> x\<^isub>2 = x\<^isub>1"
  1587     by simp
  1588 
  1589 * Inner syntax includes (*(*nested*) comments*).
  1590 
  1591 * Pretty printer now supports unbreakable blocks, specified in mixfix
  1592 annotations as "(00...)".
  1593 
  1594 * Clear separation of logical types and nonterminals, where the latter
  1595 may only occur in 'syntax' specifications or type abbreviations.
  1596 Before that distinction was only partially implemented via type class
  1597 "logic" vs. "{}".  Potential INCOMPATIBILITY in rare cases of improper
  1598 use of 'types'/'consts' instead of 'nonterminals'/'syntax'.  Some very
  1599 exotic syntax specifications may require further adaption
  1600 (e.g. Cube/Cube.thy).
  1601 
  1602 * Removed obsolete type class "logic", use the top sort {} instead.
  1603 Note that non-logical types should be declared as 'nonterminals'
  1604 rather than 'types'.  INCOMPATIBILITY for new object-logic
  1605 specifications.
  1606 
  1607 * Attributes 'induct' and 'cases': type or set names may now be
  1608 locally fixed variables as well.
  1609 
  1610 * Simplifier: can now control the depth to which conditional rewriting
  1611 is traced via the PG menu Isabelle -> Settings -> Trace Simp Depth
  1612 Limit.
  1613 
  1614 * Simplifier: simplification procedures may now take the current
  1615 simpset into account (cf. Simplifier.simproc(_i) / mk_simproc
  1616 interface), which is very useful for calling the Simplifier
  1617 recursively.  Minor INCOMPATIBILITY: the 'prems' argument of simprocs
  1618 is gone -- use prems_of_ss on the simpset instead.  Moreover, the
  1619 low-level mk_simproc no longer applies Logic.varify internally, to
  1620 allow for use in a context of fixed variables.
  1621 
  1622 * thin_tac now works even if the assumption being deleted contains !!
  1623 or ==>.  More generally, erule now works even if the major premise of
  1624 the elimination rule contains !! or ==>.
  1625 
  1626 * Method 'rules' has been renamed to 'iprover'. INCOMPATIBILITY.
  1627 
  1628 * Reorganized bootstrapping of the Pure theories; CPure is now derived
  1629 from Pure, which contains all common declarations already.  Both
  1630 theories are defined via plain Isabelle/Isar .thy files.
  1631 INCOMPATIBILITY: elements of CPure (such as the CPure.intro /
  1632 CPure.elim / CPure.dest attributes) now appear in the Pure name space;
  1633 use isatool fixcpure to adapt your theory and ML sources.
  1634 
  1635 * New syntax 'name(i-j, i-, i, ...)' for referring to specific
  1636 selections of theorems in named facts via index ranges.
  1637 
  1638 * 'print_theorems': in theory mode, really print the difference
  1639 wrt. the last state (works for interactive theory development only),
  1640 in proof mode print all local facts (cf. 'print_facts');
  1641 
  1642 * 'hide': option '(open)' hides only base names.
  1643 
  1644 * More efficient treatment of intermediate checkpoints in interactive
  1645 theory development.
  1646 
  1647 * Code generator is now invoked via code_module (incremental code
  1648 generation) and code_library (modular code generation, ML structures
  1649 for each theory).  INCOMPATIBILITY: new keywords 'file' and 'contains'
  1650 must be quoted when used as identifiers.
  1651 
  1652 * New 'value' command for reading, evaluating and printing terms using
  1653 the code generator.  INCOMPATIBILITY: command keyword 'value' must be
  1654 quoted when used as identifier.
  1655 
  1656 
  1657 *** Locales ***
  1658 
  1659 * New commands for the interpretation of locale expressions in
  1660 theories (1), locales (2) and proof contexts (3).  These generate
  1661 proof obligations from the expression specification.  After the
  1662 obligations have been discharged, theorems of the expression are added
  1663 to the theory, target locale or proof context.  The synopsis of the
  1664 commands is a follows:
  1665 
  1666   (1) interpretation expr inst
  1667   (2) interpretation target < expr
  1668   (3) interpret expr inst
  1669 
  1670 Interpretation in theories and proof contexts require a parameter
  1671 instantiation of terms from the current context.  This is applied to
  1672 specifications and theorems of the interpreted expression.
  1673 Interpretation in locales only permits parameter renaming through the
  1674 locale expression.  Interpretation is smart in that interpretations
  1675 that are active already do not occur in proof obligations, neither are
  1676 instantiated theorems stored in duplicate.  Use 'print_interps' to
  1677 inspect active interpretations of a particular locale.  For details,
  1678 see the Isar Reference manual.  Examples can be found in
  1679 HOL/Finite_Set.thy and HOL/Algebra/UnivPoly.thy.
  1680 
  1681 INCOMPATIBILITY: former 'instantiate' has been withdrawn, use
  1682 'interpret' instead.
  1683 
  1684 * New context element 'constrains' for adding type constraints to
  1685 parameters.
  1686 
  1687 * Context expressions: renaming of parameters with syntax
  1688 redeclaration.
  1689 
  1690 * Locale declaration: 'includes' disallowed.
  1691 
  1692 * Proper static binding of attribute syntax -- i.e. types / terms /
  1693 facts mentioned as arguments are always those of the locale definition
  1694 context, independently of the context of later invocations.  Moreover,
  1695 locale operations (renaming and type / term instantiation) are applied
  1696 to attribute arguments as expected.
  1697 
  1698 INCOMPATIBILITY of the ML interface: always pass Attrib.src instead of
  1699 actual attributes; rare situations may require Attrib.attribute to
  1700 embed those attributes into Attrib.src that lack concrete syntax.
  1701 Attribute implementations need to cooperate properly with the static
  1702 binding mechanism.  Basic parsers Args.XXX_typ/term/prop and
  1703 Attrib.XXX_thm etc. already do the right thing without further
  1704 intervention.  Only unusual applications -- such as "where" or "of"
  1705 (cf. src/Pure/Isar/attrib.ML), which process arguments depending both
  1706 on the context and the facts involved -- may have to assign parsed
  1707 values to argument tokens explicitly.
  1708 
  1709 * Changed parameter management in theorem generation for long goal
  1710 statements with 'includes'.  INCOMPATIBILITY: produces a different
  1711 theorem statement in rare situations.
  1712 
  1713 * Locale inspection command 'print_locale' omits notes elements.  Use
  1714 'print_locale!' to have them included in the output.
  1715 
  1716 
  1717 *** Provers ***
  1718 
  1719 * Provers/hypsubst.ML: improved version of the subst method, for
  1720 single-step rewriting: it now works in bound variable contexts. New is
  1721 'subst (asm)', for rewriting an assumption.  INCOMPATIBILITY: may
  1722 rewrite a different subterm than the original subst method, which is
  1723 still available as 'simplesubst'.
  1724 
  1725 * Provers/quasi.ML: new transitivity reasoners for transitivity only
  1726 and quasi orders.
  1727 
  1728 * Provers/trancl.ML: new transitivity reasoner for transitive and
  1729 reflexive-transitive closure of relations.
  1730 
  1731 * Provers/blast.ML: new reference depth_limit to make blast's depth
  1732 limit (previously hard-coded with a value of 20) user-definable.
  1733 
  1734 * Provers/simplifier.ML has been moved to Pure, where Simplifier.setup
  1735 is peformed already.  Object-logics merely need to finish their
  1736 initial simpset configuration as before.  INCOMPATIBILITY.
  1737 
  1738 
  1739 *** HOL ***
  1740 
  1741 * Symbolic syntax of Hilbert Choice Operator is now as follows:
  1742 
  1743   syntax (epsilon)
  1744     "_Eps" :: "[pttrn, bool] => 'a"    ("(3\<some>_./ _)" [0, 10] 10)
  1745 
  1746 The symbol \<some> is displayed as the alternative epsilon of LaTeX
  1747 and x-symbol; use option '-m epsilon' to get it actually printed.
  1748 Moreover, the mathematically important symbolic identifier \<epsilon>
  1749 becomes available as variable, constant etc.  INCOMPATIBILITY,
  1750 
  1751 * "x > y" abbreviates "y < x" and "x >= y" abbreviates "y <= x".
  1752 Similarly for all quantifiers: "ALL x > y" etc.  The x-symbol for >=
  1753 is \<ge>. New transitivity rules have been added to HOL/Orderings.thy to
  1754 support corresponding Isar calculations.
  1755 
  1756 * "{x:A. P}" abbreviates "{x. x:A & P}", and similarly for "\<in>"
  1757 instead of ":".
  1758 
  1759 * theory SetInterval: changed the syntax for open intervals:
  1760 
  1761   Old       New
  1762   {..n(}    {..<n}
  1763   {)n..}    {n<..}
  1764   {m..n(}   {m..<n}
  1765   {)m..n}   {m<..n}
  1766   {)m..n(}  {m<..<n}
  1767 
  1768 The old syntax is still supported but will disappear in the next
  1769 release.  For conversion use the following Emacs search and replace
  1770 patterns (these are not perfect but work quite well):
  1771 
  1772   {)\([^\.]*\)\.\.  ->  {\1<\.\.}
  1773   \.\.\([^(}]*\)(}  ->  \.\.<\1}
  1774 
  1775 * Theory Commutative_Ring (in Library): method comm_ring for proving
  1776 equalities in commutative rings; method 'algebra' provides a generic
  1777 interface.
  1778 
  1779 * Theory Finite_Set: changed the syntax for 'setsum', summation over
  1780 finite sets: "setsum (%x. e) A", which used to be "\<Sum>x:A. e", is
  1781 now either "SUM x:A. e" or "\<Sum>x \<in> A. e". The bound variable can
  1782 be a tuple pattern.
  1783 
  1784 Some new syntax forms are available:
  1785 
  1786   "\<Sum>x | P. e"      for     "setsum (%x. e) {x. P}"
  1787   "\<Sum>x = a..b. e"   for     "setsum (%x. e) {a..b}"
  1788   "\<Sum>x = a..<b. e"  for     "setsum (%x. e) {a..<b}"
  1789   "\<Sum>x < k. e"      for     "setsum (%x. e) {..<k}"
  1790 
  1791 The latter form "\<Sum>x < k. e" used to be based on a separate
  1792 function "Summation", which has been discontinued.
  1793 
  1794 * theory Finite_Set: in structured induction proofs, the insert case
  1795 is now 'case (insert x F)' instead of the old counterintuitive 'case
  1796 (insert F x)'.
  1797 
  1798 * The 'refute' command has been extended to support a much larger
  1799 fragment of HOL, including axiomatic type classes, constdefs and
  1800 typedefs, inductive datatypes and recursion.
  1801 
  1802 * New tactics 'sat' and 'satx' to prove propositional tautologies.
  1803 Requires zChaff with proof generation to be installed.  See
  1804 HOL/ex/SAT_Examples.thy for examples.
  1805 
  1806 * Datatype induction via method 'induct' now preserves the name of the
  1807 induction variable. For example, when proving P(xs::'a list) by
  1808 induction on xs, the induction step is now P(xs) ==> P(a#xs) rather
  1809 than P(list) ==> P(a#list) as previously.  Potential INCOMPATIBILITY
  1810 in unstructured proof scripts.
  1811 
  1812 * Reworked implementation of records.  Improved scalability for
  1813 records with many fields, avoiding performance problems for type
  1814 inference. Records are no longer composed of nested field types, but
  1815 of nested extension types. Therefore the record type only grows linear
  1816 in the number of extensions and not in the number of fields.  The
  1817 top-level (users) view on records is preserved.  Potential
  1818 INCOMPATIBILITY only in strange cases, where the theory depends on the
  1819 old record representation. The type generated for a record is called
  1820 <record_name>_ext_type.
  1821 
  1822 Flag record_quick_and_dirty_sensitive can be enabled to skip the
  1823 proofs triggered by a record definition or a simproc (if
  1824 quick_and_dirty is enabled).  Definitions of large records can take
  1825 quite long.
  1826 
  1827 New simproc record_upd_simproc for simplification of multiple record
  1828 updates enabled by default.  Moreover, trivial updates are also
  1829 removed: r(|x := x r|) = r.  INCOMPATIBILITY: old proofs break
  1830 occasionally, since simplification is more powerful by default.
  1831 
  1832 * typedef: proper support for polymorphic sets, which contain extra
  1833 type-variables in the term.
  1834 
  1835 * Simplifier: automatically reasons about transitivity chains
  1836 involving "trancl" (r^+) and "rtrancl" (r^*) by setting up tactics
  1837 provided by Provers/trancl.ML as additional solvers.  INCOMPATIBILITY:
  1838 old proofs break occasionally as simplification may now solve more
  1839 goals than previously.
  1840 
  1841 * Simplifier: converts x <= y into x = y if assumption y <= x is
  1842 present.  Works for all partial orders (class "order"), in particular
  1843 numbers and sets.  For linear orders (e.g. numbers) it treats ~ x < y
  1844 just like y <= x.
  1845 
  1846 * Simplifier: new simproc for "let x = a in f x".  If a is a free or
  1847 bound variable or a constant then the let is unfolded.  Otherwise
  1848 first a is simplified to b, and then f b is simplified to g. If
  1849 possible we abstract b from g arriving at "let x = b in h x",
  1850 otherwise we unfold the let and arrive at g.  The simproc can be
  1851 enabled/disabled by the reference use_let_simproc.  Potential
  1852 INCOMPATIBILITY since simplification is more powerful by default.
  1853 
  1854 * Classical reasoning: the meson method now accepts theorems as arguments.
  1855 
  1856 * Prover support: pre-release of the Isabelle-ATP linkup, which runs background
  1857 jobs to provide advice on the provability of subgoals.
  1858 
  1859 * Theory OrderedGroup and Ring_and_Field: various additions and
  1860 improvements to faciliate calculations involving equalities and
  1861 inequalities.
  1862 
  1863 The following theorems have been eliminated or modified
  1864 (INCOMPATIBILITY):
  1865 
  1866   abs_eq             now named abs_of_nonneg
  1867   abs_of_ge_0        now named abs_of_nonneg
  1868   abs_minus_eq       now named abs_of_nonpos
  1869   imp_abs_id         now named abs_of_nonneg
  1870   imp_abs_neg_id     now named abs_of_nonpos
  1871   mult_pos           now named mult_pos_pos
  1872   mult_pos_le        now named mult_nonneg_nonneg
  1873   mult_pos_neg_le    now named mult_nonneg_nonpos
  1874   mult_pos_neg2_le   now named mult_nonneg_nonpos2
  1875   mult_neg           now named mult_neg_neg
  1876   mult_neg_le        now named mult_nonpos_nonpos
  1877 
  1878 * The following lemmas in Ring_and_Field have been added to the simplifier:
  1879      
  1880      zero_le_square
  1881      not_square_less_zero 
  1882 
  1883   The following lemmas have been deleted from Real/RealPow:
  1884   
  1885      realpow_zero_zero
  1886      realpow_two
  1887      realpow_less
  1888      zero_le_power
  1889      realpow_two_le
  1890      abs_realpow_two
  1891      realpow_two_abs     
  1892 
  1893 * Theory Parity: added rules for simplifying exponents.
  1894 
  1895 * Theory List:
  1896 
  1897 The following theorems have been eliminated or modified
  1898 (INCOMPATIBILITY):
  1899 
  1900   list_all_Nil       now named list_all.simps(1)
  1901   list_all_Cons      now named list_all.simps(2)
  1902   list_all_conv      now named list_all_iff
  1903   set_mem_eq         now named mem_iff
  1904 
  1905 * Theories SetsAndFunctions and BigO (see HOL/Library) support
  1906 asymptotic "big O" calculations.  See the notes in BigO.thy.
  1907 
  1908 
  1909 *** HOL-Complex ***
  1910 
  1911 * Theory RealDef: better support for embedding natural numbers and
  1912 integers in the reals.
  1913 
  1914 The following theorems have been eliminated or modified
  1915 (INCOMPATIBILITY):
  1916 
  1917   exp_ge_add_one_self  now requires no hypotheses
  1918   real_of_int_add      reversed direction of equality (use [symmetric])
  1919   real_of_int_minus    reversed direction of equality (use [symmetric])
  1920   real_of_int_diff     reversed direction of equality (use [symmetric])
  1921   real_of_int_mult     reversed direction of equality (use [symmetric])
  1922 
  1923 * Theory RComplete: expanded support for floor and ceiling functions.
