NEWS
author haftmann
Thu Jul 19 21:47:34 2007 +0200 (2007-07-19)
changeset 23850 f1434532a562
parent 23783 e4d514f81d95
child 23881 851c74f1bb69
permissions -rw-r--r--
updated
     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 * Legacy goal package: reduced interface to the bare minimum required
    28 to keep existing proof scripts running.  Most other user-level
    29 functions are now part of the OldGoals structure, which is *not* open
    30 by default (consider isatool expandshort before open OldGoals).
    31 Removed top_sg, prin, printyp, pprint_term/typ altogether, because
    32 these tend to cause confusion about the actual goal (!) context being
    33 used here, which is not necessarily the same as the_context().
    34 
    35 * Command 'find_theorems': supports "*" wild-card in "name:"
    36 criterion; "with_dups" option.  Certain ProofGeneral versions might
    37 support a specific search form (see ProofGeneral/CHANGES).
    38 
    39 * The ``prems limit'' option (cf. ProofContext.prems_limit) is now -1
    40 by default, which means that "prems" (and also "fixed variables") are
    41 suppressed from proof state output.  Note that the ProofGeneral
    42 settings mechanism allows to change and save options persistently, but
    43 older versions of Isabelle will fail to start up if a negative prems
    44 limit is imposed.
    45 
    46 * Local theory targets may be specified by non-nested blocks of
    47 ``context/locale/class ... begin'' followed by ``end''.  The body may
    48 contain definitions, theorems etc., including any derived mechanism
    49 that has been implemented on top of these primitives.  This concept
    50 generalizes the existing ``theorem (in ...)'' towards more versatility
    51 and scalability.
    52 
    53 * Proof General interface: proper undo of final 'end' command;
    54 discontinued Isabelle/classic mode (ML proof scripts).
    55 
    56 
    57 *** Document preparation ***
    58 
    59 * Added antiquotation @{theory name} which prints the given name,
    60 after checking that it refers to a valid ancestor theory in the
    61 current context.
    62 
    63 * Added antiquotations @{ML_type text} and @{ML_struct text} which
    64 check the given source text as ML type/structure, printing verbatim.
    65 
    66 * Added antiquotation @{abbrev "c args"} which prints the abbreviation
    67 "c args == rhs" given in the current context.  (Any number of
    68 arguments may be given on the LHS.)
    69 
    70 
    71 
    72 *** Pure ***
    73 
    74 * code generator: consts in 'consts_code' Isar commands are now referred
    75   to by usual term syntax (including optional type annotations).
    76 
    77 * code generator: 
    78   - Isar 'definition's, 'constdef's and primitive instance definitions are added
    79     explicitly to the table of defining equations
    80   - primitive definitions are not used as defining equations by default any longer
    81   - defining equations are now definitly restricted to meta "==" and object
    82         equality "="
    83   - HOL theories have been adopted accordingly
    84 
    85 * class_package.ML offers a combination of axclasses and locales to
    86 achieve Haskell-like type classes in Isabelle.  See
    87 HOL/ex/Classpackage.thy for examples.
    88 
    89 * Yet another code generator framework allows to generate executable
    90 code for ML and Haskell (including "class"es).  A short usage sketch:
    91 
    92     internal compilation:
    93         code_gen <list of constants (term syntax)> in SML
    94     writing SML code to a file:
    95         code_gen <list of constants (term syntax)> in SML <filename>
    96     writing OCaml code to a file:
    97         code_gen <list of constants (term syntax)> in OCaml <filename>
    98     writing Haskell code to a bunch of files:
    99         code_gen <list of constants (term syntax)> in Haskell <filename>
   100 
   101 Reasonable default setup of framework in HOL/Main.
   102 
   103 Theorem attributs for selecting and transforming function equations theorems:
   104 
   105     [code fun]:        select a theorem as function equation for a specific constant
   106     [code fun del]:    deselect a theorem as function equation for a specific constant
   107     [code inline]:     select an equation theorem for unfolding (inlining) in place
   108     [code inline del]: deselect an equation theorem for unfolding (inlining) in place
   109 
   110 User-defined serializations (target in {SML, OCaml, Haskell}):
   111 
   112     code_const <and-list of constants (term syntax)>
   113       {(target) <and-list of const target syntax>}+
   114 
   115     code_type <and-list of type constructors>
   116       {(target) <and-list of type target syntax>}+
   117 
   118     code_instance <and-list of instances>
   119       {(target)}+
   120         where instance ::= <type constructor> :: <class>
   121 
   122     code_class <and_list of classes>
   123       {(target) <and-list of class target syntax>}+
   124         where class target syntax ::= <class name> {where {<classop> == <target syntax>}+}?
   125 
   126 code_instance and code_class only apply to target Haskell.
   127 
   128 See HOL theories and HOL/ex/Codegenerator*.thy for usage examples.
   129 Doc/Isar/Advanced/Codegen/ provides a tutorial.
   130 
   131 * Command 'no_translations' removes translation rules from theory
   132 syntax.
   133 
   134 * Overloaded definitions are now actually checked for acyclic
   135 dependencies.  The overloading scheme is slightly more general than
   136 that of Haskell98, although Isabelle does not demand an exact
   137 correspondence to type class and instance declarations.
   138 INCOMPATIBILITY, use ``defs (unchecked overloaded)'' to admit more
   139 exotic versions of overloading -- at the discretion of the user!
   140 
   141 Polymorphic constants are represented via type arguments, i.e. the
   142 instantiation that matches an instance against the most general
   143 declaration given in the signature.  For example, with the declaration
   144 c :: 'a => 'a => 'a, an instance c :: nat => nat => nat is represented
   145 as c(nat).  Overloading is essentially simultaneous structural
   146 recursion over such type arguments.  Incomplete specification patterns
   147 impose global constraints on all occurrences, e.g. c('a * 'a) on the
   148 LHS means that more general c('a * 'b) will be disallowed on any RHS.
   149 Command 'print_theory' outputs the normalized system of recursive
   150 equations, see section "definitions".
   151 
   152 * Isar: method "assumption" (and implicit closing of subproofs) now
   153 takes simple non-atomic goal assumptions into account: after applying
   154 an assumption as a rule the resulting subgoals are solved by atomic
   155 assumption steps.  This is particularly useful to finish 'obtain'
   156 goals, such as "!!x. (!!x. P x ==> thesis) ==> P x ==> thesis",
   157 without referring to the original premise "!!x. P x ==> thesis" in the
   158 Isar proof context.  POTENTIAL INCOMPATIBILITY: method "assumption" is
   159 more permissive.
   160 
   161 * Isar: implicit use of prems from the Isar proof context is
   162 considered a legacy feature.  Common applications like ``have A .''
   163 may be replaced by ``have A by fact'' or ``note `A`''.  In general,
   164 referencing facts explicitly here improves readability and
   165 maintainability of proof texts.
   166 
   167 * Isar: improper proof element 'guess' is like 'obtain', but derives
   168 the obtained context from the course of reasoning!  For example:
   169 
   170   assume "EX x y. A x & B y"   -- "any previous fact"
   171   then guess x and y by clarify
   172 
   173 This technique is potentially adventurous, depending on the facts and
   174 proof tools being involved here.
   175 
   176 * Isar: known facts from the proof context may be specified as literal
   177 propositions, using ASCII back-quote syntax.  This works wherever
   178 named facts used to be allowed so far, in proof commands, proof
   179 methods, attributes etc.  Literal facts are retrieved from the context
   180 according to unification of type and term parameters.  For example,
   181 provided that "A" and "A ==> B" and "!!x. P x ==> Q x" are known
   182 theorems in the current context, then these are valid literal facts:
   183 `A` and `A ==> B` and `!!x. P x ==> Q x" as well as `P a ==> Q a` etc.
   184 
   185 There is also a proof method "fact" which does the same composition
   186 for explicit goal states, e.g. the following proof texts coincide with
   187 certain special cases of literal facts:
   188 
   189   have "A" by fact                 ==  note `A`
   190   have "A ==> B" by fact           ==  note `A ==> B`
   191   have "!!x. P x ==> Q x" by fact  ==  note `!!x. P x ==> Q x`
   192   have "P a ==> Q a" by fact       ==  note `P a ==> Q a`
   193 
   194 * Isar: ":" (colon) is no longer a symbolic identifier character in
   195 outer syntax.  Thus symbolic identifiers may be used without
   196 additional white space in declarations like this: ``assume *: A''.
   197 
   198 * Isar: 'print_facts' prints all local facts of the current context,
   199 both named and unnamed ones.
   200 
   201 * Isar: 'def' now admits simultaneous definitions, e.g.:
   202 
   203   def x == "t" and y == "u"
   204 
   205 * Isar: added command 'unfolding', which is structurally similar to
   206 'using', but affects both the goal state and facts by unfolding given
   207 rewrite rules.  Thus many occurrences of the 'unfold' method or
   208 'unfolded' attribute may be replaced by first-class proof text.
   209 
   210 * Isar: methods 'unfold' / 'fold', attributes 'unfolded' / 'folded',
   211 and command 'unfolding' now all support object-level equalities
   212 (potentially conditional).  The underlying notion of rewrite rule is
   213 analogous to the 'rule_format' attribute, but *not* that of the
   214 Simplifier (which is usually more generous).
   215 
   216 * Isar: the goal restriction operator [N] (default N = 1) evaluates a
   217 method expression within a sandbox consisting of the first N
   218 sub-goals, which need to exist.  For example, ``simp_all [3]''
   219 simplifies the first three sub-goals, while (rule foo, simp_all)[]
   220 simplifies all new goals that emerge from applying rule foo to the
   221 originally first one.
   222 
   223 * Isar: schematic goals are no longer restricted to higher-order
   224 patterns; e.g. ``lemma "?P(?x)" by (rule TrueI)'' now works as
   225 expected.
   226 
   227 * Isar: the conclusion of a long theorem statement is now either
   228 'shows' (a simultaneous conjunction, as before), or 'obtains'
   229 (essentially a disjunction of cases with local parameters and
   230 assumptions).  The latter allows to express general elimination rules
   231 adequately; in this notation common elimination rules look like this:
   232 
   233   lemma exE:    -- "EX x. P x ==> (!!x. P x ==> thesis) ==> thesis"
   234     assumes "EX x. P x"
   235     obtains x where "P x"
   236 
   237   lemma conjE:  -- "A & B ==> (A ==> B ==> thesis) ==> thesis"
   238     assumes "A & B"
   239     obtains A and B
   240 
   241   lemma disjE:  -- "A | B ==> (A ==> thesis) ==> (B ==> thesis) ==> thesis"
   242     assumes "A | B"
   243     obtains
   244       A
   245     | B
   246 
   247 The subsequent classical rules even refer to the formal "thesis"
   248 explicitly:
   249 
   250   lemma classical:     -- "(~ thesis ==> thesis) ==> thesis"
   251     obtains "~ thesis"
   252 
   253   lemma Peirce's_Law:  -- "((thesis ==> something) ==> thesis) ==> thesis"
   254     obtains "thesis ==> something"
   255 
   256 The actual proof of an 'obtains' statement is analogous to that of the
   257 Isar proof element 'obtain', only that there may be several cases.
   258 Optional case names may be specified in parentheses; these will be
   259 available both in the present proof and as annotations in the
   260 resulting rule, for later use with the 'cases' method (cf. attribute
   261 case_names).
   262 
   263 * Isar: the assumptions of a long theorem statement are available as
   264 "assms" fact in the proof context.  This is more appropriate than the
   265 (historical) "prems", which refers to all assumptions of the current
   266 context, including those from the target locale, proof body etc.
   267 
   268 * Isar: 'print_statement' prints theorems from the current theory or
   269 proof context in long statement form, according to the syntax of a
   270 top-level lemma.
   271 
   272 * Isar: 'obtain' takes an optional case name for the local context
   273 introduction rule (default "that").
   274 
   275 * Isar: removed obsolete 'concl is' patterns.  INCOMPATIBILITY, use
   276 explicit (is "_ ==> ?foo") in the rare cases where this still happens
   277 to occur.
   278 
   279 * Pure: syntax "CONST name" produces a fully internalized constant
   280 according to the current context.  This is particularly useful for
   281 syntax translations that should refer to internal constant
   282 representations independently of name spaces.
   283 
   284 * Pure: syntax constant for foo (binder "FOO ") is called "foo_binder"
   285 instead of "FOO ". This allows multiple binder declarations to coexist
   286 in the same context.  INCOMPATIBILITY.
   287 
   288 * Isar/locales: 'notation' provides a robust interface to the 'syntax'
   289 primitive that also works in a locale context (both for constants and
   290 fixed variables).  Type declaration and internal syntactic
   291 representation of given constants retrieved from the context.
   292 
   293 * Isar/locales: new derived specification elements 'axiomatization',
   294 'definition', 'abbreviation', which support type-inference, admit
   295 object-level specifications (equality, equivalence).  See also the
   296 isar-ref manual.  Examples:
   297 
   298   axiomatization
   299     eq  (infix "===" 50) where
   300     eq_refl: "x === x" and eq_subst: "x === y ==> P x ==> P y"
   301 
   302   definition "f x y = x + y + 1"
   303   definition g where "g x = f x x"
   304 
   305   abbreviation
   306     neq  (infix "=!=" 50) where
   307     "x =!= y == ~ (x === y)"
   308 
   309 These specifications may be also used in a locale context.  Then the
   310 constants being introduced depend on certain fixed parameters, and the
   311 constant name is qualified by the locale base name.  An internal
   312 abbreviation takes care for convenient input and output, making the
   313 parameters implicit and using the original short name.  See also
   314 HOL/ex/Abstract_NAT.thy for an example of deriving polymorphic
   315 entities from a monomorphic theory.
   316 
   317 Presently, abbreviations are only available 'in' a target locale, but
   318 not inherited by general import expressions.  Also note that
   319 'abbreviation' may be used as a type-safe replacement for 'syntax' +
   320 'translations' in common applications.
   321 
   322 Concrete syntax is attached to specified constants in internal form,
   323 independently of name spaces.  The parse tree representation is
   324 slightly different -- use 'notation' instead of raw 'syntax', and
   325 'translations' with explicit "CONST" markup to accommodate this.
   326 
   327 * Pure: command 'print_abbrevs' prints all constant abbreviations of
   328 the current context.  Print mode "no_abbrevs" prevents inversion of
   329 abbreviations on output.
   330 
   331 * Isar/locales: improved parameter handling:
   332 - use of locales "var" and "struct" no longer necessary;
   333 - parameter renamings are no longer required to be injective.
   334   This enables, for example, to define a locale for endomorphisms thus:
   335   locale endom = homom mult mult h.
   336 
   337 * Isar/locales: changed the way locales with predicates are defined.
   338 Instead of accumulating the specification, the imported expression is
   339 now an interpretation.  INCOMPATIBILITY: different normal form of
   340 locale expressions.  In particular, in interpretations of locales with
   341 predicates, goals repesenting already interpreted fragments are not
   342 removed automatically.  Use methods `intro_locales' and
   343 `unfold_locales'; see below.
   344 
   345 * Isar/locales: new methods `intro_locales' and `unfold_locales'
   346 provide backward reasoning on locales predicates.  The methods are
   347 aware of interpretations and discharge corresponding goals.
   348 `intro_locales' is less aggressive then `unfold_locales' and does not
   349 unfold predicates to assumptions.
   350 
   351 * Isar/locales: the order in which locale fragments are accumulated
   352 has changed.  This enables to override declarations from fragments due
   353 to interpretations -- for example, unwanted simp rules.