  1924 
  1925 * Theory Ln is new, with properties of the natural logarithm
  1926 
  1927 * Hyperreal: There is a new type constructor "star" for making
  1928 nonstandard types.  The old type names are now type synonyms:
  1929 
  1930   hypreal = real star
  1931   hypnat = nat star
  1932   hcomplex = complex star
  1933 
  1934 * Hyperreal: Many groups of similarly-defined constants have been
  1935 replaced by polymorphic versions (INCOMPATIBILITY):
  1936 
  1937   star_of <-- hypreal_of_real, hypnat_of_nat, hcomplex_of_complex
  1938 
  1939   starset      <-- starsetNat, starsetC
  1940   *s*          <-- *sNat*, *sc*
  1941   starset_n    <-- starsetNat_n, starsetC_n
  1942   *sn*         <-- *sNatn*, *scn*
  1943   InternalSets <-- InternalNatSets, InternalCSets
  1944 
  1945   starfun      <-- starfun{Nat,Nat2,C,RC,CR}
  1946   *f*          <-- *fNat*, *fNat2*, *fc*, *fRc*, *fcR*
  1947   starfun_n    <-- starfun{Nat,Nat2,C,RC,CR}_n
  1948   *fn*         <-- *fNatn*, *fNat2n*, *fcn*, *fRcn*, *fcRn*
  1949   InternalFuns <-- InternalNatFuns, InternalNatFuns2, Internal{C,RC,CR}Funs
  1950 
  1951 * Hyperreal: Many type-specific theorems have been removed in favor of
  1952 theorems specific to various axiomatic type classes (INCOMPATIBILITY):
  1953 
  1954   add_commute <-- {hypreal,hypnat,hcomplex}_add_commute
  1955   add_assoc   <-- {hypreal,hypnat,hcomplex}_add_assocs
  1956   OrderedGroup.add_0 <-- {hypreal,hypnat,hcomplex}_add_zero_left
  1957   OrderedGroup.add_0_right <-- {hypreal,hcomplex}_add_zero_right
  1958   right_minus <-- hypreal_add_minus
  1959   left_minus <-- {hypreal,hcomplex}_add_minus_left
  1960   mult_commute <-- {hypreal,hypnat,hcomplex}_mult_commute
  1961   mult_assoc <-- {hypreal,hypnat,hcomplex}_mult_assoc
  1962   mult_1_left <-- {hypreal,hypnat}_mult_1, hcomplex_mult_one_left
  1963   mult_1_right <-- hcomplex_mult_one_right
  1964   mult_zero_left <-- hcomplex_mult_zero_left
  1965   left_distrib <-- {hypreal,hypnat,hcomplex}_add_mult_distrib
  1966   right_distrib <-- hypnat_add_mult_distrib2
  1967   zero_neq_one <-- {hypreal,hypnat,hcomplex}_zero_not_eq_one
  1968   right_inverse <-- hypreal_mult_inverse
  1969   left_inverse <-- hypreal_mult_inverse_left, hcomplex_mult_inv_left
  1970   order_refl <-- {hypreal,hypnat}_le_refl
  1971   order_trans <-- {hypreal,hypnat}_le_trans
  1972   order_antisym <-- {hypreal,hypnat}_le_anti_sym
  1973   order_less_le <-- {hypreal,hypnat}_less_le
  1974   linorder_linear <-- {hypreal,hypnat}_le_linear
  1975   add_left_mono <-- {hypreal,hypnat}_add_left_mono
  1976   mult_strict_left_mono <-- {hypreal,hypnat}_mult_less_mono2
  1977   add_nonneg_nonneg <-- hypreal_le_add_order
  1978 
  1979 * Hyperreal: Separate theorems having to do with type-specific
  1980 versions of constants have been merged into theorems that apply to the
  1981 new polymorphic constants (INCOMPATIBILITY):
  1982 
  1983   STAR_UNIV_set <-- {STAR_real,NatStar_real,STARC_complex}_set
  1984   STAR_empty_set <-- {STAR,NatStar,STARC}_empty_set
  1985   STAR_Un <-- {STAR,NatStar,STARC}_Un
  1986   STAR_Int <-- {STAR,NatStar,STARC}_Int
  1987   STAR_Compl <-- {STAR,NatStar,STARC}_Compl
  1988   STAR_subset <-- {STAR,NatStar,STARC}_subset
  1989   STAR_mem <-- {STAR,NatStar,STARC}_mem
  1990   STAR_mem_Compl <-- {STAR,STARC}_mem_Compl
  1991   STAR_diff <-- {STAR,STARC}_diff
  1992   STAR_star_of_image_subset <-- {STAR_hypreal_of_real, NatStar_hypreal_of_real,
  1993     STARC_hcomplex_of_complex}_image_subset
  1994   starset_n_Un <-- starset{Nat,C}_n_Un
  1995   starset_n_Int <-- starset{Nat,C}_n_Int
  1996   starset_n_Compl <-- starset{Nat,C}_n_Compl
  1997   starset_n_diff <-- starset{Nat,C}_n_diff
  1998   InternalSets_Un <-- Internal{Nat,C}Sets_Un
  1999   InternalSets_Int <-- Internal{Nat,C}Sets_Int
  2000   InternalSets_Compl <-- Internal{Nat,C}Sets_Compl
  2001   InternalSets_diff <-- Internal{Nat,C}Sets_diff
  2002   InternalSets_UNIV_diff <-- Internal{Nat,C}Sets_UNIV_diff
  2003   InternalSets_starset_n <-- Internal{Nat,C}Sets_starset{Nat,C}_n
  2004   starset_starset_n_eq <-- starset{Nat,C}_starset{Nat,C}_n_eq
  2005   starset_n_starset <-- starset{Nat,C}_n_starset{Nat,C}
  2006   starfun_n_starfun <-- starfun{Nat,Nat2,C,RC,CR}_n_starfun{Nat,Nat2,C,RC,CR}
  2007   starfun <-- starfun{Nat,Nat2,C,RC,CR}
  2008   starfun_mult <-- starfun{Nat,Nat2,C,RC,CR}_mult
  2009   starfun_add <-- starfun{Nat,Nat2,C,RC,CR}_add
  2010   starfun_minus <-- starfun{Nat,Nat2,C,RC,CR}_minus
  2011   starfun_diff <-- starfun{C,RC,CR}_diff
  2012   starfun_o <-- starfun{NatNat2,Nat2,_stafunNat,C,C_starfunRC,_starfunCR}_o
  2013   starfun_o2 <-- starfun{NatNat2,_stafunNat,C,C_starfunRC,_starfunCR}_o2
  2014   starfun_const_fun <-- starfun{Nat,Nat2,C,RC,CR}_const_fun
  2015   starfun_inverse <-- starfun{Nat,C,RC,CR}_inverse
  2016   starfun_eq <-- starfun{Nat,Nat2,C,RC,CR}_eq
  2017   starfun_eq_iff <-- starfun{C,RC,CR}_eq_iff
  2018   starfun_Id <-- starfunC_Id
  2019   starfun_approx <-- starfun{Nat,CR}_approx
  2020   starfun_capprox <-- starfun{C,RC}_capprox
  2021   starfun_abs <-- starfunNat_rabs
  2022   starfun_lambda_cancel <-- starfun{C,CR,RC}_lambda_cancel
  2023   starfun_lambda_cancel2 <-- starfun{C,CR,RC}_lambda_cancel2
  2024   starfun_mult_HFinite_approx <-- starfunCR_mult_HFinite_capprox
  2025   starfun_mult_CFinite_capprox <-- starfun{C,RC}_mult_CFinite_capprox
  2026   starfun_add_capprox <-- starfun{C,RC}_add_capprox
  2027   starfun_add_approx <-- starfunCR_add_approx
  2028   starfun_inverse_inverse <-- starfunC_inverse_inverse
  2029   starfun_divide <-- starfun{C,CR,RC}_divide
  2030   starfun_n <-- starfun{Nat,C}_n
  2031   starfun_n_mult <-- starfun{Nat,C}_n_mult
  2032   starfun_n_add <-- starfun{Nat,C}_n_add
  2033   starfun_n_add_minus <-- starfunNat_n_add_minus
  2034   starfun_n_const_fun <-- starfun{Nat,C}_n_const_fun
  2035   starfun_n_minus <-- starfun{Nat,C}_n_minus
  2036   starfun_n_eq <-- starfun{Nat,C}_n_eq
  2037 
  2038   star_n_add <-- {hypreal,hypnat,hcomplex}_add
  2039   star_n_minus <-- {hypreal,hcomplex}_minus
  2040   star_n_diff <-- {hypreal,hcomplex}_diff
  2041   star_n_mult <-- {hypreal,hcomplex}_mult
  2042   star_n_inverse <-- {hypreal,hcomplex}_inverse
  2043   star_n_le <-- {hypreal,hypnat}_le
  2044   star_n_less <-- {hypreal,hypnat}_less
  2045   star_n_zero_num <-- {hypreal,hypnat,hcomplex}_zero_num
  2046   star_n_one_num <-- {hypreal,hypnat,hcomplex}_one_num
  2047   star_n_abs <-- hypreal_hrabs
  2048   star_n_divide <-- hcomplex_divide
  2049 
  2050   star_of_add <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_add
  2051   star_of_minus <-- {hypreal_of_real,hcomplex_of_complex}_minus
  2052   star_of_diff <-- hypreal_of_real_diff
  2053   star_of_mult <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_mult
  2054   star_of_one <-- {hypreal_of_real,hcomplex_of_complex}_one
  2055   star_of_zero <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_zero
  2056   star_of_le <-- {hypreal_of_real,hypnat_of_nat}_le_iff
  2057   star_of_less <-- {hypreal_of_real,hypnat_of_nat}_less_iff
  2058   star_of_eq <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_eq_iff
  2059   star_of_inverse <-- {hypreal_of_real,hcomplex_of_complex}_inverse
  2060   star_of_divide <-- {hypreal_of_real,hcomplex_of_complex}_divide
  2061   star_of_of_nat <-- {hypreal_of_real,hcomplex_of_complex}_of_nat
  2062   star_of_of_int <-- {hypreal_of_real,hcomplex_of_complex}_of_int
  2063   star_of_number_of <-- {hypreal,hcomplex}_number_of
  2064   star_of_number_less <-- number_of_less_hypreal_of_real_iff
  2065   star_of_number_le <-- number_of_le_hypreal_of_real_iff
  2066   star_of_eq_number <-- hypreal_of_real_eq_number_of_iff
  2067   star_of_less_number <-- hypreal_of_real_less_number_of_iff
  2068   star_of_le_number <-- hypreal_of_real_le_number_of_iff
  2069   star_of_power <-- hypreal_of_real_power
  2070   star_of_eq_0 <-- hcomplex_of_complex_zero_iff
  2071 
  2072 * Hyperreal: new method "transfer" that implements the transfer
  2073 principle of nonstandard analysis. With a subgoal that mentions
  2074 nonstandard types like "'a star", the command "apply transfer"
  2075 replaces it with an equivalent one that mentions only standard types.
  2076 To be successful, all free variables must have standard types; non-
  2077 standard variables must have explicit universal quantifiers.
  2078 
  2079 * Hyperreal: A theory of Taylor series.
  2080 
  2081 
  2082 *** HOLCF ***
  2083 
  2084 * Discontinued special version of 'constdefs' (which used to support
  2085 continuous functions) in favor of the general Pure one with full
  2086 type-inference.
  2087 
  2088 * New simplification procedure for solving continuity conditions; it
  2089 is much faster on terms with many nested lambda abstractions (cubic
  2090 instead of exponential time).
  2091 
  2092 * New syntax for domain package: selector names are now optional.
  2093 Parentheses should be omitted unless argument is lazy, for example:
  2094 
  2095   domain 'a stream = cons "'a" (lazy "'a stream")
  2096 
  2097 * New command 'fixrec' for defining recursive functions with pattern
  2098 matching; defining multiple functions with mutual recursion is also
  2099 supported.  Patterns may include the constants cpair, spair, up, sinl,
  2100 sinr, or any data constructor defined by the domain package. The given
  2101 equations are proven as rewrite rules. See HOLCF/ex/Fixrec_ex.thy for
  2102 syntax and examples.
  2103 
  2104 * New commands 'cpodef' and 'pcpodef' for defining predicate subtypes
  2105 of cpo and pcpo types. Syntax is exactly like the 'typedef' command,
  2106 but the proof obligation additionally includes an admissibility
  2107 requirement. The packages generate instances of class cpo or pcpo,
  2108 with continuity and strictness theorems for Rep and Abs.
  2109 
  2110 * HOLCF: Many theorems have been renamed according to a more standard naming
  2111 scheme (INCOMPATIBILITY):
  2112 
  2113   foo_inject:  "foo$x = foo$y ==> x = y"
  2114   foo_eq:      "(foo$x = foo$y) = (x = y)"
  2115   foo_less:    "(foo$x << foo$y) = (x << y)"
  2116   foo_strict:  "foo$UU = UU"
  2117   foo_defined: "... ==> foo$x ~= UU"
  2118   foo_defined_iff: "(foo$x = UU) = (x = UU)"
  2119 
  2120 
  2121 *** ZF ***
  2122 
  2123 * ZF/ex: theories Group and Ring provide examples in abstract algebra,
  2124 including the First Isomorphism Theorem (on quotienting by the kernel
  2125 of a homomorphism).
  2126 
  2127 * ZF/Simplifier: install second copy of type solver that actually
  2128 makes use of TC rules declared to Isar proof contexts (or locales);
  2129 the old version is still required for ML proof scripts.
  2130 
  2131 
  2132 *** Cube ***
  2133 
  2134 * Converted to Isar theory format; use locales instead of axiomatic
  2135 theories.
  2136 
  2137 
  2138 *** ML ***
  2139 
  2140 * Pure/library.ML: added ##>, ##>>, #>> -- higher-order counterparts
  2141 for ||>, ||>>, |>>,
  2142 
  2143 * Pure/library.ML no longer defines its own option datatype, but uses
  2144 that of the SML basis, which has constructors NONE and SOME instead of
  2145 None and Some, as well as exception Option.Option instead of OPTION.
  2146 The functions the, if_none, is_some, is_none have been adapted
  2147 accordingly, while Option.map replaces apsome.
  2148 
  2149 * Pure/library.ML: the exception LIST has been given up in favour of
  2150 the standard exceptions Empty and Subscript, as well as
  2151 Library.UnequalLengths.  Function like Library.hd and Library.tl are
  2152 superceded by the standard hd and tl functions etc.
  2153 
  2154 A number of basic list functions are no longer exported to the ML
  2155 toplevel, as they are variants of predefined functions.  The following
  2156 suggests how one can translate existing code:
  2157 
  2158     rev_append xs ys = List.revAppend (xs, ys)
  2159     nth_elem (i, xs) = List.nth (xs, i)
  2160     last_elem xs = List.last xs
  2161     flat xss = List.concat xss
  2162     seq fs = List.app fs
  2163     partition P xs = List.partition P xs
  2164     mapfilter f xs = List.mapPartial f xs
  2165 
  2166 * Pure/library.ML: several combinators for linear functional
  2167 transformations, notably reverse application and composition:
  2168 
  2169   x |> f                f #> g
  2170   (x, y) |-> f          f #-> g
  2171 
  2172 * Pure/library.ML: introduced/changed precedence of infix operators:
  2173 
  2174   infix 1 |> |-> ||> ||>> |>> |>>> #> #->;
  2175   infix 2 ?;
  2176   infix 3 o oo ooo oooo;
  2177   infix 4 ~~ upto downto;
  2178 
  2179 Maybe INCOMPATIBILITY when any of those is used in conjunction with other
  2180 infix operators.
  2181 
  2182 * Pure/library.ML: natural list combinators fold, fold_rev, and
  2183 fold_map support linear functional transformations and nesting.  For
  2184 example:
  2185 
  2186   fold f [x1, ..., xN] y =
  2187     y |> f x1 |> ... |> f xN
  2188 
  2189   (fold o fold) f [xs1, ..., xsN] y =
  2190     y |> fold f xs1 |> ... |> fold f xsN
  2191 
  2192   fold f [x1, ..., xN] =
  2193     f x1 #> ... #> f xN
  2194 
  2195   (fold o fold) f [xs1, ..., xsN] =
  2196     fold f xs1 #> ... #> fold f xsN
  2197 
  2198 * Pure/library.ML: the following selectors on type 'a option are
  2199 available:
  2200 
  2201   the:               'a option -> 'a  (*partial*)
  2202   these:             'a option -> 'a  where 'a = 'b list
  2203   the_default: 'a -> 'a option -> 'a
  2204   the_list:          'a option -> 'a list
  2205 
  2206 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  2207 basic operations for association lists, following natural argument
  2208 order; moreover the explicit equality predicate passed here avoids
  2209 potentially expensive polymorphic runtime equality checks.
  2210 The old functions may be expressed as follows:
  2211 
  2212   assoc = uncurry (AList.lookup (op =))
  2213   assocs = these oo AList.lookup (op =)
  2214   overwrite = uncurry (AList.update (op =)) o swap
  2215 
  2216 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  2217 
  2218   val make: ('a -> 'b) -> 'a list -> ('a * 'b) list
  2219   val find: ('a * 'b -> bool) -> ('c * 'b) list -> 'a -> 'c list
  2220 
  2221 replacing make_keylist and keyfilter (occassionally used)
  2222 Naive rewrites:
  2223 
  2224   make_keylist = AList.make
  2225   keyfilter = AList.find (op =)
  2226 
  2227 * eq_fst and eq_snd now take explicit equality parameter, thus
  2228   avoiding eqtypes. Naive rewrites:
  2229 
  2230     eq_fst = eq_fst (op =)
  2231     eq_snd = eq_snd (op =)
  2232 
  2233 * Removed deprecated apl and apr (rarely used).
  2234   Naive rewrites:
  2235 
  2236     apl (n, op) =>>= curry op n
  2237     apr (op, m) =>>= fn n => op (n, m)
  2238 
  2239 * Pure/General: structure OrdList (cf. Pure/General/ord_list.ML)
  2240 provides a reasonably efficient light-weight implementation of sets as
  2241 lists.
  2242 
  2243 * Pure/General: generic tables (cf. Pure/General/table.ML) provide a
  2244 few new operations; existing lookup and update are now curried to
  2245 follow natural argument order (for use with fold etc.);
  2246 INCOMPATIBILITY, use (uncurry Symtab.lookup) etc. as last resort.
  2247 
  2248 * Pure/General: output via the Isabelle channels of
  2249 writeln/warning/error etc. is now passed through Output.output, with a
  2250 hook for arbitrary transformations depending on the print_mode
  2251 (cf. Output.add_mode -- the first active mode that provides a output
  2252 function wins).  Already formatted output may be embedded into further
  2253 text via Output.raw; the result of Pretty.string_of/str_of and derived
  2254 functions (string_of_term/cterm/thm etc.) is already marked raw to
  2255 accommodate easy composition of diagnostic messages etc.  Programmers
  2256 rarely need to care about Output.output or Output.raw at all, with
  2257 some notable exceptions: Output.output is required when bypassing the
  2258 standard channels (writeln etc.), or in token translations to produce
  2259 properly formatted results; Output.raw is required when capturing
  2260 already output material that will eventually be presented to the user
  2261 a second time.  For the default print mode, both Output.output and
  2262 Output.raw have no effect.
  2263 
  2264 * Pure/General: Output.time_accumulator NAME creates an operator ('a
  2265 -> 'b) -> 'a -> 'b to measure runtime and count invocations; the
  2266 cumulative results are displayed at the end of a batch session.
  2267 
  2268 * Pure/General: File.sysify_path and File.quote_sysify path have been
  2269 replaced by File.platform_path and File.shell_path (with appropriate
  2270 hooks).  This provides a clean interface for unusual systems where the
  2271 internal and external process view of file names are different.
  2272 
  2273 * Pure: more efficient orders for basic syntactic entities: added
  2274 fast_string_ord, fast_indexname_ord, fast_term_ord; changed sort_ord
  2275 and typ_ord to use fast_string_ord and fast_indexname_ord (term_ord is
  2276 NOT affected); structures Symtab, Vartab, Typtab, Termtab use the fast
  2277 orders now -- potential INCOMPATIBILITY for code that depends on a
  2278 particular order for Symtab.keys, Symtab.dest, etc. (consider using
  2279 Library.sort_strings on result).
  2280 
  2281 * Pure/term.ML: combinators fold_atyps, fold_aterms, fold_term_types,
  2282 fold_types traverse types/terms from left to right, observing natural
  2283 argument order.  Supercedes previous foldl_XXX versions, add_frees,
  2284 add_vars etc. have been adapted as well: INCOMPATIBILITY.
  2285 
  2286 * Pure: name spaces have been refined, with significant changes of the
  2287 internal interfaces -- INCOMPATIBILITY.  Renamed cond_extern(_table)
  2288 to extern(_table).  The plain name entry path is superceded by a
  2289 general 'naming' context, which also includes the 'policy' to produce
  2290 a fully qualified name and external accesses of a fully qualified
  2291 name; NameSpace.extend is superceded by context dependent
  2292 Sign.declare_name.  Several theory and proof context operations modify
  2293 the naming context.  Especially note Theory.restore_naming and
  2294 ProofContext.restore_naming to get back to a sane state; note that
  2295 Theory.add_path is no longer sufficient to recover from
  2296 Theory.absolute_path in particular.
  2297 
  2298 * Pure: new flags short_names (default false) and unique_names
  2299 (default true) for controlling output of qualified names.  If
  2300 short_names is set, names are printed unqualified.  If unique_names is
  2301 reset, the name prefix is reduced to the minimum required to achieve
  2302 the original result when interning again, even if there is an overlap
  2303 with earlier declarations.
  2304 
  2305 * Pure/TheoryDataFun: change of the argument structure; 'prep_ext' is
  2306 now 'extend', and 'merge' gets an additional Pretty.pp argument
  2307 (useful for printing error messages).  INCOMPATIBILITY.