   354 
   355 * Provers/induct: improved internal context management to support
   356 local fixes and defines on-the-fly.  Thus explicit meta-level
   357 connectives !! and ==> are rarely required anymore in inductive goals
   358 (using object-logic connectives for this purpose has been long
   359 obsolete anyway).  The subsequent proof patterns illustrate advanced
   360 techniques of natural induction; general datatypes and inductive sets
   361 work analogously (see also src/HOL/Lambda for realistic examples).
   362 
   363 (1) This is how to ``strengthen'' an inductive goal wrt. certain
   364 parameters:
   365 
   366   lemma
   367     fixes n :: nat and x :: 'a
   368     assumes a: "A n x"
   369     shows "P n x"
   370     using a                     -- {* make induct insert fact a *}
   371   proof (induct n arbitrary: x) -- {* generalize goal to "!!x. A n x ==> P n x" *}
   372     case 0
   373     show ?case sorry
   374   next
   375     case (Suc n)
   376     note `!!x. A n x ==> P n x` -- {* induction hypothesis, according to induction rule *}
   377     note `A (Suc n) x`          -- {* induction premise, stemming from fact a *}
   378     show ?case sorry
   379   qed
   380 
   381 (2) This is how to perform induction over ``expressions of a certain
   382 form'', using a locally defined inductive parameter n == "a x"
   383 together with strengthening (the latter is usually required to get
   384 sufficiently flexible induction hypotheses):
   385 
   386   lemma
   387     fixes a :: "'a => nat"
   388     assumes a: "A (a x)"
   389     shows "P (a x)"
   390     using a
   391   proof (induct n == "a x" arbitrary: x)
   392     ...
   393 
   394 See also HOL/Isar_examples/Puzzle.thy for an application of the this
   395 particular technique.
   396 
   397 (3) This is how to perform existential reasoning ('obtains' or
   398 'obtain') by induction, while avoiding explicit object-logic
   399 encodings:
   400 
   401   lemma
   402     fixes n :: nat
   403     obtains x :: 'a where "P n x" and "Q n x"
   404   proof (induct n arbitrary: thesis)
   405     case 0
   406     obtain x where "P 0 x" and "Q 0 x" sorry
   407     then show thesis by (rule 0)
   408   next
   409     case (Suc n)
   410     obtain x where "P n x" and "Q n x" by (rule Suc.hyps)
   411     obtain x where "P (Suc n) x" and "Q (Suc n) x" sorry
   412     then show thesis by (rule Suc.prems)
   413   qed
   414 
   415 Here the 'arbitrary: thesis' specification essentially modifies the
   416 scope of the formal thesis parameter, in order to the get the whole
   417 existence statement through the induction as expected.
   418 
   419 * Provers/induct: mutual induction rules are now specified as a list
   420 of rule sharing the same induction cases.  HOL packages usually
   421 provide foo_bar.inducts for mutually defined items foo and bar
   422 (e.g. inductive sets or datatypes).  INCOMPATIBILITY, users need to
   423 specify mutual induction rules differently, i.e. like this:
   424 
   425   (induct rule: foo_bar.inducts)
   426   (induct set: foo bar)
   427   (induct type: foo bar)
   428 
   429 The ML function ProjectRule.projections turns old-style rules into the
   430 new format.
   431 
   432 * Provers/induct: improved handling of simultaneous goals.  Instead of
   433 introducing object-level conjunction, the statement is now split into
   434 several conclusions, while the corresponding symbolic cases are
   435 nested accordingly.  INCOMPATIBILITY, proofs need to be structured
   436 explicitly.  For example:
   437 
   438   lemma
   439     fixes n :: nat
   440     shows "P n" and "Q n"
   441   proof (induct n)
   442     case 0 case 1
   443     show "P 0" sorry
   444   next
   445     case 0 case 2
   446     show "Q 0" sorry
   447   next
   448     case (Suc n) case 1
   449     note `P n` and `Q n`
   450     show "P (Suc n)" sorry
   451   next
   452     case (Suc n) case 2
   453     note `P n` and `Q n`
   454     show "Q (Suc n)" sorry
   455   qed
   456 
   457 The split into subcases may be deferred as follows -- this is
   458 particularly relevant for goal statements with local premises.
   459 
   460   lemma
   461     fixes n :: nat
   462     shows "A n ==> P n" and "B n ==> Q n"
   463   proof (induct n)
   464     case 0
   465     {
   466       case 1
   467       note `A 0`
   468       show "P 0" sorry
   469     next
   470       case 2
   471       note `B 0`
   472       show "Q 0" sorry
   473     }
   474   next
   475     case (Suc n)
   476     note `A n ==> P n` and `B n ==> Q n`
   477     {
   478       case 1
   479       note `A (Suc n)`
   480       show "P (Suc n)" sorry
   481     next
   482       case 2
   483       note `B (Suc n)`
   484       show "Q (Suc n)" sorry
   485     }
   486   qed
   487 
   488 If simultaneous goals are to be used with mutual rules, the statement
   489 needs to be structured carefully as a two-level conjunction, using
   490 lists of propositions separated by 'and':
   491 
   492   lemma
   493     shows "a : A ==> P1 a"
   494           "a : A ==> P2 a"
   495       and "b : B ==> Q1 b"
   496           "b : B ==> Q2 b"
   497           "b : B ==> Q3 b"
   498   proof (induct set: A B)
   499 
   500 * Provers/induct: support coinduction as well.  See
   501 src/HOL/Library/Coinductive_List.thy for various examples.
   502 
   503 * Attribute "symmetric" produces result with standardized schematic
   504 variables (index 0).  Potential INCOMPATIBILITY.
   505 
   506 * Simplifier: by default the simplifier trace only shows top level
   507 rewrites now. That is, trace_simp_depth_limit is set to 1 by
   508 default. Thus there is less danger of being flooded by the trace. The
   509 trace indicates where parts have been suppressed.
   510   
   511 * Provers/classical: removed obsolete classical version of elim_format
   512 attribute; classical elim/dest rules are now treated uniformly when
   513 manipulating the claset.
   514 
   515 * Provers/classical: stricter checks to ensure that supplied intro,
   516 dest and elim rules are well-formed; dest and elim rules must have at
   517 least one premise.
   518 
   519 * Provers/classical: attributes dest/elim/intro take an optional
   520 weight argument for the rule (just as the Pure versions).  Weights are
   521 ignored by automated tools, but determine the search order of single
   522 rule steps.
   523 
   524 * Syntax: input syntax now supports dummy variable binding "%_. b",
   525 where the body does not mention the bound variable.  Note that dummy
   526 patterns implicitly depend on their context of bounds, which makes
   527 "{_. _}" match any set comprehension as expected.  Potential
   528 INCOMPATIBILITY -- parse translations need to cope with syntactic
   529 constant "_idtdummy" in the binding position.
   530 
   531 * Syntax: removed obsolete syntactic constant "_K" and its associated
   532 parse translation.  INCOMPATIBILITY -- use dummy abstraction instead,
   533 for example "A -> B" => "Pi A (%_. B)".
   534 
   535 * Pure: 'class_deps' command visualizes the subclass relation, using
   536 the graph browser tool.
   537 
   538 * Pure: 'print_theory' now suppresses entities with internal name
   539 (trailing "_") by default; use '!' option for full details.
   540 
   541 
   542 *** HOL ***
   543 
   544 * Code generator library theories:
   545   * Pretty_Int represents HOL integers by big integer literals in target
   546     languages.
   547   * Pretty_Char represents HOL characters by character literals in target
   548     languages.
   549   * Pretty_Char_chr like Pretty_Char, but also offers treatment of character
   550     codes; includes Pretty_Int.
   551   * Executable_Set allows to generate code for finite sets using lists.
   552   * Executable_Rat implements rational numbers as triples (sign, enumerator,
   553     denominator).