  2308 
  2309 * Pure: major reorganization of the theory context.  Type Sign.sg and
  2310 Theory.theory are now identified, referring to the universal
  2311 Context.theory (see Pure/context.ML).  Actual signature and theory
  2312 content is managed as theory data.  The old code and interfaces were
  2313 spread over many files and structures; the new arrangement introduces
  2314 considerable INCOMPATIBILITY to gain more clarity:
  2315 
  2316   Context -- theory management operations (name, identity, inclusion,
  2317     parents, ancestors, merge, etc.), plus generic theory data;
  2318 
  2319   Sign -- logical signature and syntax operations (declaring consts,
  2320     types, etc.), plus certify/read for common entities;
  2321 
  2322   Theory -- logical theory operations (stating axioms, definitions,
  2323     oracles), plus a copy of logical signature operations (consts,
  2324     types, etc.); also a few basic management operations (Theory.copy,
  2325     Theory.merge, etc.)
  2326 
  2327 The most basic sign_of operations (Theory.sign_of, Thm.sign_of_thm
  2328 etc.) as well as the sign field in Thm.rep_thm etc. have been retained
  2329 for convenience -- they merely return the theory.
  2330 
  2331 * Pure: type Type.tsig is superceded by theory in most interfaces.
  2332 
  2333 * Pure: the Isar proof context type is already defined early in Pure
  2334 as Context.proof (note that ProofContext.context and Proof.context are
  2335 aliases, where the latter is the preferred name).  This enables other
  2336 Isabelle components to refer to that type even before Isar is present.
  2337 
  2338 * Pure/sign/theory: discontinued named name spaces (i.e. classK,
  2339 typeK, constK, axiomK, oracleK), but provide explicit operations for
  2340 any of these kinds.  For example, Sign.intern typeK is now
  2341 Sign.intern_type, Theory.hide_space Sign.typeK is now
  2342 Theory.hide_types.  Also note that former
  2343 Theory.hide_classes/types/consts are now
  2344 Theory.hide_classes_i/types_i/consts_i, while the non '_i' versions
  2345 internalize their arguments!  INCOMPATIBILITY.
  2346 
  2347 * Pure: get_thm interface (of PureThy and ProofContext) expects
  2348 datatype thmref (with constructors Name and NameSelection) instead of
  2349 plain string -- INCOMPATIBILITY;
  2350 
  2351 * Pure: cases produced by proof methods specify options, where NONE
  2352 means to remove case bindings -- INCOMPATIBILITY in
  2353 (RAW_)METHOD_CASES.
  2354 
  2355 * Pure: the following operations retrieve axioms or theorems from a
  2356 theory node or theory hierarchy, respectively:
  2357 
  2358   Theory.axioms_of: theory -> (string * term) list
  2359   Theory.all_axioms_of: theory -> (string * term) list
  2360   PureThy.thms_of: theory -> (string * thm) list
  2361   PureThy.all_thms_of: theory -> (string * thm) list
  2362 
  2363 * Pure: print_tac now outputs the goal through the trace channel.
  2364 
  2365 * Isar toplevel: improved diagnostics, mostly for Poly/ML only.
  2366 Reference Toplevel.debug (default false) controls detailed printing
  2367 and tracing of low-level exceptions; Toplevel.profiling (default 0)
  2368 controls execution profiling -- set to 1 for time and 2 for space
  2369 (both increase the runtime).
  2370 
  2371 * Isar session: The initial use of ROOT.ML is now always timed,
  2372 i.e. the log will show the actual process times, in contrast to the
  2373 elapsed wall-clock time that the outer shell wrapper produces.
  2374 
  2375 * Simplifier: improved handling of bound variables (nameless
  2376 representation, avoid allocating new strings).  Simprocs that invoke
  2377 the Simplifier recursively should use Simplifier.inherit_bounds to
  2378 avoid local name clashes.  Failure to do so produces warnings
  2379 "Simplifier: renamed bound variable ..."; set Simplifier.debug_bounds
  2380 for further details.
  2381 
  2382 * ML functions legacy_bindings and use_legacy_bindings produce ML fact
  2383 bindings for all theorems stored within a given theory; this may help
  2384 in porting non-Isar theories to Isar ones, while keeping ML proof
  2385 scripts for the time being.
  2386 
  2387 * ML operator HTML.with_charset specifies the charset begin used for
  2388 generated HTML files.  For example:
  2389 
  2390   HTML.with_charset "utf-8" use_thy "Hebrew";
  2391   HTML.with_charset "utf-8" use_thy "Chinese";
  2392 
  2393 
  2394 *** System ***
  2395 
  2396 * Allow symlinks to all proper Isabelle executables (Isabelle,
  2397 isabelle, isatool etc.).
  2398 
  2399 * ISABELLE_DOC_FORMAT setting specifies preferred document format (for
  2400 isatool doc, isatool mkdir, display_drafts etc.).
  2401 
  2402 * isatool usedir: option -f allows specification of the ML file to be
  2403 used by Isabelle; default is ROOT.ML.
  2404 
  2405 * New isatool version outputs the version identifier of the Isabelle
  2406 distribution being used.
  2407 
  2408 * HOL: new isatool dimacs2hol converts files in DIMACS CNF format
  2409 (containing Boolean satisfiability problems) into Isabelle/HOL
  2410 theories.
  2411 
  2412 
  2413 
  2414 New in Isabelle2004 (April 2004)
  2415 --------------------------------
  2416 
  2417 *** General ***
  2418 
  2419 * Provers/order.ML:  new efficient reasoner for partial and linear orders.
  2420   Replaces linorder.ML.
  2421 
  2422 * Pure: Greek letters (except small lambda, \<lambda>), as well as Gothic
  2423   (\<aa>...\<zz>\<AA>...\<ZZ>), calligraphic (\<A>...\<Z>), and Euler
  2424   (\<a>...\<z>), are now considered normal letters, and can therefore
  2425   be used anywhere where an ASCII letter (a...zA...Z) has until
  2426   now. COMPATIBILITY: This obviously changes the parsing of some
  2427   terms, especially where a symbol has been used as a binder, say
  2428   '\<Pi>x. ...', which is now a type error since \<Pi>x will be parsed
  2429   as an identifier.  Fix it by inserting a space around former
  2430   symbols.  Call 'isatool fixgreek' to try to fix parsing errors in
  2431   existing theory and ML files.
  2432 
  2433 * Pure: Macintosh and Windows line-breaks are now allowed in theory files.
  2434 
  2435 * Pure: single letter sub/superscripts (\<^isub> and \<^isup>) are now
  2436   allowed in identifiers. Similar to Greek letters \<^isub> is now considered
  2437   a normal (but invisible) letter. For multiple letter subscripts repeat
  2438   \<^isub> like this: x\<^isub>1\<^isub>2.
  2439 
  2440 * Pure: There are now sub-/superscripts that can span more than one
  2441   character. Text between \<^bsub> and \<^esub> is set in subscript in
  2442   ProofGeneral and LaTeX, text between \<^bsup> and \<^esup> in
  2443   superscript. The new control characters are not identifier parts.
  2444 
  2445 * Pure: Control-symbols of the form \<^raw:...> will literally print the
  2446   content of "..." to the latex file instead of \isacntrl... . The "..."
  2447   may consist of any printable characters excluding the end bracket >.
  2448 
  2449 * Pure: Using new Isar command "finalconsts" (or the ML functions
  2450   Theory.add_finals or Theory.add_finals_i) it is now possible to
  2451   declare constants "final", which prevents their being given a definition
  2452   later.  It is useful for constants whose behaviour is fixed axiomatically
  2453   rather than definitionally, such as the meta-logic connectives.
  2454 
  2455 * Pure: 'instance' now handles general arities with general sorts
  2456   (i.e. intersections of classes),
  2457 
  2458 * Presentation: generated HTML now uses a CSS style sheet to make layout
  2459   (somewhat) independent of content. It is copied from lib/html/isabelle.css.
  2460   It can be changed to alter the colors/layout of generated pages.
  2461 
  2462 
  2463 *** Isar ***
  2464 
  2465 * Tactic emulation methods rule_tac, erule_tac, drule_tac, frule_tac,
  2466   cut_tac, subgoal_tac and thin_tac:
  2467   - Now understand static (Isar) contexts.  As a consequence, users of Isar
  2468     locales are no longer forced to write Isar proof scripts.
  2469     For details see Isar Reference Manual, paragraph 4.3.2: Further tactic
  2470     emulations.
  2471   - INCOMPATIBILITY: names of variables to be instantiated may no
  2472     longer be enclosed in quotes.  Instead, precede variable name with `?'.
  2473     This is consistent with the instantiation attribute "where".
  2474 
  2475 * Attributes "where" and "of":
  2476   - Now take type variables of instantiated theorem into account when reading
  2477     the instantiation string.  This fixes a bug that caused instantiated
  2478     theorems to have too special types in some circumstances.
  2479   - "where" permits explicit instantiations of type variables.
  2480 
  2481 * Calculation commands "moreover" and "also" no longer interfere with
  2482   current facts ("this"), admitting arbitrary combinations with "then"
  2483   and derived forms.
  2484 
  2485 * Locales:
  2486   - Goal statements involving the context element "includes" no longer
  2487     generate theorems with internal delta predicates (those ending on
  2488     "_axioms") in the premise.
  2489     Resolve particular premise with <locale>.intro to obtain old form.
  2490   - Fixed bug in type inference ("unify_frozen") that prevented mix of target
  2491     specification and "includes" elements in goal statement.
  2492   - Rule sets <locale>.intro and <locale>.axioms no longer declared as
  2493     [intro?] and [elim?] (respectively) by default.
  2494   - Experimental command for instantiation of locales in proof contexts:
  2495         instantiate <label>[<attrs>]: <loc>
  2496     Instantiates locale <loc> and adds all its theorems to the current context
  2497     taking into account their attributes.  Label and attrs are optional
  2498     modifiers, like in theorem declarations.  If present, names of
  2499     instantiated theorems are qualified with <label>, and the attributes
  2500     <attrs> are applied after any attributes these theorems might have already.
  2501       If the locale has assumptions, a chained fact of the form
  2502     "<loc> t1 ... tn" is expected from which instantiations of the parameters
  2503     are derived.  The command does not support old-style locales declared
  2504     with "locale (open)".
  2505       A few (very simple) examples can be found in FOL/ex/LocaleInst.thy.
  2506 
  2507 * HOL: Tactic emulation methods induct_tac and case_tac understand static
  2508   (Isar) contexts.
  2509 
  2510 
  2511 *** HOL ***
  2512 
  2513 * Proof import: new image HOL4 contains the imported library from
  2514   the HOL4 system with about 2500 theorems. It is imported by
  2515   replaying proof terms produced by HOL4 in Isabelle. The HOL4 image
  2516   can be used like any other Isabelle image.  See
  2517   HOL/Import/HOL/README for more information.
  2518 
  2519 * Simplifier:
  2520   - Much improved handling of linear and partial orders.
  2521     Reasoners for linear and partial orders are set up for type classes
  2522     "linorder" and "order" respectively, and are added to the default simpset
  2523     as solvers.  This means that the simplifier can build transitivity chains
  2524     to solve goals from the assumptions.
  2525   - INCOMPATIBILITY: old proofs break occasionally.  Typically, applications
  2526     of blast or auto after simplification become unnecessary because the goal
  2527     is solved by simplification already.
  2528 
  2529 * Numerics: new theory Ring_and_Field contains over 250 basic numerical laws,
  2530     all proved in axiomatic type classes for semirings, rings and fields.
  2531 
  2532 * Numerics:
  2533   - Numeric types (nat, int, and in HOL-Complex rat, real, complex, etc.) are
  2534     now formalized using the Ring_and_Field theory mentioned above.
  2535   - INCOMPATIBILITY: simplification and arithmetic behaves somewhat differently
  2536     than before, because now they are set up once in a generic manner.
  2537   - INCOMPATIBILITY: many type-specific arithmetic laws have gone.
  2538     Look for the general versions in Ring_and_Field (and Power if they concern
  2539     exponentiation).
  2540 
  2541 * Type "rat" of the rational numbers is now available in HOL-Complex.
  2542 
  2543 * Records:
  2544   - Record types are now by default printed with their type abbreviation
  2545     instead of the list of all field types. This can be configured via
  2546     the reference "print_record_type_abbr".
  2547   - Simproc "record_upd_simproc" for simplification of multiple updates added
  2548     (not enabled by default).
  2549   - Simproc "record_ex_sel_eq_simproc" to simplify EX x. sel r = x resp.
  2550     EX x. x = sel r to True (not enabled by default).
  2551   - Tactic "record_split_simp_tac" to split and simplify records added.
  2552 
  2553 * 'specification' command added, allowing for definition by
  2554   specification.  There is also an 'ax_specification' command that
  2555   introduces the new constants axiomatically.
  2556 
  2557 * arith(_tac) is now able to generate counterexamples for reals as well.
  2558 
  2559 * HOL-Algebra: new locale "ring" for non-commutative rings.
  2560 
  2561 * HOL-ex: InductiveInvariant_examples illustrates advanced recursive function
  2562   definitions, thanks to Sava Krsti\'{c} and John Matthews.
  2563 
  2564 * HOL-Matrix: a first theory for matrices in HOL with an application of
  2565   matrix theory to linear programming.
  2566 
  2567 * Unions and Intersections:
  2568   The latex output syntax of UN and INT has been changed
  2569   from "\Union x \in A. B" to "\Union_{x \in A} B"
  2570   i.e. the index formulae has become a subscript.
  2571   Similarly for "\Union x. B", and for \Inter instead of \Union.
  2572 
  2573 * Unions and Intersections over Intervals:
  2574   There is new short syntax "UN i<=n. A" for "UN i:{0..n}. A". There is
  2575   also an x-symbol version with subscripts "\<Union>\<^bsub>i <= n\<^esub>. A"
  2576   like in normal math, and corresponding versions for < and for intersection.
  2577 
  2578 * HOL/List: Ordering "lexico" is renamed "lenlex" and the standard
  2579   lexicographic dictonary ordering has been added as "lexord".
  2580 
  2581 * ML: the legacy theory structures Int and List have been removed. They had
  2582   conflicted with ML Basis Library structures having the same names.
  2583 
  2584 * 'refute' command added to search for (finite) countermodels.  Only works
  2585   for a fragment of HOL.  The installation of an external SAT solver is
  2586   highly recommended.  See "HOL/Refute.thy" for details.
  2587 
  2588 * 'quickcheck' command: Allows to find counterexamples by evaluating
  2589   formulae under an assignment of free variables to random values.
  2590   In contrast to 'refute', it can deal with inductive datatypes,
  2591   but cannot handle quantifiers. See "HOL/ex/Quickcheck_Examples.thy"
  2592   for examples.
  2593 
  2594 
  2595 *** HOLCF ***
  2596 
  2597 * Streams now come with concatenation and are part of the HOLCF image
  2598 
  2599 
  2600 
  2601 New in Isabelle2003 (May 2003)
  2602 ------------------------------
  2603 
  2604 *** General ***
  2605 
  2606 * Provers/simplifier:
  2607 
  2608   - Completely reimplemented method simp (ML: Asm_full_simp_tac):
  2609     Assumptions are now subject to complete mutual simplification,
  2610     not just from left to right. The simplifier now preserves
  2611     the order of assumptions.
  2612 
  2613     Potential INCOMPATIBILITY:
  2614 
  2615     -- simp sometimes diverges where the old version did
  2616        not, e.g. invoking simp on the goal
  2617 
  2618         [| P (f x); y = x; f x = f y |] ==> Q
  2619 
  2620        now gives rise to the infinite reduction sequence
  2621 
  2622         P(f x) --(f x = f y)--> P(f y) --(y = x)--> P(f x) --(f x = f y)--> ...
  2623 
  2624        Using "simp (asm_lr)" (ML: Asm_lr_simp_tac) instead often solves this
  2625        kind of problem.
  2626 
  2627     -- Tactics combining classical reasoner and simplification (such as auto)
  2628        are also affected by this change, because many of them rely on
  2629        simp. They may sometimes diverge as well or yield a different numbers
  2630        of subgoals. Try to use e.g. force, fastsimp, or safe instead of auto
  2631        in case of problems. Sometimes subsequent calls to the classical
  2632        reasoner will fail because a preceeding call to the simplifier too
  2633        eagerly simplified the goal, e.g. deleted redundant premises.
  2634 
  2635   - The simplifier trace now shows the names of the applied rewrite rules
  2636 
  2637   - You can limit the number of recursive invocations of the simplifier
  2638     during conditional rewriting (where the simplifie tries to solve the
  2639     conditions before applying the rewrite rule):
  2640     ML "simp_depth_limit := n"
  2641     where n is an integer. Thus you can force termination where previously
  2642     the simplifier would diverge.
  2643 
  2644   - Accepts free variables as head terms in congruence rules.  Useful in Isar.
  2645 
  2646   - No longer aborts on failed congruence proof.  Instead, the
  2647     congruence is ignored.
  2648 
  2649 * Pure: New generic framework for extracting programs from constructive
  2650   proofs. See HOL/Extraction.thy for an example instantiation, as well
  2651   as HOL/Extraction for some case studies.
  2652 
  2653 * Pure: The main goal of the proof state is no longer shown by default, only
  2654 the subgoals. This behaviour is controlled by a new flag.
  2655    PG menu: Isabelle/Isar -> Settings -> Show Main Goal
  2656 (ML: Proof.show_main_goal).
  2657 
  2658 * Pure: You can find all matching introduction rules for subgoal 1, i.e. all
  2659 rules whose conclusion matches subgoal 1:
  2660       PG menu: Isabelle/Isar -> Show me -> matching rules
  2661 The rules are ordered by how closely they match the subgoal.
  2662 In particular, rules that solve a subgoal outright are displayed first
  2663 (or rather last, the way they are printed).
  2664 (ML: ProofGeneral.print_intros())
  2665 
  2666 * Pure: New flag trace_unify_fail causes unification to print
  2667 diagnostic information (PG: in trace buffer) when it fails. This is
  2668 useful for figuring out why single step proofs like rule, erule or
  2669 assumption failed.
  2670 
  2671 * Pure: Locale specifications now produce predicate definitions
  2672 according to the body of text (covering assumptions modulo local
  2673 definitions); predicate "loc_axioms" covers newly introduced text,
  2674 while "loc" is cumulative wrt. all included locale expressions; the
  2675 latter view is presented only on export into the global theory
  2676 context; potential INCOMPATIBILITY, use "(open)" option to fall back
  2677 on the old view without predicates;
  2678 
  2679 * Pure: predefined locales "var" and "struct" are useful for sharing
  2680 parameters (as in CASL, for example); just specify something like
  2681 ``var x + var y + struct M'' as import;
  2682 
  2683 * Pure: improved thms_containing: proper indexing of facts instead of
  2684 raw theorems; check validity of results wrt. current name space;
  2685 include local facts of proof configuration (also covers active
  2686 locales), cover fixed variables in index; may use "_" in term
  2687 specification; an optional limit for the number of printed facts may
  2688 be given (the default is 40);
  2689 
  2690 * Pure: disallow duplicate fact bindings within new-style theory files
  2691 (batch-mode only);
  2692 
  2693 * Provers: improved induct method: assumptions introduced by case
  2694 "foo" are split into "foo.hyps" (from the rule) and "foo.prems" (from
  2695 the goal statement); "foo" still refers to all facts collectively;
  2696 
  2697 * Provers: the function blast.overloaded has been removed: all constants
  2698 are regarded as potentially overloaded, which improves robustness in exchange
  2699 for slight decrease in efficiency;
  2700 
  2701 * Provers/linorder: New generic prover for transitivity reasoning over
  2702 linear orders.  Note: this prover is not efficient!