   554   * Executable_Real implements a subset of real numbers, namly those
   555     representable by rational numbers.
   556   * Efficient_Nat implements natural numbers by integers, which in general will
   557     result in higher efficency; pattern matching with 0/Suc is eliminated;
   558     includes Pretty_Int.
   559   * ML_String provides an additional datatype ml_string; in the HOL default
   560     setup, strings in HOL are mapped to lists of HOL characters in SML; values
   561     of type ml_string are mapped to strings in SML.
   562   * ML_Int provides an additional datatype ml_int which is mapped to to SML
   563     built-in integers.
   564 
   565 * New package for inductive predicates
   566 
   567   An n-ary predicate p with m parameters z_1, ..., z_m can now be defined via
   568 
   569     inductive
   570       p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   571       for z_1 :: U_1 and ... and z_n :: U_m
   572     where
   573       rule_1: "... ==> p z_1 ... z_m t_1_1 ... t_1_n"
   574     | ...
   575 
   576   rather than
   577 
   578     consts s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   579 
   580     abbreviation p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   581     where "p z_1 ... z_m x_1 ... x_n == (x_1, ..., x_n) : s z_1 ... z_m"
   582 
   583     inductive "s z_1 ... z_m"
   584     intros
   585       rule_1: "... ==> (t_1_1, ..., t_1_n) : s z_1 ... z_m"
   586       ...
   587 
   588   For backward compatibility, there is a wrapper allowing inductive
   589   sets to be defined with the new package via
   590 
   591     inductive_set
   592       s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   593       for z_1 :: U_1 and ... and z_n :: U_m
   594     where
   595       rule_1: "... ==> (t_1_1, ..., t_1_n) : s z_1 ... z_m"
   596     | ...
   597 
   598   or
   599 
   600     inductive_set
   601       s :: "U_1 => ... => U_m => (T_1 * ... * T_n) set"
   602       and p :: "U_1 => ... => U_m => T_1 => ... => T_n => bool"
   603       for z_1 :: U_1 and ... and z_n :: U_m
   604     where
   605       "p z_1 ... z_m x_1 ... x_n == (x_1, ..., x_n) : s z_1 ... z_m"
   606     | rule_1: "... ==> p z_1 ... z_m t_1_1 ... t_1_n"
   607     | ...
   608 
   609   if the additional syntax "p ..." is required.
   610 
   611   Many examples can be found in the subdirectories Auth, Bali, Induct,
   612   or MicroJava.
   613 
   614   INCOMPATIBILITIES:
   615 
   616   - Since declaration and definition of inductive sets or predicates
   617     is no longer separated, abbreviations involving the newly introduced
   618     sets or predicates must be specified together with the introduction
   619     rules after the "where" keyword (see example above), rather than before
   620     the actual inductive definition.
   621 
   622   - The variables in induction and elimination rules are now quantified
   623     in the order of their occurrence in the introduction rules, rather than
   624     in alphabetical order. Since this may break some proofs, these proofs
   625     either have to be repaired, e.g. by reordering the variables
   626     a_i_1 ... a_i_{k_i} in Isar "case" statements of the form
   627 
   628       case (rule_i a_i_1 ... a_i_{k_i})
   629 
   630     or the old order of quantification has to be restored by explicitly adding
   631     meta-level quantifiers in the introduction rules, i.e.
   632 
   633       | rule_i: "!!a_i_1 ... a_i_{k_i}. ... ==> p z_1 ... z_m t_i_1 ... t_i_n"
   634 
   635   - The format of the elimination rules is now
   636 
   637       p z_1 ... z_m x_1 ... x_n ==>
   638         (!!a_1_1 ... a_1_{k_1}. x_1 = t_1_1 ==> ... ==> x_n = t_1_n ==> ... ==> P)
   639         ==> ... ==> P
   640 
   641     for predicates and
   642 
   643       (x_1, ..., x_n) : s z_1 ... z_m ==>
   644         (!!a_1_1 ... a_1_{k_1}. x_1 = t_1_1 ==> ... ==> x_n = t_1_n ==> ... ==> P)
   645         ==> ... ==> P
   646 
   647     for sets rather than
   648 
   649       x : s z_1 ... z_m ==>
   650         (!!a_1_1 ... a_1_{k_1}. x = (t_1_1, ..., t_1_n) ==> ... ==> P)
   651         ==> ... ==> P
   652 
   653     This may require terms in goals to be expanded to n-tuples (e.g. using case_tac
   654     or simplification with the split_paired_all rule) before the above elimination
   655     rule is applicable.
   656 
   657   - The elimination or case analysis rules for (mutually) inductive sets or
   658     predicates are now called "p_1.cases" ... "p_k.cases". The list of rules
   659     "p_1_..._p_k.elims" is no longer available.
   660 
   661 * Method "metis" proves goals by applying the Metis general-purpose
   662 resolution prover.  Examples are in the directory MetisExamples.  See
   663 also http://gilith.com/software/metis/
   664   
   665 * Command 'sledgehammer' invokes external automatic theorem provers as
   666 background processes.  It generates calls to the "metis" method if
   667 successful. These can be pasted into the proof.  Users do not have to
   668 wait for the automatic provers to return.
   669 
   670 * Case-expressions allow arbitrary constructor-patterns (including "_") and
   671   take their order into account, like in functional programming.
   672   Internally, this is translated into nested case-expressions; missing cases
   673   are added and mapped to the predefined constant "undefined". In complicated
   674   cases printing may no longer show the original input but the internal
   675   form. Lambda-abstractions allow the same form of pattern matching:
   676   "% pat1 => e1 | ..." is an abbreviation for
   677   "%x. case x of pat1 => e1 | ..." where x is a new variable.
   678 
   679 * IntDef: The constant "int :: nat => int" has been removed; now "int"
   680   is an abbreviation for "of_nat :: nat => int". The simplification rules
   681   for "of_nat" have been changed to work like "int" did previously.
   682   (potential INCOMPATIBILITY)
   683   - "of_nat (Suc m)" simplifies to "1 + of_nat m" instead of "of_nat m + 1"
   684   - of_nat_diff and of_nat_mult are no longer default simp rules
   685 
   686 * Method "algebra" solves polynomial equations over (semi)rings using
   687   Groebner bases. The (semi)ring structure is defined by locales and
   688   the tool setup depends on that generic context. Installing the
   689   method for a specific type involves instantiating the locale and
   690   possibly adding declarations for computation on the coefficients.
   691   The method is already instantiated for natural numbers and for the
   692   axiomatic class of idoms with numerals.  See also the paper by
   693   Chaieb and Wenzel at CALCULEMUS 2007 for the general principles
   694   underlying this architecture of context-aware proof-tools.
   695 
   696 * constant "List.op @" now named "List.append".  Use ML antiquotations
   697 @{const_name List.append} or @{term " ... @ ... "} to circumvent
   698 possible incompatibilities when working on ML level.
   699 
   700 * Constant renames due to introduction of canonical name prefixing for
   701   class package:
   702 
   703     HOL.abs ~> HOL.minus_class.abs
   704     HOL.divide ~> HOL.divide_class.divide
   705     Nat.power ~> Nat.power_class.power
   706     Nat.size ~> Nat.size_class.size
   707     Numeral.number_of ~> Numeral.number_class.number_of
   708     FixedPoint.Inf ~> FixedPoint.complete_lattice_class.Inf
   709 
   710 * Rudimentary class target mechanism involves constant renames:
   711 
   712     Orderings.min ~> Orderings.ord_class.min
   713     Orderings.max ~> Orderings.ord_class.max
   714     FixedPoint.Sup ~> FixedPoint.complete_lattice_class.Sup
   715 
   716 * primrec: missing cases mapped to "undefined" instead of "arbitrary"
   717 
   718 * new constant "undefined" with axiom "undefined x = undefined"
   719 
   720 * new class "default" with associated constant "default"
   721 
   722 * new function listsum :: 'a list => 'a for arbitrary monoids.