  2703 
  2704 * Isar: preview of problems to finish 'show' now produce an error
  2705 rather than just a warning (in interactive mode);
  2706 
  2707 
  2708 *** HOL ***
  2709 
  2710 * arith(_tac)
  2711 
  2712  - Produces a counter example if it cannot prove a goal.
  2713    Note that the counter example may be spurious if the goal is not a formula
  2714    of quantifier-free linear arithmetic.
  2715    In ProofGeneral the counter example appears in the trace buffer.
  2716 
  2717  - Knows about div k and mod k where k is a numeral of type nat or int.
  2718 
  2719  - Calls full Presburger arithmetic (by Amine Chaieb) if quantifier-free
  2720    linear arithmetic fails. This takes account of quantifiers and divisibility.
  2721    Presburger arithmetic can also be called explicitly via presburger(_tac).
  2722 
  2723 * simp's arithmetic capabilities have been enhanced a bit: it now
  2724 takes ~= in premises into account (by performing a case split);
  2725 
  2726 * simp reduces "m*(n div m) + n mod m" to n, even if the two summands
  2727 are distributed over a sum of terms;
  2728 
  2729 * New tactic "trans_tac" and method "trans" instantiate
  2730 Provers/linorder.ML for axclasses "order" and "linorder" (predicates
  2731 "<=", "<" and "=").
  2732 
  2733 * function INCOMPATIBILITIES: Pi-sets have been redefined and moved from main
  2734 HOL to Library/FuncSet; constant "Fun.op o" is now called "Fun.comp";
  2735 
  2736 * 'typedef' command has new option "open" to suppress the set
  2737 definition;
  2738 
  2739 * functions Min and Max on finite sets have been introduced (theory
  2740 Finite_Set);
  2741 
  2742 * attribute [symmetric] now works for relations as well; it turns
  2743 (x,y) : R^-1 into (y,x) : R, and vice versa;
  2744 
  2745 * induct over a !!-quantified statement (say !!x1..xn):
  2746   each "case" automatically performs "fix x1 .. xn" with exactly those names.
  2747 
  2748 * Map: `empty' is no longer a constant but a syntactic abbreviation for
  2749 %x. None. Warning: empty_def now refers to the previously hidden definition
  2750 of the empty set.
  2751 
  2752 * Algebra: formalization of classical algebra.  Intended as base for
  2753 any algebraic development in Isabelle.  Currently covers group theory
  2754 (up to Sylow's theorem) and ring theory (Universal Property of
  2755 Univariate Polynomials).  Contributions welcome;
  2756 
  2757 * GroupTheory: deleted, since its material has been moved to Algebra;
  2758 
  2759 * Complex: new directory of the complex numbers with numeric constants,
  2760 nonstandard complex numbers, and some complex analysis, standard and
  2761 nonstandard (Jacques Fleuriot);
  2762 
  2763 * HOL-Complex: new image for analysis, replacing HOL-Real and HOL-Hyperreal;
  2764 
  2765 * Hyperreal: introduced Gauge integration and hyperreal logarithms (Jacques
  2766 Fleuriot);
  2767 
  2768 * Real/HahnBanach: updated and adapted to locales;
  2769 
  2770 * NumberTheory: added Gauss's law of quadratic reciprocity (by Avigad,
  2771 Gray and Kramer);
  2772 
  2773 * UNITY: added the Meier-Sanders theory of progress sets;
  2774 
  2775 * MicroJava: bytecode verifier and lightweight bytecode verifier
  2776 as abstract algorithms, instantiated to the JVM;
  2777 
  2778 * Bali: Java source language formalization. Type system, operational
  2779 semantics, axiomatic semantics. Supported language features:
  2780 classes, interfaces, objects,virtual methods, static methods,
  2781 static/instance fields, arrays, access modifiers, definite
  2782 assignment, exceptions.
  2783 
  2784 
  2785 *** ZF ***
  2786 
  2787 * ZF/Constructible: consistency proof for AC (Gdel's constructible
  2788 universe, etc.);
  2789 
  2790 * Main ZF: virtually all theories converted to new-style format;
  2791 
  2792 
  2793 *** ML ***
  2794 
  2795 * Pure: Tactic.prove provides sane interface for internal proofs;
  2796 omits the infamous "standard" operation, so this is more appropriate
  2797 than prove_goalw_cterm in many situations (e.g. in simprocs);
  2798 
  2799 * Pure: improved error reporting of simprocs;
  2800 
  2801 * Provers: Simplifier.simproc(_i) provides sane interface for setting
  2802 up simprocs;
  2803 
  2804 
  2805 *** Document preparation ***
  2806 
  2807 * uses \par instead of \\ for line breaks in theory text. This may
  2808 shift some page breaks in large documents. To get the old behaviour
  2809 use \renewcommand{\isanewline}{\mbox{}\\\mbox{}} in root.tex.
  2810 
  2811 * minimized dependencies of isabelle.sty and isabellesym.sty on
  2812 other packages
  2813 
  2814 * \<euro> now needs package babel/greek instead of marvosym (which
  2815 broke \Rightarrow)
  2816 
  2817 * normal size for \<zero>...\<nine> (uses \mathbf instead of
  2818 textcomp package)
  2819 
  2820 
  2821 
  2822 New in Isabelle2002 (March 2002)
  2823 --------------------------------
  2824 
  2825 *** Document preparation ***
  2826 
  2827 * greatly simplified document preparation setup, including more
  2828 graceful interpretation of isatool usedir -i/-d/-D options, and more
  2829 instructive isatool mkdir; users should basically be able to get
  2830 started with "isatool mkdir HOL Test && isatool make"; alternatively,
  2831 users may run a separate document processing stage manually like this:
  2832 "isatool usedir -D output HOL Test && isatool document Test/output";
  2833 
  2834 * theory dependency graph may now be incorporated into documents;
  2835 isatool usedir -g true will produce session_graph.eps/.pdf for use
  2836 with \includegraphics of LaTeX;
  2837 
  2838 * proper spacing of consecutive markup elements, especially text
  2839 blocks after section headings;
  2840 
  2841 * support bold style (for single symbols only), input syntax is like
  2842 this: "\<^bold>\<alpha>" or "\<^bold>A";
  2843 
  2844 * \<bullet> is now output as bold \cdot by default, which looks much
  2845 better in printed text;
  2846 
  2847 * added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
  2848 note that these symbols are currently unavailable in Proof General /
  2849 X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
  2850 
  2851 * isatool latex no longer depends on changed TEXINPUTS, instead
  2852 isatool document copies the Isabelle style files to the target
  2853 location;
  2854 
  2855 
  2856 *** Isar ***
  2857 
  2858 * Pure/Provers: improved proof by cases and induction;
  2859   - 'case' command admits impromptu naming of parameters (such as
  2860     "case (Suc n)");
  2861   - 'induct' method divinates rule instantiation from the inductive
  2862     claim; no longer requires excessive ?P bindings for proper
  2863     instantiation of cases;
  2864   - 'induct' method properly enumerates all possibilities of set/type
  2865     rules; as a consequence facts may be also passed through *type*
  2866     rules without further ado;
  2867   - 'induct' method now derives symbolic cases from the *rulified*
  2868     rule (before it used to rulify cases stemming from the internal
  2869     atomized version); this means that the context of a non-atomic
  2870     statement becomes is included in the hypothesis, avoiding the
  2871     slightly cumbersome show "PROP ?case" form;
  2872   - 'induct' may now use elim-style induction rules without chaining
  2873     facts, using ``missing'' premises from the goal state; this allows
  2874     rules stemming from inductive sets to be applied in unstructured
  2875     scripts, while still benefitting from proper handling of non-atomic
  2876     statements; NB: major inductive premises need to be put first, all
  2877     the rest of the goal is passed through the induction;
  2878   - 'induct' proper support for mutual induction involving non-atomic
  2879     rule statements (uses the new concept of simultaneous goals, see
  2880     below);
  2881   - append all possible rule selections, but only use the first
  2882     success (no backtracking);
  2883   - removed obsolete "(simplified)" and "(stripped)" options of methods;
  2884   - undeclared rule case names default to numbers 1, 2, 3, ...;
  2885   - added 'print_induct_rules' (covered by help item in recent Proof
  2886     General versions);
  2887   - moved induct/cases attributes to Pure, methods to Provers;
  2888   - generic method setup instantiated for FOL and HOL;
  2889 
  2890 * Pure: support multiple simultaneous goal statements, for example
  2891 "have a: A and b: B" (same for 'theorem' etc.); being a pure
  2892 meta-level mechanism, this acts as if several individual goals had
  2893 been stated separately; in particular common proof methods need to be
  2894 repeated in order to cover all claims; note that a single elimination
  2895 step is *not* sufficient to establish the two conjunctions, so this
  2896 fails:
  2897 
  2898   assume "A & B" then have A and B ..   (*".." fails*)
  2899 
  2900 better use "obtain" in situations as above; alternative refer to
  2901 multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
  2902 
  2903 * Pure: proper integration with ``locales''; unlike the original
  2904 version by Florian Kammller, Isar locales package high-level proof
  2905 contexts rather than raw logical ones (e.g. we admit to include
  2906 attributes everywhere); operations on locales include merge and
  2907 rename; support for implicit arguments (``structures''); simultaneous
  2908 type-inference over imports and text; see also HOL/ex/Locales.thy for
  2909 some examples;
  2910 
  2911 * Pure: the following commands have been ``localized'', supporting a
  2912 target locale specification "(in name)": 'lemma', 'theorem',
  2913 'corollary', 'lemmas', 'theorems', 'declare'; the results will be
  2914 stored both within the locale and at the theory level (exported and
  2915 qualified by the locale name);
  2916 
  2917 * Pure: theory goals may now be specified in ``long'' form, with
  2918 ad-hoc contexts consisting of arbitrary locale elements. for example
  2919 ``lemma foo: fixes x assumes "A x" shows "B x"'' (local syntax and
  2920 definitions may be given, too); the result is a meta-level rule with
  2921 the context elements being discharged in the obvious way;
  2922 
  2923 * Pure: new proof command 'using' allows to augment currently used
  2924 facts after a goal statement ('using' is syntactically analogous to
  2925 'apply', but acts on the goal's facts only); this allows chained facts
  2926 to be separated into parts given before and after a claim, as in
  2927 ``from a and b have C using d and e <proof>'';
  2928 
  2929 * Pure: renamed "antecedent" case to "rule_context";
  2930 
  2931 * Pure: new 'judgment' command records explicit information about the
  2932 object-logic embedding (used by several tools internally); no longer
  2933 use hard-wired "Trueprop";
  2934 
  2935 * Pure: added 'corollary' command;
  2936 
  2937 * Pure: fixed 'token_translation' command;
  2938 
  2939 * Pure: removed obsolete 'exported' attribute;
  2940 
  2941 * Pure: dummy pattern "_" in is/let is now automatically lifted over
  2942 bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
  2943 supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
  2944 
  2945 * Pure: method 'atomize' presents local goal premises as object-level
  2946 statements (atomic meta-level propositions); setup controlled via
  2947 rewrite rules declarations of 'atomize' attribute; example
  2948 application: 'induct' method with proper rule statements in improper
  2949 proof *scripts*;
  2950 
  2951 * Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
  2952 now consider the syntactic context of assumptions, giving a better
  2953 chance to get type-inference of the arguments right (this is
  2954 especially important for locales);
  2955 
  2956 * Pure: "sorry" no longer requires quick_and_dirty in interactive
  2957 mode;
  2958 
  2959 * Pure/obtain: the formal conclusion "thesis", being marked as
  2960 ``internal'', may no longer be reference directly in the text;
  2961 potential INCOMPATIBILITY, may need to use "?thesis" in rare
  2962 situations;
  2963 
  2964 * Pure: generic 'sym' attribute which declares a rule both as pure
  2965 'elim?' and for the 'symmetric' operation;
  2966 
  2967 * Pure: marginal comments ``--'' may now occur just anywhere in the
  2968 text; the fixed correlation with particular command syntax has been
  2969 discontinued;
  2970 
  2971 * Pure: new method 'rules' is particularly well-suited for proof
  2972 search in intuitionistic logic; a bit slower than 'blast' or 'fast',
  2973 but often produces more compact proof terms with less detours;
  2974 
  2975 * Pure/Provers/classical: simplified integration with pure rule
  2976 attributes and methods; the classical "intro?/elim?/dest?"
  2977 declarations coincide with the pure ones; the "rule" method no longer
  2978 includes classically swapped intros; "intro" and "elim" methods no
  2979 longer pick rules from the context; also got rid of ML declarations
  2980 AddXIs/AddXEs/AddXDs; all of this has some potential for
  2981 INCOMPATIBILITY;
  2982 
  2983 * Provers/classical: attribute 'swapped' produces classical inversions
  2984 of introduction rules;
  2985 
  2986 * Provers/simplifier: 'simplified' attribute may refer to explicit
  2987 rules instead of full simplifier context; 'iff' attribute handles
  2988 conditional rules;
  2989 
  2990 * HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
  2991 
  2992 * HOL: 'recdef' now fails on unfinished automated proofs, use
  2993 "(permissive)" option to recover old behavior;
  2994 
  2995 * HOL: 'inductive' no longer features separate (collective) attributes
  2996 for 'intros' (was found too confusing);
  2997 
  2998 * HOL: properly declared induction rules less_induct and
  2999 wf_induct_rule;
  3000 
  3001 
  3002 *** HOL ***
  3003 
  3004 * HOL: moved over to sane numeral syntax; the new policy is as
  3005 follows:
  3006 
  3007   - 0 and 1 are polymorphic constants, which are defined on any
  3008   numeric type (nat, int, real etc.);
  3009 
  3010   - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
  3011   binary representation internally;
  3012 
  3013   - type nat has special constructor Suc, and generally prefers Suc 0
  3014   over 1::nat and Suc (Suc 0) over 2::nat;
  3015 
  3016 This change may cause significant problems of INCOMPATIBILITY; here
  3017 are some hints on converting existing sources:
  3018 
  3019   - due to the new "num" token, "-0" and "-1" etc. are now atomic
  3020   entities, so expressions involving "-" (unary or binary minus) need
  3021   to be spaced properly;
  3022 
  3023   - existing occurrences of "1" may need to be constraint "1::nat" or
  3024   even replaced by Suc 0; similar for old "2";
  3025 
  3026   - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
  3027 
  3028   - remove all special provisions on numerals in proofs;
  3029 
  3030 * HOL: simp rules nat_number expand numerals on nat to Suc/0
  3031 representation (depends on bin_arith_simps in the default context);
  3032 
  3033 * HOL: symbolic syntax for x^2 (numeral 2);
  3034 
  3035 * HOL: the class of all HOL types is now called "type" rather than
  3036 "term"; INCOMPATIBILITY, need to adapt references to this type class
  3037 in axclass/classes, instance/arities, and (usually rare) occurrences
  3038 in typings (of consts etc.); internally the class is called
  3039 "HOL.type", ML programs should refer to HOLogic.typeS;
  3040 
  3041 * HOL/record package improvements:
  3042   - new derived operations "fields" to build a partial record section,
  3043     "extend" to promote a fixed record to a record scheme, and
  3044     "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
  3045     declared as simp by default;
  3046   - shared operations ("more", "fields", etc.) now need to be always
  3047     qualified) --- potential INCOMPATIBILITY;
  3048   - removed "make_scheme" operations (use "make" with "extend") --
  3049     INCOMPATIBILITY;
  3050   - removed "more" class (simply use "term") -- INCOMPATIBILITY;
  3051   - provides cases/induct rules for use with corresponding Isar
  3052     methods (for concrete records, record schemes, concrete more
  3053     parts, and schematic more parts -- in that order);
  3054   - internal definitions directly based on a light-weight abstract
  3055     theory of product types over typedef rather than datatype;
  3056 
  3057 * HOL: generic code generator for generating executable ML code from
  3058 specifications; specific support for HOL constructs such as inductive
  3059 datatypes and sets, as well as recursive functions; can be invoked
  3060 via 'generate_code' theory section;
  3061 
  3062 * HOL: canonical cases/induct rules for n-tuples (n = 3..7);
  3063 
  3064 * HOL: consolidated and renamed several theories.  In particular:
  3065         Ord.thy has been absorbed into HOL.thy
  3066         String.thy has been absorbed into List.thy
  3067 
  3068 * HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
  3069 (beware of argument permutation!);
  3070 
  3071 * HOL: linorder_less_split superseded by linorder_cases;
  3072 
  3073 * HOL/List: "nodups" renamed to "distinct";
  3074 
  3075 * HOL: added "The" definite description operator; move Hilbert's "Eps"
  3076 to peripheral theory "Hilbert_Choice"; some INCOMPATIBILITIES:
  3077   - Ex_def has changed, now need to use some_eq_ex
  3078 
  3079 * HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
  3080 in this (rare) case use:
  3081 
  3082   delSWrapper "split_all_tac"
  3083   addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
  3084 
  3085 * HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
  3086 MAY FAIL;
  3087 
  3088 * HOL: introduced f^n = f o ... o f; warning: due to the limits of
  3089 Isabelle's type classes, ^ on functions and relations has too general
  3090 a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
  3091 necessary to attach explicit type constraints;
  3092 
  3093 * HOL/Relation: the prefix name of the infix "O" has been changed from
  3094 "comp" to "rel_comp"; INCOMPATIBILITY: a few theorems have been
  3095 renamed accordingly (eg "compI" -> "rel_compI").