   723   Special syntax: "SUM x <- xs. f x" (and latex variants)
   724 
   725 * new (input only) syntax for Haskell-like list comprehension, eg
   726   [(x,y). x <- xs, y <- ys, x ~= y]
   727   For details see List.thy.
   728 
   729 * The special syntax for function "filter" has changed from [x : xs. P] to
   730   [x <- xs. P] to avoid an ambiguity caused by list comprehension syntax,
   731   and for uniformity. INCOMPATIBILITY
   732 
   733 * New lemma collection field_simps (an extension of ring_simps)
   734   for manipulating (in)equations involving division. Multiplies
   735   with all denominators that can be proved to be non-zero (in equations)
   736   or positive/negative (in inequations).
   737 
   738 * Lemma collections ring_eq_simps, group_eq_simps and ring_distrib
   739   have been improved and renamed to ring_simps, group_simps and ring_distribs.
   740   Removed lemmas field_xyz in Ring_and_Field
   741   because they were subsumed by lemmas xyz.
   742 INCOMPATIBILITY.
   743 
   744 * Library/Pretty_Int.thy: maps HOL numerals on target language integer literals
   745   when generating code.
   746 
   747 * Library/Pretty_Char.thy: maps HOL characters on target language character literals
   748   when generating code.
   749 
   750 * Library/Commutative_Ring.thy: switched from recdef to function package;
   751   constants add, mul, pow now curried.  Infix syntax for algebraic operations.
   752 
   753 * Some steps towards more uniform lattice theory development in HOL.
   754 
   755     constants "meet" and "join" now named "inf" and "sup"
   756     constant "Meet" now named "Inf"
   757 
   758     classes "meet_semilorder" and "join_semilorder" now named
   759       "lower_semilattice" and "upper_semilattice"
   760     class "lorder" now named "lattice"
   761     class "comp_lat" now named "complete_lattice"
   762 
   763     Instantiation of lattice classes allows explicit definitions
   764     for "inf" and "sup" operations.
   765 
   766   INCOMPATIBILITY.  Theorem renames:
   767 
   768     meet_left_le            ~> inf_le1
   769     meet_right_le           ~> inf_le2
   770     join_left_le            ~> sup_ge1
   771     join_right_le           ~> sup_ge2
   772     meet_join_le            ~> inf_sup_ord
   773     le_meetI                ~> le_infI
   774     join_leI                ~> le_supI
   775     le_meet                 ~> le_inf_iff
   776     le_join                 ~> ge_sup_conv
   777     meet_idempotent         ~> inf_idem
   778     join_idempotent         ~> sup_idem
   779     meet_comm               ~> inf_commute
   780     join_comm               ~> sup_commute
   781     meet_leI1               ~> le_infI1
   782     meet_leI2               ~> le_infI2
   783     le_joinI1               ~> le_supI1
   784     le_joinI2               ~> le_supI2
   785     meet_assoc              ~> inf_assoc
   786     join_assoc              ~> sup_assoc
   787     meet_left_comm          ~> inf_left_commute
   788     meet_left_idempotent    ~> inf_left_idem
   789     join_left_comm          ~> sup_left_commute
   790     join_left_idempotent    ~> sup_left_idem
   791     meet_aci                ~> inf_aci
   792     join_aci                ~> sup_aci
   793     le_def_meet             ~> le_iff_inf
   794     le_def_join             ~> le_iff_sup
   795     join_absorp2            ~> sup_absorb2
   796     join_absorp1            ~> sup_absorb1
   797     meet_absorp1            ~> inf_absorb1
   798     meet_absorp2            ~> inf_absorb2
   799     meet_join_absorp        ~> inf_sup_absorb
   800     join_meet_absorp        ~> sup_inf_absorb
   801     distrib_join_le         ~> distrib_sup_le
   802     distrib_meet_le         ~> distrib_inf_le
   803 
   804     add_meet_distrib_left   ~> add_inf_distrib_left
   805     add_join_distrib_left   ~> add_sup_distrib_left
   806     is_join_neg_meet        ~> is_join_neg_inf
   807     is_meet_neg_join        ~> is_meet_neg_sup
   808     add_meet_distrib_right  ~> add_inf_distrib_right
   809     add_join_distrib_right  ~> add_sup_distrib_right
   810     add_meet_join_distribs  ~> add_sup_inf_distribs
   811     join_eq_neg_meet        ~> sup_eq_neg_inf
   812     meet_eq_neg_join        ~> inf_eq_neg_sup
   813     add_eq_meet_join        ~> add_eq_inf_sup
   814     meet_0_imp_0            ~> inf_0_imp_0
   815     join_0_imp_0            ~> sup_0_imp_0
   816     meet_0_eq_0             ~> inf_0_eq_0
   817     join_0_eq_0             ~> sup_0_eq_0
   818     neg_meet_eq_join        ~> neg_inf_eq_sup
   819     neg_join_eq_meet        ~> neg_sup_eq_inf
   820     join_eq_if              ~> sup_eq_if
   821 
   822     mono_meet               ~> mono_inf
   823     mono_join               ~> mono_sup
   824     meet_bool_eq            ~> inf_bool_eq
   825     join_bool_eq            ~> sup_bool_eq
   826     meet_fun_eq             ~> inf_fun_eq
   827     join_fun_eq             ~> sup_fun_eq
   828     meet_set_eq             ~> inf_set_eq
   829     join_set_eq             ~> sup_set_eq
   830     meet1_iff               ~> inf1_iff
   831     meet2_iff               ~> inf2_iff
   832     meet1I                  ~> inf1I
   833     meet2I                  ~> inf2I
   834     meet1D1                 ~> inf1D1
   835     meet2D1                 ~> inf2D1
   836     meet1D2                 ~> inf1D2
   837     meet2D2                 ~> inf2D2
   838     meet1E                  ~> inf1E
   839     meet2E                  ~> inf2E
   840     join1_iff               ~> sup1_iff
   841     join2_iff               ~> sup2_iff
   842     join1I1                 ~> sup1I1
   843     join2I1                 ~> sup2I1
   844     join1I1                 ~> sup1I1
   845     join2I2                 ~> sup1I2
   846     join1CI                 ~> sup1CI
   847     join2CI                 ~> sup2CI
   848     join1E                  ~> sup1E
   849     join2E                  ~> sup2E
   850 
   851     is_meet_Meet            ~> is_meet_Inf
   852     Meet_bool_def           ~> Inf_bool_def
   853     Meet_fun_def            ~> Inf_fun_def
   854     Meet_greatest           ~> Inf_greatest
   855     Meet_lower              ~> Inf_lower
   856     Meet_set_def            ~> Inf_set_def
   857 
   858     listsp_meetI            ~> listsp_infI
   859     listsp_meet_eq          ~> listsp_inf_eq
   860 
   861     meet_min                ~> inf_min
   862     join_max                ~> sup_max
   863 
   864 * Classes "order" and "linorder": facts "refl", "trans" and
   865 "cases" renamed ro "order_refl", "order_trans" and "linorder_cases", to
   866 avoid clashes with HOL "refl" and "trans". INCOMPATIBILITY.
   867 
   868 * Classes "order" and "linorder": 
   869 potential INCOMPATIBILITY: order of proof goals in order/linorder instance
   870 proofs changed.
   871 
   872 * Dropped lemma duplicate def_imp_eq in favor of meta_eq_to_obj_eq.
   873 INCOMPATIBILITY.
   874 
   875 * Dropped lemma duplicate if_def2 in favor of if_bool_eq_conj.
   876 INCOMPATIBILITY.
   877 
   878 * Added syntactic class "size"; overloaded constant "size" now has
   879 type "'a::size ==> bool"
   880 
   881 * Renamed constants "Divides.op div", "Divides.op mod" and "Divides.op
   882 dvd" to "Divides.div_class.div", "Divides.div_class.mod" and "Divides.dvd". INCOMPATIBILITY.
   883 
   884 * Added method "lexicographic_order" automatically synthesizes
   885 termination relations as lexicographic combinations of size measures
   886 -- 'function' package.
   887 
   888 * HOL/records: generalised field-update to take a function on the
   889 field rather than the new value: r(|A := x|) is translated to A_update
   890 (K x) r The K-combinator that is internally used is called K_record.
   891 INCOMPATIBILITY: Usage of the plain update functions has to be
   892 adapted.
   893  
   894 * axclass "semiring_0" now contains annihilation axioms x * 0 = 0 and
   895 0 * x = 0, which are required for a semiring.  Richer structures do
   896 not inherit from semiring_0 anymore, because this property is a
   897 theorem there, not an axiom.  INCOMPATIBILITY: In instances of
   898 semiring_0, there is more to prove, but this is mostly trivial.
   899 
   900 * axclass "recpower" was generalized to arbitrary monoids, not just
   901 commutative semirings.  INCOMPATIBILITY: If you use recpower and need
   902 commutativity or a semiring property, add the corresponding classes.
   903 
   904 * Unified locale partial_order with class definition (cf. theory
   905 Orderings), added parameter ``less''.  INCOMPATIBILITY.
   906 
   907 * Constant "List.list_all2" in List.thy now uses authentic syntax.
   908 INCOMPATIBILITY: translations containing list_all2 may go wrong.  On
   909 Isar level, use abbreviations instead.
   910 
   911 * Renamed constant "List.op mem" to "List.memberl" INCOMPATIBILITY:
   912 rarely occuring name references (e.g. ``List.op mem.simps'') require
   913 renaming (e.g. ``List.memberl.simps'').
   914 
   915 * Renamed constants "0" to "HOL.zero_class.zero" and "1" to "HOL.one_class.one".
   916 INCOMPATIBILITY.
   917 
   918 * Added class "HOL.eq", allowing for code generation with polymorphic equality.
   919 
   920 * Numeral syntax: type 'bin' which was a mere type copy of 'int' has
   921 been abandoned in favour of plain 'int'. INCOMPATIBILITY --
   922 significant changes for setting up numeral syntax for types:
   923 
   924   - new constants Numeral.pred and Numeral.succ instead
   925       of former Numeral.bin_pred and Numeral.bin_succ.
   926   - Use integer operations instead of bin_add, bin_mult and so on.
   927   - Numeral simplification theorems named Numeral.numeral_simps instead of Bin_simps.
   928   - ML structure Bin_Simprocs now named Int_Numeral_Base_Simprocs.
   929 
   930 See HOL/Integ/IntArith.thy for an example setup.
   931 
   932 * New top level command 'normal_form' computes the normal form of a
   933 term that may contain free variables. For example ``normal_form
   934 "rev[a,b,c]"'' produces ``[b,c,a]'' (without proof).  This command is
   935 suitable for heavy-duty computations because the functions are
   936 compiled to ML first.
   937 
   938 * Alternative iff syntax "A <-> B" for equality on bool (with priority
   939 25 like -->); output depends on the "iff" print_mode, the default is
   940 "A = B" (with priority 50).
   941 
   942 * Renamed constants in HOL.thy and Orderings.thy:
   943     op +   ~> HOL.plus_class.plus
   944     op -   ~> HOL.minus_class.minus
   945     uminus ~> HOL.minus_class.uminus
   946     op *   ~> HOL.times_class.times
   947     op <   ~> Orderings.ord_class.less
   948     op <=  ~> Orderings.ord_class.less_eq
   949 
   950 Adaptions may be required in the following cases:
   951 
   952 a) User-defined constants using any of the names "plus", "minus", "times",
   953 "less" or "less_eq". The standard syntax translations for "+", "-" and "*"
   954 may go wrong.
   955 INCOMPATIBILITY: use more specific names.
   956 
   957 b) Variables named "plus", "minus", "times", "less", "less_eq"
   958 INCOMPATIBILITY: use more specific names.
   959 
   960 c) Permutative equations (e.g. "a + b = b + a")
   961 Since the change of names also changes the order of terms, permutative
   962 rewrite rules may get applied in a different order. Experience shows that
   963 this is rarely the case (only two adaptions in the whole Isabelle
   964 distribution).
   965 INCOMPATIBILITY: rewrite proofs
   966 
   967 d) ML code directly refering to constant names
   968 This in general only affects hand-written proof tactics, simprocs and so on.
   969 INCOMPATIBILITY: grep your sourcecode and replace names.  Consider use
   970 of const_name ML antiquotations.
   971 
   972 * Relations less (<) and less_eq (<=) are also available on type bool.
   973 Modified syntax to disallow nesting without explicit parentheses,
   974 e.g. "(x < y) < z" or "x < (y < z)", but NOT "x < y < z".
   975 
   976 * "LEAST x:A. P" expands to "LEAST x. x:A & P" (input only).
   977 
   978 * Relation composition operator "op O" now has precedence 75 and binds
   979 stronger than union and intersection. INCOMPATIBILITY.
   980 
   981 * The old set interval syntax "{m..n(}" (and relatives) has been
   982 removed.  Use "{m..<n}" (and relatives) instead.
   983 
   984 * In the context of the assumption "~(s = t)" the Simplifier rewrites
   985 "t = s" to False (by simproc "neq_simproc").  For backward
   986 compatibility this can be disabled by ML "reset use_neq_simproc".
   987 
   988 * "m dvd n" where m and n are numbers is evaluated to True/False by
   989 simp.
   990 
   991 * Theorem Cons_eq_map_conv no longer declared as ``simp''.
   992 
   993 * Theorem setsum_mult renamed to setsum_right_distrib.
   994 
   995 * Prefer ex1I over ex_ex1I in single-step reasoning, e.g. by the
   996 ``rule'' method.
   997 
   998 * Reimplemented methods ``sat'' and ``satx'', with several
   999 improvements: goals no longer need to be stated as "<prems> ==>
  1000 False", equivalences (i.e. "=" on type bool) are handled, variable
  1001 names of the form "lit_<n>" are no longer reserved, significant
  1002 speedup.
  1003 
  1004 * Methods ``sat'' and ``satx'' can now replay MiniSat proof traces.
  1005 zChaff is still supported as well.
  1006 
  1007 * 'inductive' and 'datatype': provide projections of mutual rules,
  1008 bundled as foo_bar.inducts;
  1009 
  1010 * Library: moved theories Parity, GCD, Binomial, Infinite_Set to
  1011 Library.
  1012 
  1013 * Library: moved theory Accessible_Part to main HOL.
  1014 
  1015 * Library: added theory Coinductive_List of potentially infinite lists
  1016 as greatest fixed-point.
  1017 
  1018 * Library: added theory AssocList which implements (finite) maps as
  1019 association lists.
  1020 
  1021 * Added proof method ``evaluation'' for efficiently solving a goal
  1022 (i.e. a boolean expression) by compiling it to ML. The goal is
  1023 "proved" (via an oracle) if it evaluates to True.
  1024 
  1025 * Linear arithmetic now splits certain operators (e.g. min, max, abs)
  1026 also when invoked by the simplifier.  This results in the simplifier
  1027 being more powerful on arithmetic goals.  INCOMPATIBILITY.  Set
  1028 fast_arith_split_limit to 0 to obtain the old behavior.
  1029 
  1030 * Support for hex (0x20) and binary (0b1001) numerals.
  1031 
  1032 * New method: reify eqs (t), where eqs are equations for an
  1033 interpretation I :: 'a list => 'b => 'c and t::'c is an optional
  1034 parameter, computes a term s::'b and a list xs::'a list and proves the
  1035 theorem I xs s = t. This is also known as reification or quoting. The
  1036 resulting theorem is applied to the subgoal to substitute t with I xs
  1037 s.  If t is omitted, the subgoal itself is reified.