  3096 
  3097 * HOL: syntax translations now work properly with numerals and records
  3098 expressions;
  3099 
  3100 * HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
  3101 of "lam" -- INCOMPATIBILITY;
  3102 
  3103 * HOL: got rid of some global declarations (potential INCOMPATIBILITY
  3104 for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
  3105 renamed "Product_Type.unit";
  3106 
  3107 * HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
  3108 
  3109 * HOL: removed obsolete theorem "optionE" (use "option.exhaust", or
  3110 the "cases" method);
  3111 
  3112 * HOL/GroupTheory: group theory examples including Sylow's theorem (by
  3113 Florian Kammller);
  3114 
  3115 * HOL/IMP: updated and converted to new-style theory format; several
  3116 parts turned into readable document, with proper Isar proof texts and
  3117 some explanations (by Gerwin Klein);
  3118 
  3119 * HOL-Real: added Complex_Numbers (by Gertrud Bauer);
  3120 
  3121 * HOL-Hyperreal is now a logic image;
  3122 
  3123 
  3124 *** HOLCF ***
  3125 
  3126 * Isar: consts/constdefs supports mixfix syntax for continuous
  3127 operations;
  3128 
  3129 * Isar: domain package adapted to new-style theory format, e.g. see
  3130 HOLCF/ex/Dnat.thy;
  3131 
  3132 * theory Lift: proper use of rep_datatype lift instead of ML hacks --
  3133 potential INCOMPATIBILITY; now use plain induct_tac instead of former
  3134 lift.induct_tac, always use UU instead of Undef;
  3135 
  3136 * HOLCF/IMP: updated and converted to new-style theory;
  3137 
  3138 
  3139 *** ZF ***
  3140 
  3141 * Isar: proper integration of logic-specific tools and packages,
  3142 including theory commands '(co)inductive', '(co)datatype',
  3143 'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
  3144 'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
  3145 
  3146 * theory Main no longer includes AC; for the Axiom of Choice, base
  3147 your theory on Main_ZFC;
  3148 
  3149 * the integer library now covers quotients and remainders, with many
  3150 laws relating division to addition, multiplication, etc.;
  3151 
  3152 * ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
  3153 typeless version of the formalism;
  3154 
  3155 * ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory
  3156 format;
  3157 
  3158 * ZF/Induct: new directory for examples of inductive definitions,
  3159 including theory Multiset for multiset orderings; converted to
  3160 new-style theory format;
  3161 
  3162 * ZF: many new theorems about lists, ordinals, etc.;
  3163 
  3164 
  3165 *** General ***
  3166 
  3167 * Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
  3168 variable proof controls level of detail: 0 = no proofs (only oracle
  3169 dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
  3170 also ref manual for further ML interfaces;
  3171 
  3172 * Pure/axclass: removed obsolete ML interface
  3173 goal_subclass/goal_arity;
  3174 
  3175 * Pure/syntax: new token syntax "num" for plain numerals (without "#"
  3176 of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
  3177 separate tokens, so expressions involving minus need to be spaced
  3178 properly;
  3179 
  3180 * Pure/syntax: support non-oriented infixes, using keyword "infix"
  3181 rather than "infixl" or "infixr";
  3182 
  3183 * Pure/syntax: concrete syntax for dummy type variables admits genuine
  3184 sort constraint specifications in type inference; e.g. "x::_::foo"
  3185 ensures that the type of "x" is of sort "foo" (but not necessarily a
  3186 type variable);
  3187 
  3188 * Pure/syntax: print modes "type_brackets" and "no_type_brackets"
  3189 control output of nested => (types); the default behavior is
  3190 "type_brackets";
  3191 
  3192 * Pure/syntax: builtin parse translation for "_constify" turns valued
  3193 tokens into AST constants;
  3194 
  3195 * Pure/syntax: prefer later declarations of translations and print
  3196 translation functions; potential INCOMPATIBILITY: need to reverse
  3197 multiple declarations for same syntax element constant;
  3198 
  3199 * Pure/show_hyps reset by default (in accordance to existing Isar
  3200 practice);
  3201 
  3202 * Provers/classical: renamed addaltern to addafter, addSaltern to
  3203 addSafter;
  3204 
  3205 * Provers/clasimp: ``iff'' declarations now handle conditional rules
  3206 as well;
  3207 
  3208 * system: tested support for MacOS X; should be able to get Isabelle +
  3209 Proof General to work in a plain Terminal after installing Poly/ML
  3210 (e.g. from the Isabelle distribution area) and GNU bash alone
  3211 (e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
  3212 support requires further installations, e.g. from
  3213 http://fink.sourceforge.net/);
  3214 
  3215 * system: support Poly/ML 4.1.1 (able to manage larger heaps);
  3216 
  3217 * system: reduced base memory usage by Poly/ML (approx. 20 MB instead
  3218 of 40 MB), cf. ML_OPTIONS;
  3219 
  3220 * system: Proof General keywords specification is now part of the
  3221 Isabelle distribution (see etc/isar-keywords.el);
  3222 
  3223 * system: support for persistent Proof General sessions (refrain from
  3224 outdating all loaded theories on startup); user may create writable
  3225 logic images like this: ``isabelle -q HOL Test'';
  3226 
  3227 * system: smart selection of Isabelle process versus Isabelle
  3228 interface, accommodates case-insensitive file systems (e.g. HFS+); may
  3229 run both "isabelle" and "Isabelle" even if file names are badly
  3230 damaged (executable inspects the case of the first letter of its own
  3231 name); added separate "isabelle-process" and "isabelle-interface";
  3232 
  3233 * system: refrain from any attempt at filtering input streams; no
  3234 longer support ``8bit'' encoding of old isabelle font, instead proper
  3235 iso-latin characters may now be used; the related isatools
  3236 "symbolinput" and "nonascii" have disappeared as well;
  3237 
  3238 * system: removed old "xterm" interface (the print modes "xterm" and
  3239 "xterm_color" are still available for direct use in a suitable
  3240 terminal);
  3241 
  3242 
  3243 
  3244 New in Isabelle99-2 (February 2001)
  3245 -----------------------------------
  3246 
  3247 *** Overview of INCOMPATIBILITIES ***
  3248 
  3249 * HOL: please note that theories in the Library and elsewhere often use the
  3250 new-style (Isar) format; to refer to their theorems in an ML script you must
  3251 bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
  3252 
  3253 * HOL: inductive package no longer splits induction rule aggressively,
  3254 but only as far as specified by the introductions given; the old
  3255 format may be recovered via ML function complete_split_rule or attribute
  3256 'split_rule (complete)';
  3257 
  3258 * HOL: induct renamed to lfp_induct, lfp_Tarski to lfp_unfold,
  3259 gfp_Tarski to gfp_unfold;
  3260 
  3261 * HOL: contrapos, contrapos2 renamed to contrapos_nn, contrapos_pp;
  3262 
  3263 * HOL: infix "dvd" now has priority 50 rather than 70 (because it is a
  3264 relation); infix "^^" has been renamed "``"; infix "``" has been
  3265 renamed "`"; "univalent" has been renamed "single_valued";
  3266 
  3267 * HOL/Real: "rinv" and "hrinv" replaced by overloaded "inverse"
  3268 operation;
  3269 
  3270 * HOLCF: infix "`" has been renamed "$"; the symbol syntax is \<cdot>;
  3271 
  3272 * Isar: 'obtain' no longer declares "that" fact as simp/intro;
  3273 
  3274 * Isar/HOL: method 'induct' now handles non-atomic goals; as a
  3275 consequence, it is no longer monotonic wrt. the local goal context
  3276 (which is now passed through the inductive cases);
  3277 
  3278 * Document preparation: renamed standard symbols \<ll> to \<lless> and
  3279 \<gg> to \<ggreater>;
  3280 
  3281 
  3282 *** Document preparation ***
  3283 
  3284 * \isabellestyle{NAME} selects version of Isabelle output (currently
  3285 available: are "it" for near math-mode best-style output, "sl" for
  3286 slanted text style, and "tt" for plain type-writer; if no
  3287 \isabellestyle command is given, output is according to slanted
  3288 type-writer);
  3289 
  3290 * support sub/super scripts (for single symbols only), input syntax is
  3291 like this: "A\<^sup>*" or "A\<^sup>\<star>";
  3292 
  3293 * some more standard symbols; see Appendix A of the system manual for
  3294 the complete list of symbols defined in isabellesym.sty;
  3295 
  3296 * improved isabelle style files; more abstract symbol implementation
  3297 (should now use \isamath{...} and \isatext{...} in custom symbol
  3298 definitions);
  3299 
  3300 * antiquotation @{goals} and @{subgoals} for output of *dynamic* goals
  3301 state; Note that presentation of goal states does not conform to
  3302 actual human-readable proof documents.  Please do not include goal
  3303 states into document output unless you really know what you are doing!
  3304 
  3305 * proper indentation of antiquoted output with proportional LaTeX
  3306 fonts;
  3307 
  3308 * no_document ML operator temporarily disables LaTeX document
  3309 generation;
  3310 
  3311 * isatool unsymbolize tunes sources for plain ASCII communication;
  3312 
  3313 
  3314 *** Isar ***
  3315 
  3316 * Pure: Isar now suffers initial goal statements to contain unbound
  3317 schematic variables (this does not conform to actual readable proof
  3318 documents, due to unpredictable outcome and non-compositional proof
  3319 checking); users who know what they are doing may use schematic goals
  3320 for Prolog-style synthesis of proven results;
  3321 
  3322 * Pure: assumption method (an implicit finishing) now handles actual
  3323 rules as well;
  3324 
  3325 * Pure: improved 'obtain' --- moved to Pure, insert "that" into
  3326 initial goal, declare "that" only as Pure intro (only for single
  3327 steps); the "that" rule assumption may now be involved in implicit
  3328 finishing, thus ".." becomes a feasible for trivial obtains;
  3329 
  3330 * Pure: default proof step now includes 'intro_classes'; thus trivial
  3331 instance proofs may be performed by "..";
  3332 
  3333 * Pure: ?thesis / ?this / "..." now work for pure meta-level
  3334 statements as well;
  3335 
  3336 * Pure: more robust selection of calculational rules;
  3337 
  3338 * Pure: the builtin notion of 'finished' goal now includes the ==-refl
  3339 rule (as well as the assumption rule);
  3340 
  3341 * Pure: 'thm_deps' command visualizes dependencies of theorems and
  3342 lemmas, using the graph browser tool;
  3343 
  3344 * Pure: predict failure of "show" in interactive mode;
  3345 
  3346 * Pure: 'thms_containing' now takes actual terms as arguments;
  3347 
  3348 * HOL: improved method 'induct' --- now handles non-atomic goals
  3349 (potential INCOMPATIBILITY); tuned error handling;
  3350 
  3351 * HOL: cases and induct rules now provide explicit hints about the
  3352 number of facts to be consumed (0 for "type" and 1 for "set" rules);
  3353 any remaining facts are inserted into the goal verbatim;
  3354 
  3355 * HOL: local contexts (aka cases) may now contain term bindings as
  3356 well; the 'cases' and 'induct' methods new provide a ?case binding for
  3357 the result to be shown in each case;
  3358 
  3359 * HOL: added 'recdef_tc' command;
  3360 
  3361 * isatool convert assists in eliminating legacy ML scripts;
  3362 
  3363 
  3364 *** HOL ***
  3365 
  3366 * HOL/Library: a collection of generic theories to be used together
  3367 with main HOL; the theory loader path already includes this directory
  3368 by default; the following existing theories have been moved here:
  3369 HOL/Induct/Multiset, HOL/Induct/Acc (as Accessible_Part), HOL/While
  3370 (as While_Combinator), HOL/Lex/Prefix (as List_Prefix);
  3371 
  3372 * HOL/Unix: "Some aspects of Unix file-system security", a typical
  3373 modelling and verification task performed in Isabelle/HOL +
  3374 Isabelle/Isar + Isabelle document preparation (by Markus Wenzel).
  3375 
  3376 * HOL/Algebra: special summation operator SUM no longer exists, it has
  3377 been replaced by setsum; infix 'assoc' now has priority 50 (like
  3378 'dvd'); axiom 'one_not_zero' has been moved from axclass 'ring' to
  3379 'domain', this makes the theory consistent with mathematical
  3380 literature;
  3381 
  3382 * HOL basics: added overloaded operations "inverse" and "divide"
  3383 (infix "/"), syntax for generic "abs" operation, generic summation
  3384 operator \<Sum>;
  3385 
  3386 * HOL/typedef: simplified package, provide more useful rules (see also
  3387 HOL/subset.thy);
  3388 
  3389 * HOL/datatype: induction rule for arbitrarily branching datatypes is
  3390 now expressed as a proper nested rule (old-style tactic scripts may
  3391 require atomize_strip_tac to cope with non-atomic premises);
  3392 
  3393 * HOL: renamed theory "Prod" to "Product_Type", renamed "split" rule
  3394 to "split_conv" (old name still available for compatibility);
  3395 
  3396 * HOL: improved concrete syntax for strings (e.g. allows translation
  3397 rules with string literals);
  3398 
  3399 * HOL-Real-Hyperreal: this extends HOL-Real with the hyperreals
  3400  and Fleuriot's mechanization of analysis, including the transcendental
  3401  functions for the reals;
  3402 
  3403 * HOL/Real, HOL/Hyperreal: improved arithmetic simplification;
  3404 
  3405 
  3406 *** CTT ***
  3407 
  3408 * CTT: x-symbol support for Pi, Sigma, -->, : (membership); note that
  3409 "lam" is displayed as TWO lambda-symbols
  3410 
  3411 * CTT: theory Main now available, containing everything (that is, Bool
  3412 and Arith);
  3413 
  3414 
  3415 *** General ***
  3416 
  3417 * Pure: the Simplifier has been implemented properly as a derived rule
  3418 outside of the actual kernel (at last!); the overall performance
  3419 penalty in practical applications is about 50%, while reliability of
  3420 the Isabelle inference kernel has been greatly improved;
  3421 
  3422 * print modes "brackets" and "no_brackets" control output of nested =>
  3423 (types) and ==> (props); the default behaviour is "brackets";
  3424 
  3425 * Provers: fast_tac (and friends) now handle actual object-logic rules
  3426 as assumptions as well;
  3427 
  3428 * system: support Poly/ML 4.0;
  3429 
  3430 * system: isatool install handles KDE version 1 or 2;
  3431 
  3432 
  3433 
  3434 New in Isabelle99-1 (October 2000)
  3435 ----------------------------------
  3436 
  3437 *** Overview of INCOMPATIBILITIES ***
  3438 
  3439 * HOL: simplification of natural numbers is much changed; to partly
  3440 recover the old behaviour (e.g. to prevent n+n rewriting to #2*n)
  3441 issue the following ML commands:
  3442 
  3443   Delsimprocs Nat_Numeral_Simprocs.cancel_numerals;
  3444   Delsimprocs [Nat_Numeral_Simprocs.combine_numerals];
  3445 
  3446 * HOL: simplification no longer dives into case-expressions; this is
  3447 controlled by "t.weak_case_cong" for each datatype t;
  3448 
  3449 * HOL: nat_less_induct renamed to less_induct;
  3450 
  3451 * HOL: systematic renaming of the SOME (Eps) rules, may use isatool
  3452 fixsome to patch .thy and .ML sources automatically;
  3453 
  3454   select_equality  -> some_equality
  3455   select_eq_Ex     -> some_eq_ex
  3456   selectI2EX       -> someI2_ex
  3457   selectI2         -> someI2
  3458   selectI          -> someI
  3459   select1_equality -> some1_equality
  3460   Eps_sym_eq       -> some_sym_eq_trivial
  3461   Eps_eq           -> some_eq_trivial
  3462 
  3463 * HOL: exhaust_tac on datatypes superceded by new generic case_tac;
  3464 
  3465 * HOL: removed obsolete theorem binding expand_if (refer to split_if
  3466 instead);
  3467 
  3468 * HOL: the recursion equations generated by 'recdef' are now called
  3469 f.simps instead of f.rules;
  3470 
  3471 * HOL: qed_spec_mp now also handles bounded ALL as well;
  3472 
  3473 * HOL: 0 is now overloaded, so the type constraint ":: nat" may
  3474 sometimes be needed;
  3475 
  3476 * HOL: the constant for "f``x" is now "image" rather than "op ``";
  3477 
  3478 * HOL: the constant for "f-``x" is now "vimage" rather than "op -``";
  3479 
  3480 * HOL: the disjoint sum is now "<+>" instead of "Plus"; the cartesian
  3481 product is now "<*>" instead of "Times"; the lexicographic product is
  3482 now "<*lex*>" instead of "**";
  3483 
  3484 * HOL: theory Sexp is now in HOL/Induct examples (it used to be part
  3485 of main HOL, but was unused); better use HOL's datatype package;
  3486 
  3487 * HOL: removed "symbols" syntax for constant "override" of theory Map;
  3488 the old syntax may be recovered as follows:
  3489 
  3490   syntax (symbols)
  3491     override  :: "('a ~=> 'b) => ('a ~=> 'b) => ('a ~=> 'b)"
  3492       (infixl "\\<oplus>" 100)
  3493 
  3494 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3495 
  3496 * HOL/ML: even fewer consts are declared as global (see theories Ord,
  3497 Lfp, Gfp, WF); this only affects ML packages that refer to const names
  3498 internally;
  3499 
  3500 * HOL and ZF: syntax for quotienting wrt an equivalence relation
  3501 changed from A/r to A//r;
  3502 
  3503 * ZF: new treatment of arithmetic (nat & int) may break some old
  3504 proofs;
  3505 
  3506 * Isar: renamed some attributes (RS -> THEN, simplify -> simplified,
  3507 rulify -> rule_format, elimify -> elim_format, ...);
  3508 
  3509 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3510 intro/intro!/intro!! flags to intro!/intro/intro? (in most cases, one
  3511 should have to change intro!! to intro? only); replaced "delrule" by
  3512 "rule del";
  3513 
  3514 * Isar/HOL: renamed "intrs" to "intros" in inductive definitions;
  3515 
  3516 * Provers: strengthened force_tac by using new first_best_tac;
  3517 
  3518 * LaTeX document preparation: several changes of isabelle.sty (see
  3519 lib/texinputs);
  3520 
  3521 
  3522 *** Document preparation ***
  3523 
  3524 * formal comments (text blocks etc.) in new-style theories may now
  3525 contain antiquotations of thm/prop/term/typ/text to be presented
  3526 according to latex print mode; concrete syntax is like this:
  3527 @{term[show_types] "f(x) = a + x"};
  3528 
  3529 * isatool mkdir provides easy setup of Isabelle session directories,
  3530 including proper document sources;
  3531 
  3532 * generated LaTeX sources are now deleted after successful run
  3533 (isatool document -c); may retain a copy somewhere else via -D option
  3534 of isatool usedir;
  3535 
  3536 * isatool usedir -D now lets isatool latex -o sty update the Isabelle
  3537 style files, achieving self-contained LaTeX sources and simplifying
  3538 LaTeX debugging;
  3539 
  3540 * old-style theories now produce (crude) LaTeX output as well;
  3541 
  3542 * browser info session directories are now self-contained (may be put
  3543 on WWW server seperately); improved graphs of nested sessions; removed
  3544 graph for 'all sessions';
  3545 
  3546 * several improvements in isabelle style files; \isabellestyle{it}
  3547 produces fake math mode output; \isamarkupheader is now \section by
  3548 default; see lib/texinputs/isabelle.sty etc.;
  3549 
  3550 
  3551 *** Isar ***
  3552 
  3553 * Isar/Pure: local results and corresponding term bindings are now
  3554 subject to Hindley-Milner polymorphism (similar to ML); this
  3555 accommodates incremental type-inference very nicely;
  3556 
  3557 * Isar/Pure: new derived language element 'obtain' supports
  3558 generalized existence reasoning;
  3559 
  3560 * Isar/Pure: new calculational elements 'moreover' and 'ultimately'
  3561 support accumulation of results, without applying any rules yet;
  3562 useful to collect intermediate results without explicit name
  3563 references, and for use with transitivity rules with more than 2
  3564 premises;
  3565 
  3566 * Isar/Pure: scalable support for case-analysis type proofs: new
  3567 'case' language element refers to local contexts symbolically, as
  3568 produced by certain proof methods; internally, case names are attached
  3569 to theorems as "tags";
  3570 
  3571 * Isar/Pure: theory command 'hide' removes declarations from
  3572 class/type/const name spaces;
  3573 
  3574 * Isar/Pure: theory command 'defs' supports option "(overloaded)" to
  3575 indicate potential overloading;
  3576 
  3577 * Isar/Pure: changed syntax of local blocks from {{ }} to { };
  3578 
  3579 * Isar/Pure: syntax of sorts made 'inner', i.e. have to write
  3580 "{a,b,c}" instead of {a,b,c};
  3581 
  3582 * Isar/Pure now provides its own version of intro/elim/dest
  3583 attributes; useful for building new logics, but beware of confusion
  3584 with the version in Provers/classical;
  3585 
  3586 * Isar/Pure: the local context of (non-atomic) goals is provided via
  3587 case name 'antecedent';
  3588 
  3589 * Isar/Pure: removed obsolete 'transfer' attribute (transfer of thms
  3590 to the current context is now done automatically);
  3591 
  3592 * Isar/Pure: theory command 'method_setup' provides a simple interface
  3593 for definining proof methods in ML;
  3594 
  3595 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3596 intro/intro!/intro!! flags to intro!/intro/intro? (INCOMPATIBILITY, in
  3597 most cases, one should have to change intro!! to intro? only);
  3598 replaced "delrule" by "rule del";
  3599 
  3600 * Isar/Provers: new 'hypsubst' method, plain 'subst' method and
  3601 'symmetric' attribute (the latter supercedes [RS sym]);
  3602 
  3603 * Isar/Provers: splitter support (via 'split' attribute and 'simp'
  3604 method modifier); 'simp' method: 'only:' modifier removes loopers as
  3605 well (including splits);
  3606 
  3607 * Isar/Provers: Simplifier and Classical methods now support all kind
  3608 of modifiers used in the past, including 'cong', 'iff', etc.