  1038 
  1039 * New method: reflection corr_thm eqs (t). The parameters eqs and (t)
  1040 are as explained above. corr_thm is a theorem for I vs (f t) = I vs t,
  1041 where f is supposed to be a computable function (in the sense of code
  1042 generattion). The method uses reify to compute s and xs as above then
  1043 applies corr_thm and uses normalization by evaluation to "prove" f s =
  1044 r and finally gets the theorem t = r, which is again applied to the
  1045 subgoal. An Example is available in HOL/ex/ReflectionEx.thy.
  1046 
  1047 * Reflection: Automatic refification now handels binding, an example
  1048 is available in HOL/ex/ReflectionEx.thy
  1049 
  1050 
  1051 *** HOL-Algebra ***
  1052 
  1053 * Formalisation of ideals and the quotient construction over rings.
  1054 
  1055 * Order and lattice theory no longer based on records.
  1056 INCOMPATIBILITY.
  1057 
  1058 * Renamed lemmas least_carrier -> least_closed and greatest_carrier ->
  1059 greatest_closed.  INCOMPATIBILITY.
  1060 
  1061 * Method algebra is now set up via an attribute.  For examples see
  1062 Ring.thy.  INCOMPATIBILITY: the method is now weaker on combinations
  1063 of algebraic structures.
  1064 
  1065 * Renamed theory CRing to Ring.
  1066 
  1067 
  1068 *** HOL-Complex ***
  1069 
  1070 * Theory Real: new method ferrack implements quantifier elimination
  1071 for linear arithmetic over the reals. The quantifier elimination
  1072 feature is used only for decision, for compatibility with arith. This
  1073 means a goal is either solved or left unchanged, no simplification.
  1074 
  1075 * Hyperreal: Functions root and sqrt are now defined on negative real
  1076 inputs so that root n (- x) = - root n x and sqrt (- x) = - sqrt x.
  1077 Nonnegativity side conditions have been removed from many lemmas, so
  1078 that more subgoals may now be solved by simplification; potential
  1079 INCOMPATIBILITY.
  1080 
  1081 * Real: New axiomatic classes formalize real normed vector spaces and
  1082 algebras, using new overloaded constants scaleR :: real => 'a => 'a
  1083 and norm :: 'a => real.
  1084 
  1085 * Real: New constant of_real :: real => 'a::real_algebra_1 injects
  1086 from reals into other types. The overloaded constant Reals :: 'a set
  1087 is now defined as range of_real; potential INCOMPATIBILITY.
  1088 
  1089 * Real: ML code generation is supported now and hence also quickcheck.
  1090 Reals are implemented as arbitrary precision rationals.
  1091 
  1092 * Hyperreal: Several constants that previously worked only for the
  1093 reals have been generalized, so they now work over arbitrary vector
  1094 spaces. Type annotations may need to be added in some cases; potential
  1095 INCOMPATIBILITY.
  1096 
  1097   Infinitesimal  :: ('a::real_normed_vector) star set
  1098   HFinite        :: ('a::real_normed_vector) star set
  1099   HInfinite      :: ('a::real_normed_vector) star set
  1100   approx         :: ('a::real_normed_vector) star => 'a star => bool
  1101   monad          :: ('a::real_normed_vector) star => 'a star set
  1102   galaxy         :: ('a::real_normed_vector) star => 'a star set
  1103   (NS)LIMSEQ     :: [nat => 'a::real_normed_vector, 'a] => bool
  1104   (NS)convergent :: (nat => 'a::real_normed_vector) => bool
  1105   (NS)Bseq       :: (nat => 'a::real_normed_vector) => bool
  1106   (NS)Cauchy     :: (nat => 'a::real_normed_vector) => bool
  1107   (NS)LIM        :: ['a::real_normed_vector => 'b::real_normed_vector, 'a, 'b] => bool
  1108   is(NS)Cont     :: ['a::real_normed_vector => 'b::real_normed_vector, 'a] => bool
  1109   deriv          :: ['a::real_normed_field => 'a, 'a, 'a] => bool
  1110   sgn            :: 'a::real_normed_vector => 'a
  1111   exp            :: 'a::{recpower,real_normed_field,banach} => 'a
  1112 
  1113 * Complex: Some complex-specific constants are now abbreviations for
  1114 overloaded ones: complex_of_real = of_real, cmod = norm, hcmod =
  1115 hnorm.  Other constants have been entirely removed in favor of the
  1116 polymorphic versions (INCOMPATIBILITY):
  1117 
  1118   approx        <-- capprox
  1119   HFinite       <-- CFinite
  1120   HInfinite     <-- CInfinite
  1121   Infinitesimal <-- CInfinitesimal
  1122   monad         <-- cmonad
  1123   galaxy        <-- cgalaxy
  1124   (NS)LIM       <-- (NS)CLIM, (NS)CRLIM
  1125   is(NS)Cont    <-- is(NS)Contc, is(NS)contCR
  1126   (ns)deriv     <-- (ns)cderiv
  1127 
  1128 
  1129 *** ML ***
  1130 
  1131 * Generic arithmetic modules: Tools/integer.ML, Tools/rat.ML, Tools/float.ML
  1132 
  1133 * Context data interfaces (Theory/Proof/GenericDataFun): removed
  1134 name/print, uninitialized data defaults to ad-hoc copy of empty value,
  1135 init only required for impure data.  INCOMPATIBILITY: empty really
  1136 need to be empty (no dependencies on theory content!)
  1137 
  1138 * ML within Isar: antiquotations allow to embed statically-checked
  1139 formal entities in the source, referring to the context available at
  1140 compile-time.  For example:
  1141 
  1142 ML {* @{typ "'a => 'b"} *}
  1143 ML {* @{term "%x. x"} *}
  1144 ML {* @{prop "x == y"} *}
  1145 ML {* @{ctyp "'a => 'b"} *}
  1146 ML {* @{cterm "%x. x"} *}
  1147 ML {* @{cprop "x == y"} *}
  1148 ML {* @{thm asm_rl} *}
  1149 ML {* @{thms asm_rl} *}
  1150 ML {* @{const_name c} *}
  1151 ML {* @{const_syntax c} *}
  1152 ML {* @{context} *}
  1153 ML {* @{theory} *}
  1154 ML {* @{theory Pure} *}
  1155 ML {* @{simpset} *}
  1156 ML {* @{claset} *}
  1157 ML {* @{clasimpset} *}
  1158 
  1159 The same works for sources being ``used'' within an Isar context.
  1160 
  1161 * ML in Isar: improved error reporting; extra verbosity with
  1162 Toplevel.debug enabled.
  1163 
  1164 * Pure/library:
  1165 
  1166   val burrow: ('a list -> 'b list) -> 'a list list -> 'b list list
  1167   val fold_burrow: ('a list -> 'c -> 'b list * 'd) -> 'a list list -> 'c -> 'b list list * 'd
  1168 
  1169 The semantics of "burrow" is: "take a function with *simulatanously*
  1170 transforms a list of value, and apply it *simulatanously* to a list of
  1171 list of values of the appropriate type". Compare this with "map" which
  1172 would *not* apply its argument function simulatanously but in
  1173 sequence; "fold_burrow" has an additional context.
  1174 
  1175 * Pure/library: functions map2 and fold2 with curried syntax for
  1176 simultanous mapping and folding:
  1177 
  1178     val map2: ('a -> 'b -> 'c) -> 'a list -> 'b list -> 'c list
  1179     val fold2: ('a -> 'b -> 'c -> 'c) -> 'a list -> 'b list -> 'c -> 'c
  1180 
  1181 * Pure/library: indexed lists - some functions in the Isabelle library
  1182 treating lists over 'a as finite mappings from [0...n] to 'a have been
  1183 given more convenient names and signatures reminiscent of similar
  1184 functions for alists, tables, etc:
  1185 
  1186   val nth: 'a list -> int -> 'a 
  1187   val nth_map: int -> ('a -> 'a) -> 'a list -> 'a list
  1188   val fold_index: (int * 'a -> 'b -> 'b) -> 'a list -> 'b -> 'b
  1189 
  1190 Note that fold_index starts counting at index 0, not 1 like foldln
  1191 used to.