  3609 
  3610 * Isar/Provers: added 'fastsimp' and 'clarsimp' methods (combination
  3611 of Simplifier and Classical reasoner);
  3612 
  3613 * Isar/HOL: new proof method 'cases' and improved version of 'induct'
  3614 now support named cases; major packages (inductive, datatype, primrec,
  3615 recdef) support case names and properly name parameters;
  3616 
  3617 * Isar/HOL: new transitivity rules for substitution in inequalities --
  3618 monotonicity conditions are extracted to be proven at end of
  3619 calculations;
  3620 
  3621 * Isar/HOL: removed 'case_split' thm binding, should use 'cases' proof
  3622 method anyway;
  3623 
  3624 * Isar/HOL: removed old expand_if = split_if; theorems if_splits =
  3625 split_if split_if_asm; datatype package provides theorems foo.splits =
  3626 foo.split foo.split_asm for each datatype;
  3627 
  3628 * Isar/HOL: tuned inductive package, rename "intrs" to "intros"
  3629 (potential INCOMPATIBILITY), emulation of mk_cases feature for proof
  3630 scripts: new 'inductive_cases' command and 'ind_cases' method; (Note:
  3631 use "(cases (simplified))" method in proper proof texts);
  3632 
  3633 * Isar/HOL: added global 'arith_split' attribute for 'arith' method;
  3634 
  3635 * Isar: names of theorems etc. may be natural numbers as well;
  3636 
  3637 * Isar: 'pr' command: optional arguments for goals_limit and
  3638 ProofContext.prems_limit; no longer prints theory contexts, but only
  3639 proof states;
  3640 
  3641 * Isar: diagnostic commands 'pr', 'thm', 'prop', 'term', 'typ' admit
  3642 additional print modes to be specified; e.g. "pr(latex)" will print
  3643 proof state according to the Isabelle LaTeX style;
  3644 
  3645 * Isar: improved support for emulating tactic scripts, including proof
  3646 methods 'rule_tac' etc., 'cut_tac', 'thin_tac', 'subgoal_tac',
  3647 'rename_tac', 'rotate_tac', 'tactic', and 'case_tac' / 'induct_tac'
  3648 (for HOL datatypes);
  3649 
  3650 * Isar: simplified (more robust) goal selection of proof methods: 1st
  3651 goal, all goals, or explicit goal specifier (tactic emulation); thus
  3652 'proof method scripts' have to be in depth-first order;
  3653 
  3654 * Isar: tuned 'let' syntax: replaced 'as' keyword by 'and';
  3655 
  3656 * Isar: removed 'help' command, which hasn't been too helpful anyway;
  3657 should instead use individual commands for printing items
  3658 (print_commands, print_methods etc.);
  3659 
  3660 * Isar: added 'nothing' --- the empty list of theorems;
  3661 
  3662 
  3663 *** HOL ***
  3664 
  3665 * HOL/MicroJava: formalization of a fragment of Java, together with a
  3666 corresponding virtual machine and a specification of its bytecode
  3667 verifier and a lightweight bytecode verifier, including proofs of
  3668 type-safety; by Gerwin Klein, Tobias Nipkow, David von Oheimb, and
  3669 Cornelia Pusch (see also the homepage of project Bali at
  3670 http://isabelle.in.tum.de/Bali/);
  3671 
  3672 * HOL/Algebra: new theory of rings and univariate polynomials, by
  3673 Clemens Ballarin;
  3674 
  3675 * HOL/NumberTheory: fundamental Theorem of Arithmetic, Chinese
  3676 Remainder Theorem, Fermat/Euler Theorem, Wilson's Theorem, by Thomas M
  3677 Rasmussen;
  3678 
  3679 * HOL/Lattice: fundamental concepts of lattice theory and order
  3680 structures, including duals, properties of bounds versus algebraic
  3681 laws, lattice operations versus set-theoretic ones, the Knaster-Tarski
  3682 Theorem for complete lattices etc.; may also serve as a demonstration
  3683 for abstract algebraic reasoning using axiomatic type classes, and
  3684 mathematics-style proof in Isabelle/Isar; by Markus Wenzel;
  3685 
  3686 * HOL/Prolog: a (bare-bones) implementation of Lambda-Prolog, by David
  3687 von Oheimb;
  3688 
  3689 * HOL/IMPP: extension of IMP with local variables and mutually
  3690 recursive procedures, by David von Oheimb;
  3691 
  3692 * HOL/Lambda: converted into new-style theory and document;
  3693 
  3694 * HOL/ex/Multiquote: example of multiple nested quotations and
  3695 anti-quotations -- basically a generalized version of de-Bruijn
  3696 representation; very useful in avoiding lifting of operations;
  3697 
  3698 * HOL/record: added general record equality rule to simpset; fixed
  3699 select-update simplification procedure to handle extended records as
  3700 well; admit "r" as field name;
  3701 
  3702 * HOL: 0 is now overloaded over the new sort "zero", allowing its use with
  3703 other numeric types and also as the identity of groups, rings, etc.;
  3704 
  3705 * HOL: new axclass plus_ac0 for addition with the AC-laws and 0 as identity.
  3706 Types nat and int belong to this axclass;
  3707 
  3708 * HOL: greatly improved simplification involving numerals of type nat, int, real:
  3709    (i + #8 + j) = Suc k simplifies to  #7 + (i + j) = k
  3710    i*j + k + j*#3*i     simplifies to  #4*(i*j) + k
  3711   two terms #m*u and #n*u are replaced by #(m+n)*u
  3712     (where #m, #n and u can implicitly be 1; this is simproc combine_numerals)
  3713   and the term/formula #m*u+x ~~ #n*u+y simplifies simplifies to #(m-n)+x ~~ y
  3714     or x ~~ #(n-m)+y, where ~~ is one of = < <= or - (simproc cancel_numerals);
  3715 
  3716 * HOL: meson_tac is available (previously in ex/meson.ML); it is a
  3717 powerful prover for predicate logic but knows nothing of clasets; see
  3718 ex/mesontest.ML and ex/mesontest2.ML for example applications;
  3719 
  3720 * HOL: new version of "case_tac" subsumes both boolean case split and
  3721 "exhaust_tac" on datatypes; INCOMPATIBILITY: exhaust_tac no longer
  3722 exists, may define val exhaust_tac = case_tac for ad-hoc portability;
  3723 
  3724 * HOL: simplification no longer dives into case-expressions: only the
  3725 selector expression is simplified, but not the remaining arms; to
  3726 enable full simplification of case-expressions for datatype t, you may
  3727 remove t.weak_case_cong from the simpset, either globally (Delcongs
  3728 [thm"t.weak_case_cong"];) or locally (delcongs [...]).
  3729 
  3730 * HOL/recdef: the recursion equations generated by 'recdef' for
  3731 function 'f' are now called f.simps instead of f.rules; if all
  3732 termination conditions are proved automatically, these simplification
  3733 rules are added to the simpset, as in primrec; rules may be named
  3734 individually as well, resulting in a separate list of theorems for
  3735 each equation;
  3736 
  3737 * HOL/While is a new theory that provides a while-combinator. It
  3738 permits the definition of tail-recursive functions without the
  3739 provision of a termination measure. The latter is necessary once the
  3740 invariant proof rule for while is applied.
  3741 
  3742 * HOL: new (overloaded) notation for the set of elements below/above
  3743 some element: {..u}, {..u(}, {l..}, {)l..}. See theory SetInterval.
  3744 
  3745 * HOL: theorems impI, allI, ballI bound as "strip";
  3746 
  3747 * HOL: new tactic induct_thm_tac: thm -> string -> int -> tactic
  3748 induct_tac th "x1 ... xn" expects th to have a conclusion of the form
  3749 P v1 ... vn and abbreviates res_inst_tac [("v1","x1"),...,("vn","xn")] th;
  3750 
  3751 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3752 
  3753 * HOL: theory Sexp now in HOL/Induct examples (it used to be part of
  3754 main HOL, but was unused);
  3755 
  3756 * HOL: fewer consts declared as global (e.g. have to refer to
  3757 "Lfp.lfp" instead of "lfp" internally; affects ML packages only);
  3758 
  3759 * HOL: tuned AST representation of nested pairs, avoiding bogus output
  3760 in case of overlap with user translations (e.g. judgements over
  3761 tuples); (note that the underlying logical represenation is still
  3762 bogus);
  3763 
  3764 
  3765 *** ZF ***
  3766 
  3767 * ZF: simplification automatically cancels common terms in arithmetic
  3768 expressions over nat and int;
  3769 
  3770 * ZF: new treatment of nat to minimize type-checking: all operators
  3771 coerce their operands to a natural number using the function natify,
  3772 making the algebraic laws unconditional;
  3773 
  3774 * ZF: as above, for int: operators coerce their operands to an integer
  3775 using the function intify;
  3776 
  3777 * ZF: the integer library now contains many of the usual laws for the
  3778 orderings, including $<=, and monotonicity laws for $+ and $*;
  3779 
  3780 * ZF: new example ZF/ex/NatSum to demonstrate integer arithmetic
  3781 simplification;
  3782 
  3783 * FOL and ZF: AddIffs now available, giving theorems of the form P<->Q
  3784 to the simplifier and classical reasoner simultaneously;
  3785 
  3786 
  3787 *** General ***
  3788 
  3789 * Provers: blast_tac now handles actual object-logic rules as
  3790 assumptions; note that auto_tac uses blast_tac internally as well;
  3791 
  3792 * Provers: new functions rulify/rulify_no_asm: thm -> thm for turning
  3793 outer -->/All/Ball into ==>/!!; qed_spec_mp now uses rulify_no_asm;
  3794 
  3795 * Provers: delrules now handles destruct rules as well (no longer need
  3796 explicit make_elim);
  3797 
  3798 * Provers: Blast_tac now warns of and ignores "weak elimination rules" e.g.
  3799   [| inj ?f;          ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3800 use instead the strong form,
  3801   [| inj ?f; ~ ?W ==> ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3802 in HOL, FOL and ZF the function cla_make_elim will create such rules
  3803 from destruct-rules;
  3804 
  3805 * Provers: Simplifier.easy_setup provides a fast path to basic
  3806 Simplifier setup for new object-logics;
  3807 
  3808 * Pure: AST translation rules no longer require constant head on LHS;
  3809 
  3810 * Pure: improved name spaces: ambiguous output is qualified; support
  3811 for hiding of names;
  3812 
  3813 * system: smart setup of canonical ML_HOME, ISABELLE_INTERFACE, and
  3814 XSYMBOL_HOME; no longer need to do manual configuration in most
  3815 situations;
  3816 
  3817 * system: compression of ML heaps images may now be controlled via -c
  3818 option of isabelle and isatool usedir (currently only observed by
  3819 Poly/ML);
  3820 
  3821 * system: isatool installfonts may handle X-Symbol fonts as well (very
  3822 useful for remote X11);
  3823 
  3824 * system: provide TAGS file for Isabelle sources;
  3825 
  3826 * ML: infix 'OF' is a version of 'MRS' with more appropriate argument
  3827 order;
  3828 
  3829 * ML: renamed flags Syntax.trace_norm_ast to Syntax.trace_ast; global
  3830 timing flag supersedes proof_timing and Toplevel.trace;
  3831 
  3832 * ML: new combinators |>> and |>>> for incremental transformations
  3833 with secondary results (e.g. certain theory extensions):
  3834 
  3835 * ML: PureThy.add_defs gets additional argument to indicate potential
  3836 overloading (usually false);
  3837 
  3838 * ML: PureThy.add_thms/add_axioms/add_defs now return theorems as
  3839 results;
  3840 
  3841 
  3842 
  3843 New in Isabelle99 (October 1999)
  3844 --------------------------------
  3845 
  3846 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3847 
  3848 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)
  3849 are no longer simplified.  (This allows the simplifier to unfold recursive
  3850 functional programs.)  To restore the old behaviour, declare
  3851 
  3852     Delcongs [if_weak_cong];
  3853 
  3854 * HOL: Removed the obsolete syntax "Compl A"; use -A for set
  3855 complement;
  3856 
  3857 * HOL: the predicate "inj" is now defined by translation to "inj_on";
  3858 
  3859 * HOL/datatype: mutual_induct_tac no longer exists --
  3860   use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]
  3861 
  3862 * HOL/typedef: fixed type inference for representing set; type
  3863 arguments now have to occur explicitly on the rhs as type constraints;
  3864 
  3865 * ZF: The con_defs part of an inductive definition may no longer refer
  3866 to constants declared in the same theory;
  3867 
  3868 * HOL, ZF: the function mk_cases, generated by the inductive
  3869 definition package, has lost an argument.  To simplify its result, it
  3870 uses the default simpset instead of a supplied list of theorems.
  3871 
  3872 * HOL/List: the constructors of type list are now Nil and Cons;
  3873 
  3874 * Simplifier: the type of the infix ML functions
  3875         setSSolver addSSolver setSolver addSolver
  3876 is now  simpset * solver -> simpset  where `solver' is a new abstract type
  3877 for packaging solvers. A solver is created via
  3878         mk_solver: string -> (thm list -> int -> tactic) -> solver
  3879 where the string argument is only a comment.
  3880 
  3881 
  3882 *** Proof tools ***
  3883 
  3884 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a
  3885 decision procedure for linear arithmetic. Currently it is used for
  3886 types `nat', `int', and `real' in HOL (see below); it can, should and
  3887 will be instantiated for other types and logics as well.
  3888 
  3889 * The simplifier now accepts rewrite rules with flexible heads, eg
  3890      hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y
  3891   They are applied like any rule with a non-pattern lhs, i.e. by first-order
  3892   matching.
  3893 
  3894 
  3895 *** General ***
  3896 
  3897 * New Isabelle/Isar subsystem provides an alternative to traditional
  3898 tactical theorem proving; together with the ProofGeneral/isar user
  3899 interface it offers an interactive environment for developing human
  3900 readable proof documents (Isar == Intelligible semi-automated
  3901 reasoning); for further information see isatool doc isar-ref,
  3902 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/
  3903 
  3904 * improved and simplified presentation of theories: better HTML markup
  3905 (including colors), graph views in several sizes; isatool usedir now
  3906 provides a proper interface for user theories (via -P option); actual
  3907 document preparation based on (PDF)LaTeX is available as well (for
  3908 new-style theories only); see isatool doc system for more information;
  3909 
  3910 * native support for Proof General, both for classic Isabelle and
  3911 Isabelle/Isar;
  3912 
  3913 * ML function thm_deps visualizes dependencies of theorems and lemmas,
  3914 using the graph browser tool;
  3915 
  3916 * Isabelle manuals now also available as PDF;
  3917 
  3918 * theory loader rewritten from scratch (may not be fully
  3919 bug-compatible); old loadpath variable has been replaced by show_path,
  3920 add_path, del_path, reset_path functions; new operations such as
  3921 update_thy, touch_thy, remove_thy, use/update_thy_only (see also
  3922 isatool doc ref);
  3923 
  3924 * improved isatool install: option -k creates KDE application icon,
  3925 option -p DIR installs standalone binaries;
  3926 
  3927 * added ML_PLATFORM setting (useful for cross-platform installations);
  3928 more robust handling of platform specific ML images for SML/NJ;
  3929 
  3930 * the settings environment is now statically scoped, i.e. it is never
  3931 created again in sub-processes invoked from isabelle, isatool, or
  3932 Isabelle;
  3933 
  3934 * path element specification '~~' refers to '$ISABELLE_HOME';
  3935 
  3936 * in locales, the "assumes" and "defines" parts may be omitted if
  3937 empty;
  3938 
  3939 * new print_mode "xsymbols" for extended symbol support (e.g. genuine
  3940 long arrows);
  3941 
  3942 * new print_mode "HTML";
  3943 
  3944 * new flag show_tags controls display of tags of theorems (which are
  3945 basically just comments that may be attached by some tools);
  3946 
  3947 * Isamode 2.6 requires patch to accomodate change of Isabelle font
  3948 mode and goal output format:
  3949 
  3950 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el
  3951 244c244
  3952 <       (list (isa-getenv "ISABELLE") "-msymbols" logic-name)
  3953 ---
  3954 >       (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)
  3955 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el
  3956 181c181
  3957 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"
  3958 ---
  3959 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"
  3960 
  3961 * function bind_thms stores lists of theorems (cf. bind_thm);
  3962 
  3963 * new shorthand tactics ftac, eatac, datac, fatac;
  3964 
  3965 * qed (and friends) now accept "" as result name; in that case the
  3966 theorem is not stored, but proper checks and presentation of the
  3967 result still apply;
  3968 
  3969 * theorem database now also indexes constants "Trueprop", "all",
  3970 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;
  3971 
  3972 
  3973 *** HOL ***
  3974 
  3975 ** HOL arithmetic **
  3976 
  3977 * There are now decision procedures for linear arithmetic over nat and
  3978 int:
  3979 
  3980 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',
  3981 `+', `-', `Suc', `min', `max' and numerical constants; other subterms
  3982 are treated as atomic; subformulae not involving type `nat' or `int'
  3983 are ignored; quantified subformulae are ignored unless they are
  3984 positive universal or negative existential. The tactic has to be
  3985 invoked by hand and can be a little bit slow. In particular, the
  3986 running time is exponential in the number of occurrences of `min' and
  3987 `max', and `-' on `nat'.
  3988 
  3989 2. fast_arith_tac is a cut-down version of arith_tac: it only takes
  3990 (negated) (in)equalities among the premises and the conclusion into
  3991 account (i.e. no compound formulae) and does not know about `min' and
  3992 `max', and `-' on `nat'. It is fast and is used automatically by the
  3993 simplifier.
  3994 
  3995 NB: At the moment, these decision procedures do not cope with mixed
  3996 nat/int formulae where the two parts interact, such as `m < n ==>
  3997 int(m) < int(n)'.
  3998 
  3999 * HOL/Numeral provides a generic theory of numerals (encoded
  4000 efficiently as bit strings); setup for types nat/int/real is in place;
  4001 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than
  4002 int, existing theories and proof scripts may require a few additional
  4003 type constraints;
  4004 
  4005 * integer division and remainder can now be performed on constant
  4006 arguments;
  4007 
  4008 * many properties of integer multiplication, division and remainder
  4009 are now available;
  4010 
  4011 * An interface to the Stanford Validity Checker (SVC) is available through the
  4012 tactic svc_tac.  Propositional tautologies and theorems of linear arithmetic
  4013 are proved automatically.  SVC must be installed separately, and its results
  4014 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any
  4015 invocation of the underlying oracle).  For SVC see
  4016   http://verify.stanford.edu/SVC
  4017 
  4018 * IsaMakefile: the HOL-Real target now builds an actual image;
  4019 
  4020 
  4021 ** HOL misc **
  4022 
  4023 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces
  4024 (in Isabelle/Isar) -- by Gertrud Bauer;
  4025 
  4026 * HOL/BCV: generic model of bytecode verification, i.e. data-flow
  4027 analysis for assembly languages with subtypes;
  4028 
  4029 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization
  4030 -- avoids syntactic ambiguities and treats state, transition, and
  4031 temporal levels more uniformly; introduces INCOMPATIBILITIES due to
  4032 changed syntax and (many) tactics;
  4033 
  4034 * HOL/inductive: Now also handles more general introduction rules such
  4035   as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity
  4036   theorems are now maintained within the theory (maintained via the
  4037   "mono" attribute);
  4038 
  4039 * HOL/datatype: Now also handles arbitrarily branching datatypes
  4040   (using function types) such as
  4041 
  4042   datatype 'a tree = Atom 'a | Branch "nat => 'a tree"
  4043 
  4044 * HOL/record: record_simproc (part of the default simpset) takes care
  4045 of selectors applied to updated records; record_split_tac is no longer
  4046 part of the default claset; update_defs may now be removed from the
  4047 simpset in many cases; COMPATIBILITY: old behavior achieved by
  4048 
  4049   claset_ref () := claset() addSWrapper record_split_wrapper;
  4050   Delsimprocs [record_simproc]
  4051 
  4052 * HOL/typedef: fixed type inference for representing set; type
  4053 arguments now have to occur explicitly on the rhs as type constraints;
  4054 
  4055 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem
  4056 names rather than an ML expression;
  4057 
  4058 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be
  4059 supplied later.  Program schemes can be defined, such as
  4060     "While B C s = (if B s then While B C (C s) else s)"
  4061 where the well-founded relation can be chosen after B and C have been given.
  4062 
  4063 * HOL/List: the constructors of type list are now Nil and Cons;
  4064 INCOMPATIBILITY: while [] and infix # syntax is still there, of
  4065 course, ML tools referring to List.list.op # etc. have to be adapted;
  4066 
  4067 * HOL_quantifiers flag superseded by "HOL" print mode, which is
  4068 disabled by default; run isabelle with option -m HOL to get back to
  4069 the original Gordon/HOL-style output;
  4070 
  4071 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,
  4072 ALL x<=y. P, EX x<y. P, EX x<=y. P;
  4073 
  4074 * HOL basic syntax simplified (more orthogonal): all variants of
  4075 All/Ex now support plain / symbolic / HOL notation; plain syntax for
  4076 Eps operator is provided as well: "SOME x. P[x]";
  4077 
  4078 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;
  4079 
  4080 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made
  4081 thus available for user theories;
  4082 
  4083 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with
  4084 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the
  4085 time;
  4086 
  4087 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec
  4088 several times and then mp;
  4089 
  4090 
  4091 *** LK ***
  4092 
  4093 * the notation <<...>> is now available as a notation for sequences of
  4094 formulas;
  4095 
  4096 * the simplifier is now installed
  4097 
  4098 * the axiom system has been generalized (thanks to Soren Heilmann)
  4099 
  4100 * the classical reasoner now has a default rule database
  4101 
  4102 
  4103 *** ZF ***
  4104 
  4105 * new primrec section allows primitive recursive functions to be given
  4106 directly (as in HOL) over datatypes and the natural numbers;
  4107 
  4108 * new tactics induct_tac and exhaust_tac for induction (or case
  4109 analysis) over datatypes and the natural numbers;
  4110 
  4111 * the datatype declaration of type T now defines the recursor T_rec;
  4112 
  4113 * simplification automatically does freeness reasoning for datatype
  4114 constructors;
  4115 
  4116 * automatic type-inference, with AddTCs command to insert new
  4117 type-checking rules;
  4118 
  4119 * datatype introduction rules are now added as Safe Introduction rules
  4120 to the claset;
  4121 
  4122 * the syntax "if P then x else y" is now available in addition to
  4123 if(P,x,y);
  4124 
  4125 
  4126 *** Internal programming interfaces ***
  4127 
  4128 * tuned simplifier trace output; new flag debug_simp;
  4129 
  4130 * structures Vartab / Termtab (instances of TableFun) offer efficient
  4131 tables indexed by indexname_ord / term_ord (compatible with aconv);
  4132 
  4133 * AxClass.axclass_tac lost the theory argument;
  4134 
  4135 * tuned current_goals_markers semantics: begin / end goal avoids
  4136 printing empty lines;
  4137 
  4138 * removed prs and prs_fn hook, which was broken because it did not
  4139 include \n in its semantics, forcing writeln to add one
  4140 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:
  4141 string -> unit if you really want to output text without newline;
  4142 
  4143 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to
  4144 plain output, interface builders may have to enable 'isabelle_font'
  4145 mode to get Isabelle font glyphs as before;
  4146 
  4147 * refined token_translation interface; INCOMPATIBILITY: output length
  4148 now of type real instead of int;
  4149 
  4150 * theory loader actions may be traced via new ThyInfo.add_hook
  4151 interface (see src/Pure/Thy/thy_info.ML); example application: keep
  4152 your own database of information attached to *whole* theories -- as
  4153 opposed to intra-theory data slots offered via TheoryDataFun;
  4154 
  4155 * proper handling of dangling sort hypotheses (at last!);
  4156 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing
  4157 extra sort hypotheses that can be witnessed from the type signature;
  4158 the force_strip_shyps flag is gone, any remaining shyps are simply
  4159 left in the theorem (with a warning issued by strip_shyps_warning);
  4160 
  4161 
  4162 
  4163 New in Isabelle98-1 (October 1998)
  4164 ----------------------------------
  4165 
  4166 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  4167 
  4168 * several changes of automated proof tools;
  4169 
  4170 * HOL: major changes to the inductive and datatype packages, including
  4171 some minor incompatibilities of theory syntax;
  4172 
  4173 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now
  4174 called `inj_on';
  4175 
  4176 * HOL: removed duplicate thms in Arith:
  4177   less_imp_add_less  should be replaced by  trans_less_add1
  4178   le_imp_add_le      should be replaced by  trans_le_add1
  4179 
  4180 * HOL: unary minus is now overloaded (new type constraints may be
  4181 required);
  4182 
  4183 * HOL and ZF: unary minus for integers is now #- instead of #~.  In
  4184 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is
  4185 now taken as an integer constant.
  4186 
  4187 * Pure: ML function 'theory_of' renamed to 'theory';
  4188 
  4189 
  4190 *** Proof tools ***
  4191 
  4192 * Simplifier:
  4193   1. Asm_full_simp_tac is now more aggressive.
  4194      1. It will sometimes reorient premises if that increases their power to
  4195         simplify.
  4196      2. It does no longer proceed strictly from left to right but may also
  4197         rotate premises to achieve further simplification.
  4198      For compatibility reasons there is now Asm_lr_simp_tac which is like the
  4199      old Asm_full_simp_tac in that it does not rotate premises.
  4200   2. The simplifier now knows a little bit about nat-arithmetic.
  4201 
  4202 * Classical reasoner: wrapper mechanism for the classical reasoner now
  4203 allows for selected deletion of wrappers, by introduction of names for
  4204 wrapper functionals.  This implies that addbefore, addSbefore,
  4205 addaltern, and addSaltern now take a pair (name, tactic) as argument,
  4206 and that adding two tactics with the same name overwrites the first
  4207 one (emitting a warning).
  4208   type wrapper = (int -> tactic) -> (int -> tactic)
  4209   setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by
  4210   addWrapper, addSWrapper: claset * (string * wrapper) -> claset
  4211   delWrapper, delSWrapper: claset *  string            -> claset
  4212   getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;
  4213 
  4214 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE
  4215 semantics; addbefore now affects only the unsafe part of step_tac
  4216 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY
  4217 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac
  4218 by Force_tac;
  4219 
  4220 * Classical reasoner: setwrapper to setWrapper and compwrapper to
  4221 compWrapper; added safe wrapper (and access functions for it);
  4222 
  4223 * HOL/split_all_tac is now much faster and fails if there is nothing
  4224 to split.  Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order
  4225 and the names of the automatically generated variables have changed.
  4226 split_all_tac has moved within claset() from unsafe wrappers to safe
  4227 wrappers, which means that !!-bound variables are split much more
  4228 aggressively, and safe_tac and clarify_tac now split such variables.
  4229 If this splitting is not appropriate, use delSWrapper "split_all_tac".
  4230 Note: the same holds for record_split_tac, which does the job of
  4231 split_all_tac for record fields.
  4232 
  4233 * HOL/Simplifier: Rewrite rules for case distinctions can now be added
  4234 permanently to the default simpset using Addsplits just like
  4235 Addsimps. They can be removed via Delsplits just like
  4236 Delsimps. Lower-case versions are also available.
  4237 
  4238 * HOL/Simplifier: The rule split_if is now part of the default
  4239 simpset. This means that the simplifier will eliminate all occurrences
  4240 of if-then-else in the conclusion of a goal. To prevent this, you can
  4241 either remove split_if completely from the default simpset by
  4242 `Delsplits [split_if]' or remove it in a specific call of the
  4243 simplifier using `... delsplits [split_if]'.  You can also add/delete
  4244 other case splitting rules to/from the default simpset: every datatype
  4245 generates suitable rules `split_t_case' and `split_t_case_asm' (where
  4246 t is the name of the datatype).
  4247 
  4248 * Classical reasoner / Simplifier combination: new force_tac (and
  4249 derivatives Force_tac, force) combines rewriting and classical
  4250 reasoning (and whatever other tools) similarly to auto_tac, but is
  4251 aimed to solve the given subgoal completely.
  4252 
  4253 
  4254 *** General ***
  4255 
  4256 * new top-level commands `Goal' and `Goalw' that improve upon `goal'
  4257 and `goalw': the theory is no longer needed as an explicit argument -
  4258 the current theory context is used; assumptions are no longer returned
  4259 at the ML-level unless one of them starts with ==> or !!; it is
  4260 recommended to convert to these new commands using isatool fixgoal
  4261 (backup your sources first!);
  4262 
  4263 * new top-level commands 'thm' and 'thms' for retrieving theorems from
  4264 the current theory context, and 'theory' to lookup stored theories;
  4265 
  4266 * new theory section 'locale' for declaring constants, assumptions and
  4267 definitions that have local scope;
  4268 
  4269 * new theory section 'nonterminals' for purely syntactic types;
  4270 
  4271 * new theory section 'setup' for generic ML setup functions
  4272 (e.g. package initialization);
  4273 
  4274 * the distribution now includes Isabelle icons: see
  4275 lib/logo/isabelle-{small,tiny}.xpm;
  4276 
  4277 * isatool install - install binaries with absolute references to
  4278 ISABELLE_HOME/bin;
  4279 
  4280 * isatool logo -- create instances of the Isabelle logo (as EPS);
  4281 
  4282 * print mode 'emacs' reserved for Isamode;
  4283 
  4284 * support multiple print (ast) translations per constant name;
  4285 
  4286 * theorems involving oracles are now printed with a suffixed [!];
  4287 
  4288 
  4289 *** HOL ***
  4290 
  4291 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');
  4292 
  4293 * HOL/inductive package reorganized and improved: now supports mutual
  4294 definitions such as
  4295 
  4296   inductive EVEN ODD
  4297     intrs
  4298       null "0 : EVEN"
  4299       oddI "n : EVEN ==> Suc n : ODD"
  4300       evenI "n : ODD ==> Suc n : EVEN"
  4301 
  4302 new theorem list "elims" contains an elimination rule for each of the
  4303 recursive sets; inductive definitions now handle disjunctive premises
  4304 correctly (also ZF);
  4305 
  4306 INCOMPATIBILITIES: requires Inductive as an ancestor; component
  4307 "mutual_induct" no longer exists - the induction rule is always
  4308 contained in "induct";
  4309 
  4310 
  4311 * HOL/datatype package re-implemented and greatly improved: now
  4312 supports mutually recursive datatypes such as
  4313 
  4314   datatype
  4315     'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)
  4316             | SUM ('a aexp) ('a aexp)
  4317             | DIFF ('a aexp) ('a aexp)
  4318             | NUM 'a
  4319   and
  4320     'a bexp = LESS ('a aexp) ('a aexp)
  4321             | AND ('a bexp) ('a bexp)
  4322             | OR ('a bexp) ('a bexp)
  4323 
  4324 as well as indirectly recursive datatypes such as
  4325 
  4326   datatype
  4327     ('a, 'b) term = Var 'a
  4328                   | App 'b ((('a, 'b) term) list)
  4329 
  4330 The new tactic  mutual_induct_tac [<var_1>, ..., <var_n>] i  performs
  4331 induction on mutually / indirectly recursive datatypes.
  4332 
  4333 Primrec equations are now stored in theory and can be accessed via
  4334 <function_name>.simps.
  4335 
  4336 INCOMPATIBILITIES:
  4337 
  4338   - Theories using datatypes must now have theory Datatype as an
  4339     ancestor.
  4340   - The specific <typename>.induct_tac no longer exists - use the
  4341     generic induct_tac instead.
  4342   - natE has been renamed to nat.exhaust - use exhaust_tac
  4343     instead of res_inst_tac ... natE. Note that the variable
  4344     names in nat.exhaust differ from the names in natE, this
  4345     may cause some "fragile" proofs to fail.
  4346   - The theorems split_<typename>_case and split_<typename>_case_asm
  4347     have been renamed to <typename>.split and <typename>.split_asm.
  4348   - Since default sorts of type variables are now handled correctly,
  4349     some datatype definitions may have to be annotated with explicit
  4350     sort constraints.
  4351   - Primrec definitions no longer require function name and type
  4352     of recursive argument.
  4353 
  4354 Consider using isatool fixdatatype to adapt your theories and proof
  4355 scripts to the new package (backup your sources first!).
  4356 
  4357 
  4358 * HOL/record package: considerably improved implementation; now
  4359 includes concrete syntax for record types, terms, updates; theorems
  4360 for surjective pairing and splitting !!-bound record variables; proof
  4361 support is as follows:
  4362 
  4363   1) standard conversions (selectors or updates applied to record
  4364 constructor terms) are part of the standard simpset;
  4365 
  4366   2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are
  4367 made part of standard simpset and claset via addIffs;
  4368 
  4369   3) a tactic for record field splitting (record_split_tac) is part of
  4370 the standard claset (addSWrapper);
  4371 
  4372 To get a better idea about these rules you may retrieve them via
  4373 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is
  4374 the name of your record type.
  4375 
  4376 The split tactic 3) conceptually simplifies by the following rule:
  4377 
  4378   "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"
  4379 
  4380 Thus any record variable that is bound by meta-all will automatically
  4381 blow up into some record constructor term, consequently the
  4382 simplifications of 1), 2) apply.  Thus force_tac, auto_tac etc. shall
  4383 solve record problems automatically.
  4384 
  4385 
  4386 * reorganized the main HOL image: HOL/Integ and String loaded by
  4387 default; theory Main includes everything;
  4388 
  4389 * automatic simplification of integer sums and comparisons, using cancellation;
  4390 
  4391 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;
  4392 
  4393 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;
  4394 
  4395 * many new identities for unions, intersections, set difference, etc.;
  4396 
  4397 * expand_if, expand_split, expand_sum_case and expand_nat_case are now
  4398 called split_if, split_split, split_sum_case and split_nat_case (to go
  4399 with add/delsplits);
  4400 
  4401 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting
  4402 (?x::unit) = (); this is made part of the default simpset, which COULD
  4403 MAKE EXISTING PROOFS FAIL under rare circumstances (consider
  4404 'Delsimprocs [unit_eq_proc];' as last resort); also note that
  4405 unit_abs_eta_conv is added in order to counter the effect of
  4406 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by
  4407 %u.f();
  4408 
  4409 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which
  4410 makes more sense);
  4411 
  4412 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4413   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4414   disjointness reasoning but breaking a few old proofs.
  4415 
  4416 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1
  4417 to 'converse' from 'inverse' (for compatibility with ZF and some
  4418 literature);
  4419 
  4420 * HOL/recdef can now declare non-recursive functions, with {} supplied as
  4421 the well-founded relation;
  4422 
  4423 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of
  4424     Compl A.  The "Compl" syntax remains available as input syntax for this
  4425     release ONLY.
  4426 
  4427 * HOL/Update: new theory of function updates:
  4428     f(a:=b) == %x. if x=a then b else f x
  4429 may also be iterated as in f(a:=b,c:=d,...);
  4430 
  4431 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;
  4432 
  4433 * HOL/List:
  4434   - new function list_update written xs[i:=v] that updates the i-th
  4435     list position. May also be iterated as in xs[i:=a,j:=b,...].
  4436   - new function `upt' written [i..j(] which generates the list
  4437     [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper
  4438     bound write [i..j], which is a shorthand for [i..j+1(].
  4439   - new lexicographic orderings and corresponding wellfoundedness theorems.
  4440 
  4441 * HOL/Arith:
  4442   - removed 'pred' (predecessor) function;
  4443   - generalized some theorems about n-1;
  4444   - many new laws about "div" and "mod";
  4445   - new laws about greatest common divisors (see theory ex/Primes);
  4446 
  4447 * HOL/Relation: renamed the relational operator r^-1 "converse"
  4448 instead of "inverse";
  4449 
  4450 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness
  4451   of the multiset ordering;
  4452 
  4453 * directory HOL/Real: a construction of the reals using Dedekind cuts
  4454   (not included by default);
  4455 
  4456 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;
  4457 
  4458 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted
  4459   programs, i.e. different program variables may have different types.
  4460 
  4461 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4462   [previously, this check worked only if the rewrite rule was unconditional]
  4463   Now rew can involve either definitions or equalities (either == or =).
  4464 
  4465 
  4466 *** ZF ***
  4467 
  4468 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains
  4469   only the theorems proved on ZF.ML;
  4470 
  4471 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4472   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4473   disjointness reasoning but breaking a few old proofs.
  4474 
  4475 * ZF/Update: new theory of function updates
  4476     with default rewrite rule  f(x:=y) ` z = if(z=x, y, f`z)
  4477   may also be iterated as in f(a:=b,c:=d,...);
  4478 
  4479 * in  let x=t in u(x), neither t nor u(x) has to be an FOL term.
  4480 
  4481 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4482   [previously, this check worked only if the rewrite rule was unconditional]
  4483   Now rew can involve either definitions or equalities (either == or =).
  4484 
  4485 * case_tac provided for compatibility with HOL
  4486     (like the old excluded_middle_tac, but with subgoals swapped)
  4487 
  4488 
  4489 *** Internal programming interfaces ***
  4490 
  4491 * Pure: several new basic modules made available for general use, see
  4492 also src/Pure/README;
  4493 
  4494 * improved the theory data mechanism to support encapsulation (data
  4495 kind name replaced by private Object.kind, acting as authorization
  4496 key); new type-safe user interface via functor TheoryDataFun; generic
  4497 print_data function becomes basically useless;
  4498 
  4499 * removed global_names compatibility flag -- all theory declarations
  4500 are qualified by default;
  4501 
  4502 * module Pure/Syntax now offers quote / antiquote translation
  4503 functions (useful for Hoare logic etc. with implicit dependencies);
  4504 see HOL/ex/Antiquote for an example use;
  4505 
  4506 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->
  4507 cterm -> thm;
  4508 
  4509 * new tactical CHANGED_GOAL for checking that a tactic modifies a
  4510 subgoal;
  4511 
  4512 * Display.print_goals function moved to Locale.print_goals;
  4513 
  4514 * standard print function for goals supports current_goals_markers
  4515 variable for marking begin of proof, end of proof, start of goal; the
  4516 default is ("", "", ""); setting current_goals_markers := ("<proof>",
  4517 "</proof>", "<goal>") causes SGML like tagged proof state printing,
  4518 for example;
  4519 
  4520 
  4521 
  4522 New in Isabelle98 (January 1998)
  4523 --------------------------------
  4524 
  4525 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  4526 
  4527 * changed lexical syntax of terms / types: dots made part of long
  4528 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";
  4529 
  4530 * simpset (and claset) reference variable replaced by functions
  4531 simpset / simpset_ref;
  4532 
  4533 * no longer supports theory aliases (via merge) and non-trivial
  4534 implicit merge of thms' signatures;
  4535 
  4536 * most internal names of constants changed due to qualified names;
  4537 
  4538 * changed Pure/Sequence interface (see Pure/seq.ML);
  4539 
  4540 
  4541 *** General Changes ***
  4542 
  4543 * hierachically structured name spaces (for consts, types, axms, thms
  4544 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of
  4545 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:
  4546 isatool fixdots ensures space after dots (e.g. "%x. x"); set
  4547 long_names for fully qualified output names; NOTE: ML programs
  4548 (special tactics, packages etc.) referring to internal names may have
  4549 to be adapted to cope with fully qualified names; in case of severe
  4550 backward campatibility problems try setting 'global_names' at compile
  4551 time to have enrything declared within a flat name space; one may also
  4552 fine tune name declarations in theories via the 'global' and 'local'
  4553 section;
  4554 
  4555 * reimplemented the implicit simpset and claset using the new anytype
  4556 data filed in signatures; references simpset:simpset ref etc. are
  4557 replaced by functions simpset:unit->simpset and
  4558 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp
  4559 to patch your ML files accordingly;
  4560 
  4561 * HTML output now includes theory graph data for display with Java
  4562 applet or isatool browser; data generated automatically via isatool
  4563 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);
  4564 
  4565 * defs may now be conditional; improved rewrite_goals_tac to handle
  4566 conditional equations;
  4567 
  4568 * defs now admits additional type arguments, using TYPE('a) syntax;
  4569 
  4570 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always
  4571 creates a new theory node; implicit merge of thms' signatures is
  4572 restricted to 'trivial' ones; COMPATIBILITY: one may have to use
  4573 transfer:theory->thm->thm in (rare) cases;
  4574 
  4575 * improved handling of draft signatures / theories; draft thms (and
  4576 ctyps, cterms) are automatically promoted to real ones;
  4577 
  4578 * slightly changed interfaces for oracles: admit many per theory, named
  4579 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;
  4580 
  4581 * print_goals: optional output of const types (set show_consts and
  4582 show_types);
  4583 
  4584 * improved output of warnings (###) and errors (***);
  4585 
  4586 * subgoal_tac displays a warning if the new subgoal has type variables;
  4587 
  4588 * removed old README and Makefiles;
  4589 
  4590 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;
  4591 
  4592 * removed obsolete init_pps and init_database;
  4593 
  4594 * deleted the obsolete tactical STATE, which was declared by
  4595     fun STATE tacfun st = tacfun st st;
  4596 
  4597 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~
  4598 (which abbreviates $HOME);
  4599 
  4600 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:
  4601 use isatool fixseq to adapt your ML programs (this works for fully
  4602 qualified references to the Sequence structure only!);
  4603 
  4604 * use_thy no longer requires writable current directory; it always
  4605 reloads .ML *and* .thy file, if either one is out of date;
  4606 
  4607 
  4608 *** Classical Reasoner ***
  4609 
  4610 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new
  4611 tactics that use classical reasoning to simplify a subgoal without
  4612 splitting it into several subgoals;
  4613 
  4614 * Safe_tac: like safe_tac but uses the default claset;
  4615 
  4616 
  4617 *** Simplifier ***
  4618 
  4619 * added simplification meta rules:
  4620     (asm_)(full_)simplify: simpset -> thm -> thm;
  4621 
  4622 * simplifier.ML no longer part of Pure -- has to be loaded by object
  4623 logics (again);
  4624 
  4625 * added prems argument to simplification procedures;
  4626 
  4627 * HOL, FOL, ZF: added infix function `addsplits':
  4628   instead of `<simpset> setloop (split_tac <thms>)'
  4629   you can simply write `<simpset> addsplits <thms>'
  4630 
  4631 
  4632 *** Syntax ***
  4633 
  4634 * TYPE('a) syntax for type reflection terms;
  4635 
  4636 * no longer handles consts with name "" -- declare as 'syntax' instead;
  4637 
  4638 * pretty printer: changed order of mixfix annotation preference (again!);
  4639 
  4640 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;
  4641 
  4642 
  4643 *** HOL ***
  4644 
  4645 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4646   with `addloop' of the simplifier to faciliate case splitting in premises.
  4647 
  4648 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;
  4649 
  4650 * HOL/Auth: new protocol proofs including some for the Internet
  4651   protocol TLS;
  4652 
  4653 * HOL/Map: new theory of `maps' a la VDM;
  4654 
  4655 * HOL/simplifier: simplification procedures nat_cancel_sums for
  4656 cancelling out common nat summands from =, <, <= (in)equalities, or
  4657 differences; simplification procedures nat_cancel_factor for
  4658 cancelling common factor from =, <, <= (in)equalities over natural
  4659 sums; nat_cancel contains both kinds of procedures, it is installed by
  4660 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;
  4661 
  4662 * HOL/simplifier: terms of the form
  4663   `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)'  (or t=x)
  4664   are rewritten to
  4665   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',
  4666   and those of the form
  4667   `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)'  (or t=x)
  4668   are rewritten to
  4669   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',
  4670 
  4671 * HOL/datatype
  4672   Each datatype `t' now comes with a theorem `split_t_case' of the form
  4673 
  4674   P(t_case f1 ... fn x) =
  4675      ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &
  4676         ...
  4677        (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))
  4678      )
  4679 
  4680   and a theorem `split_t_case_asm' of the form
  4681 
  4682   P(t_case f1 ... fn x) =
  4683     ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |
  4684         ...
  4685        (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))
  4686      )
  4687   which can be added to a simpset via `addsplits'. The existing theorems
  4688   expand_list_case and expand_option_case have been renamed to
  4689   split_list_case and split_option_case.
  4690 
  4691 * HOL/Arithmetic:
  4692   - `pred n' is automatically converted to `n-1'.
  4693     Users are strongly encouraged not to use `pred' any longer,
  4694     because it will disappear altogether at some point.
  4695   - Users are strongly encouraged to write "0 < n" rather than
  4696     "n ~= 0". Theorems and proof tools have been modified towards this
  4697     `standard'.
  4698 
  4699 * HOL/Lists:
  4700   the function "set_of_list" has been renamed "set" (and its theorems too);
  4701   the function "nth" now takes its arguments in the reverse order and
  4702   has acquired the infix notation "!" as in "xs!n".
  4703 
  4704 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};
  4705 
  4706 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its
  4707   specialist theorems (like UN1_I) are gone.  Similarly for (INT x.B x);
  4708 
  4709 * HOL/record: extensible records with schematic structural subtyping
  4710 (single inheritance); EXPERIMENTAL version demonstrating the encoding,
  4711 still lacks various theorems and concrete record syntax;
  4712 
  4713 
  4714 *** HOLCF ***
  4715 
  4716 * removed "axioms" and "generated by" sections;
  4717 
  4718 * replaced "ops" section by extended "consts" section, which is capable of
  4719   handling the continuous function space "->" directly;
  4720 
  4721 * domain package:
  4722   . proves theorems immediately and stores them in the theory,
  4723   . creates hierachical name space,
  4724   . now uses normal mixfix annotations (instead of cinfix...),
  4725   . minor changes to some names and values (for consistency),
  4726   . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,
  4727   . separator between mutual domain defs: changed "," to "and",
  4728   . improved handling of sort constraints;  now they have to
  4729     appear on the left-hand side of the equations only;
  4730 
  4731 * fixed LAM <x,y,zs>.b syntax;
  4732 
  4733 * added extended adm_tac to simplifier in HOLCF -- can now discharge
  4734 adm (%x. P (t x)), where P is chainfinite and t continuous;
  4735 
  4736 
  4737 *** FOL and ZF ***
  4738 
  4739 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4740   with `addloop' of the simplifier to faciliate case splitting in premises.
  4741 
  4742 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as
  4743 in HOL, they strip ALL and --> from proved theorems;
  4744 
  4745 
  4746 
  4747 New in Isabelle94-8 (May 1997)
  4748 ------------------------------
  4749 
  4750 *** General Changes ***
  4751 
  4752 * new utilities to build / run / maintain Isabelle etc. (in parts
  4753 still somewhat experimental); old Makefiles etc. still functional;
  4754 
  4755 * new 'Isabelle System Manual';
  4756 
  4757 * INSTALL text, together with ./configure and ./build scripts;
  4758 
  4759 * reimplemented type inference for greater efficiency, better error
  4760 messages and clean internal interface;
  4761 
  4762 * prlim command for dealing with lots of subgoals (an easier way of
  4763 setting goals_limit);
  4764 
  4765 
  4766 *** Syntax ***
  4767 
  4768 * supports alternative (named) syntax tables (parser and pretty
  4769 printer); internal interface is provided by add_modesyntax(_i);
  4770 
  4771 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to
  4772 be used in conjunction with the Isabelle symbol font; uses the
  4773 "symbols" syntax table;
  4774 
  4775 * added token_translation interface (may translate name tokens in
  4776 arbitrary ways, dependent on their type (free, bound, tfree, ...) and
  4777 the current print_mode); IMPORTANT: user print translation functions
  4778 are responsible for marking newly introduced bounds
  4779 (Syntax.mark_boundT);
  4780 
  4781 * token translations for modes "xterm" and "xterm_color" that display
  4782 names in bold, underline etc. or colors (which requires a color
  4783 version of xterm);
  4784 
  4785 * infixes may now be declared with names independent of their syntax;
  4786 
  4787 * added typed_print_translation (like print_translation, but may
  4788 access type of constant);
  4789 
  4790 
  4791 *** Classical Reasoner ***
  4792 
  4793 Blast_tac: a new tactic!  It is often more powerful than fast_tac, but has
  4794 some limitations.  Blast_tac...
  4795   + ignores addss, addbefore, addafter; this restriction is intrinsic
  4796   + ignores elimination rules that don't have the correct format
  4797         (the conclusion MUST be a formula variable)
  4798   + ignores types, which can make HOL proofs fail
  4799   + rules must not require higher-order unification, e.g. apply_type in ZF
  4800     [message "Function Var's argument not a bound variable" relates to this]
  4801   + its proof strategy is more general but can actually be slower
  4802 
  4803 * substitution with equality assumptions no longer permutes other
  4804 assumptions;
  4805 
  4806 * minor changes in semantics of addafter (now called addaltern); renamed
  4807 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper
  4808 (and access functions for it);
  4809 
  4810 * improved combination of classical reasoner and simplifier:
  4811   + functions for handling clasimpsets
  4812   + improvement of addss: now the simplifier is called _after_ the
  4813     safe steps.
  4814   + safe variant of addss called addSss: uses safe simplifications
  4815     _during_ the safe steps. It is more complete as it allows multiple
  4816     instantiations of unknowns (e.g. with slow_tac).
  4817 
  4818 *** Simplifier ***
  4819 
  4820 * added interface for simplification procedures (functions that
  4821 produce *proven* rewrite rules on the fly, depending on current
  4822 redex);
  4823 
  4824 * ordering on terms as parameter (used for ordered rewriting);
  4825 
  4826 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;
  4827 
  4828 * the solver is now split into a safe and an unsafe part.
  4829 This should be invisible for the normal user, except that the
  4830 functions setsolver and addsolver have been renamed to setSolver and
  4831 addSolver; added safe_asm_full_simp_tac;
  4832 
  4833 
  4834 *** HOL ***
  4835 
  4836 * a generic induction tactic `induct_tac' which works for all datatypes and
  4837 also for type `nat';
  4838 
  4839 * a generic case distinction tactic `exhaust_tac' which works for all
  4840 datatypes and also for type `nat';
  4841 
  4842 * each datatype comes with a function `size';
  4843 
  4844 * patterns in case expressions allow tuple patterns as arguments to
  4845 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';
  4846 
  4847 * primrec now also works with type nat;
  4848 
  4849 * recdef: a new declaration form, allows general recursive functions to be
  4850 defined in theory files.  See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.
  4851 
  4852 * the constant for negation has been renamed from "not" to "Not" to
  4853 harmonize with FOL, ZF, LK, etc.;
  4854 
  4855 * HOL/ex/LFilter theory of a corecursive "filter" functional for
  4856 infinite lists;
  4857 
  4858 * HOL/Modelcheck demonstrates invocation of model checker oracle;
  4859 
  4860 * HOL/ex/Ring.thy declares cring_simp, which solves equational
  4861 problems in commutative rings, using axiomatic type classes for + and *;
  4862 
  4863 * more examples in HOL/MiniML and HOL/Auth;
  4864 
  4865 * more default rewrite rules for quantifiers, union/intersection;
  4866 
  4867 * a new constant `arbitrary == @x.False';
  4868 
  4869 * HOLCF/IOA replaces old HOL/IOA;
  4870 
  4871 * HOLCF changes: derived all rules and arities
  4872   + axiomatic type classes instead of classes
  4873   + typedef instead of faking type definitions
  4874   + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.
  4875   + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po
  4876   + eliminated the types void, one, tr
  4877   + use unit lift and bool lift (with translations) instead of one and tr
  4878   + eliminated blift from Lift3.thy (use Def instead of blift)
  4879   all eliminated rules are derived as theorems --> no visible changes ;
  4880 
  4881 
  4882 *** ZF ***
  4883 
  4884 * ZF now has Fast_tac, Simp_tac and Auto_tac.  Union_iff is a now a default
  4885 rewrite rule; this may affect some proofs.  eq_cs is gone but can be put back
  4886 as ZF_cs addSIs [equalityI];
  4887 
  4888 
  4889 
  4890 New in Isabelle94-7 (November 96)
  4891 ---------------------------------
  4892 
  4893 * allowing negative levels (as offsets) in prlev and choplev;
  4894 
  4895 * super-linear speedup for large simplifications;
  4896 
  4897 * FOL, ZF and HOL now use miniscoping: rewriting pushes
  4898 quantifications in as far as possible (COULD MAKE EXISTING PROOFS
  4899 FAIL); can suppress it using the command Delsimps (ex_simps @
  4900 all_simps); De Morgan laws are also now included, by default;
  4901 
  4902 * improved printing of ==>  :  ~:
  4903 
  4904 * new object-logic "Sequents" adds linear logic, while replacing LK
  4905 and Modal (thanks to Sara Kalvala);
  4906 
  4907 * HOL/Auth: correctness proofs for authentication protocols;
  4908 
  4909 * HOL: new auto_tac combines rewriting and classical reasoning (many
  4910 examples on HOL/Auth);
  4911 
  4912 * HOL: new command AddIffs for declaring theorems of the form P=Q to
  4913 the rewriter and classical reasoner simultaneously;
  4914 
  4915 * function uresult no longer returns theorems in "standard" format;
  4916 regain previous version by: val uresult = standard o uresult;
  4917 
  4918 
  4919 
  4920 New in Isabelle94-6
  4921 -------------------
  4922 
  4923 * oracles -- these establish an interface between Isabelle and trusted
  4924 external reasoners, which may deliver results as theorems;
  4925 
  4926 * proof objects (in particular record all uses of oracles);
  4927 
  4928 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;
  4929 
  4930 * "constdefs" section in theory files;
  4931 
  4932 * "primrec" section (HOL) no longer requires names;
  4933 
  4934 * internal type "tactic" now simply "thm -> thm Sequence.seq";
  4935 
  4936 
  4937 
  4938 New in Isabelle94-5
  4939 -------------------
  4940 
  4941 * reduced space requirements;
  4942 
  4943 * automatic HTML generation from theories;
  4944 
  4945 * theory files no longer require "..." (quotes) around most types;
  4946 
  4947 * new examples, including two proofs of the Church-Rosser theorem;
  4948 
  4949 * non-curried (1994) version of HOL is no longer distributed;
  4950 
  4951 
  4952 
  4953 New in Isabelle94-4
  4954 -------------------
  4955 
  4956 * greatly reduced space requirements;
  4957 
  4958 * theory files (.thy) no longer require \...\ escapes at line breaks;
  4959 
  4960 * searchable theorem database (see the section "Retrieving theorems" on
  4961 page 8 of the Reference Manual);
  4962 
  4963 * new examples, including Grabczewski's monumental case study of the
  4964 Axiom of Choice;
  4965 
  4966 * The previous version of HOL renamed to Old_HOL;
  4967 
  4968 * The new version of HOL (previously called CHOL) uses a curried syntax
  4969 for functions.  Application looks like f a b instead of f(a,b);
  4970 
  4971 * Mutually recursive inductive definitions finally work in HOL;
  4972 
  4973 * In ZF, pattern-matching on tuples is now available in all abstractions and
  4974 translates to the operator "split";
  4975 
  4976 
  4977 
  4978 New in Isabelle94-3
  4979 -------------------
  4980 
  4981 * new infix operator, addss, allowing the classical reasoner to
  4982 perform simplification at each step of its search.  Example:
  4983         fast_tac (cs addss ss)
  4984 
  4985 * a new logic, CHOL, the same as HOL, but with a curried syntax
  4986 for functions.  Application looks like f a b instead of f(a,b).  Also pairs
  4987 look like (a,b) instead of <a,b>;
  4988 
  4989 * PLEASE NOTE: CHOL will eventually replace HOL!
  4990 
  4991 * In CHOL, pattern-matching on tuples is now available in all abstractions.
  4992 It translates to the operator "split".  A new theory of integers is available;
  4993 
  4994 * In ZF, integer numerals now denote two's-complement binary integers.
  4995 Arithmetic operations can be performed by rewriting.  See ZF/ex/Bin.ML;
  4996 
  4997 * Many new examples: I/O automata, Church-Rosser theorem, equivalents
  4998 of the Axiom of Choice;
  4999 
  5000 
  5001 
  5002 New in Isabelle94-2
  5003 -------------------
  5004 
  5005 * Significantly faster resolution;
  5006 
  5007 * the different sections in a .thy file can now be mixed and repeated
  5008 freely;
  5009 
  5010 * Database of theorems for FOL, HOL and ZF.  New
  5011 commands including qed, qed_goal and bind_thm store theorems in the database.
  5012 
  5013 * Simple database queries: return a named theorem (get_thm) or all theorems of
  5014 a given theory (thms_of), or find out what theory a theorem was proved in
  5015 (theory_of_thm);
  5016 
  5017 * Bugs fixed in the inductive definition and datatype packages;
  5018 
  5019 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs
  5020 and HOL_dup_cs obsolete;
  5021 
  5022 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1
  5023 have been removed;
  5024 
  5025 * Simpler definition of function space in ZF;
  5026 
  5027 * new results about cardinal and ordinal arithmetic in ZF;
  5028 
  5029 * 'subtype' facility in HOL for introducing new types as subsets of existing
  5030 types;
  5031 
  5032 
  5033 $Id$