  1192 
  1193 * Pure/library: added general ``divide_and_conquer'' combinator on
  1194 lists.
  1195 
  1196 * Pure/General/table.ML: the join operations now works via exceptions
  1197 DUP/SAME instead of type option.  This is simpler in simple cases, and
  1198 admits slightly more efficient complex applications.
  1199 
  1200 * Pure: datatype Context.generic joins theory/Proof.context and
  1201 provides some facilities for code that works in either kind of
  1202 context, notably GenericDataFun for uniform theory and proof data.
  1203 
  1204 * Pure: 'advanced' translation functions (parse_translation etc.) now
  1205 use Context.generic instead of just theory.
  1206 
  1207 * Pure: simplified internal attribute type, which is now always
  1208 Context.generic * thm -> Context.generic * thm.  Global (theory)
  1209 vs. local (Proof.context) attributes have been discontinued, while
  1210 minimizing code duplication.  Thm.rule_attribute and
  1211 Thm.declaration_attribute build canonical attributes; see also
  1212 structure Context for further operations on Context.generic, notably
  1213 GenericDataFun.  INCOMPATIBILITY, need to adapt attribute type
  1214 declarations and definitions.
  1215 
  1216 * Pure/kernel: consts certification ignores sort constraints given in
  1217 signature declarations.  (This information is not relevant to the
  1218 logic, but only for type inference.)  IMPORTANT INTERNAL CHANGE,
  1219 potential INCOMPATIBILITY.
  1220 
  1221 * Pure: axiomatic type classes are now purely definitional, with
  1222 explicit proofs of class axioms and super class relations performed
  1223 internally.  See Pure/axclass.ML for the main internal interfaces --
  1224 notably AxClass.define_class supercedes AxClass.add_axclass, and
  1225 AxClass.axiomatize_class/classrel/arity supercede
  1226 Sign.add_classes/classrel/arities.
  1227 
  1228 * Pure/Isar: Args/Attrib parsers operate on Context.generic --
  1229 global/local versions on theory vs. Proof.context have been
  1230 discontinued; Attrib.syntax and Method.syntax have been adapted
  1231 accordingly.  INCOMPATIBILITY, need to adapt parser expressions for
  1232 attributes, methods, etc.
  1233 
  1234 * Pure: several functions of signature "... -> theory -> theory * ..."
  1235 have been reoriented to "... -> theory -> ... * theory" in order to
  1236 allow natural usage in combination with the ||>, ||>>, |-> and
  1237 fold_map combinators.
  1238 
  1239 * Pure: official theorem names (closed derivations) and additional
  1240 comments (tags) are now strictly separate.  Name hints -- which are
  1241 maintained as tags -- may be attached any time without affecting the
  1242 derivation.
  1243 
  1244 * Pure: primitive rule lift_rule now takes goal cterm instead of an
  1245 actual goal state (thm).  Use Thm.lift_rule (Thm.cprem_of st i) to
  1246 achieve the old behaviour.
  1247 
  1248 * Pure: the "Goal" constant is now called "prop", supporting a
  1249 slightly more general idea of ``protecting'' meta-level rule
  1250 statements.
  1251 
  1252 * Pure: Logic.(un)varify only works in a global context, which is now
  1253 enforced instead of silently assumed.  INCOMPATIBILITY, may use
  1254 Logic.legacy_(un)varify as temporary workaround.
  1255 
  1256 * Pure: structure Name provides scalable operations for generating
  1257 internal variable names, notably Name.variants etc.  This replaces
  1258 some popular functions from term.ML:
  1259 
  1260   Term.variant		->  Name.variant
  1261   Term.variantlist	->  Name.variant_list  (*canonical argument order*)
  1262   Term.invent_names	->  Name.invent_list
  1263 
  1264 Note that low-level renaming rarely occurs in new code -- operations
  1265 from structure Variable are used instead (see below).
  1266 
  1267 * Pure: structure Variable provides fundamental operations for proper
  1268 treatment of fixed/schematic variables in a context.  For example,
  1269 Variable.import introduces fixes for schematics of given facts and
  1270 Variable.export reverses the effect (up to renaming) -- this replaces
  1271 various freeze_thaw operations.
  1272 
  1273 * Pure: structure Goal provides simple interfaces for
  1274 init/conclude/finish and tactical prove operations (replacing former
  1275 Tactic.prove).  Goal.prove is the canonical way to prove results
  1276 within a given context; Goal.prove_global is a degraded version for
  1277 theory level goals, including a global Drule.standard.  Note that
  1278 OldGoals.prove_goalw_cterm has long been obsolete, since it is
  1279 ill-behaved in a local proof context (e.g. with local fixes/assumes or
  1280 in a locale context).
  1281 
  1282 * Isar: simplified treatment of user-level errors, using exception
  1283 ERROR of string uniformly.  Function error now merely raises ERROR,
  1284 without any side effect on output channels.  The Isar toplevel takes
  1285 care of proper display of ERROR exceptions.  ML code may use plain
  1286 handle/can/try; cat_error may be used to concatenate errors like this:
  1287 
  1288   ... handle ERROR msg => cat_error msg "..."
  1289 
  1290 Toplevel ML code (run directly or through the Isar toplevel) may be
  1291 embedded into the Isar toplevel with exception display/debug like
  1292 this:
  1293 
  1294   Isar.toplevel (fn () => ...)
  1295 
  1296 INCOMPATIBILITY, removed special transform_error facilities, removed
  1297 obsolete variants of user-level exceptions (ERROR_MESSAGE,
  1298 Context.PROOF, ProofContext.CONTEXT, Proof.STATE, ProofHistory.FAIL)
  1299 -- use plain ERROR instead.
  1300 
  1301 * Isar: theory setup now has type (theory -> theory), instead of a
  1302 list.  INCOMPATIBILITY, may use #> to compose setup functions.
  1303 
  1304 * Isar: installed ML toplevel pretty printer for type Proof.context,
  1305 subject to ProofContext.debug/verbose flags.
  1306 
  1307 * Isar: Toplevel.theory_to_proof admits transactions that modify the
  1308 theory before entering a proof state.  Transactions now always see a
  1309 quasi-functional intermediate checkpoint, both in interactive and
  1310 batch mode.
  1311 
  1312 * Simplifier: the simpset of a running simplification process now
  1313 contains a proof context (cf. Simplifier.the_context), which is the
  1314 very context that the initial simpset has been retrieved from (by
  1315 simpset_of/local_simpset_of).  Consequently, all plug-in components
  1316 (solver, looper etc.) may depend on arbitrary proof data.
  1317 
  1318 * Simplifier.inherit_context inherits the proof context (plus the
  1319 local bounds) of the current simplification process; any simproc
  1320 etc. that calls the Simplifier recursively should do this!  Removed
  1321 former Simplifier.inherit_bounds, which is already included here --
  1322 INCOMPATIBILITY.  Tools based on low-level rewriting may even have to
  1323 specify an explicit context using Simplifier.context/theory_context.
  1324 
  1325 * Simplifier/Classical Reasoner: more abstract interfaces
  1326 change_simpset/claset for modifying the simpset/claset reference of a
  1327 theory; raw versions simpset/claset_ref etc. have been discontinued --
  1328 INCOMPATIBILITY.
  1329 
  1330 * Provers: more generic wrt. syntax of object-logics, avoid hardwired
  1331 "Trueprop" etc.
  1332 
  1333 
  1334 *** System ***
  1335 
  1336 * settings: ML_IDENTIFIER -- which is appended to user specific heap
  1337 locations -- now includes the Isabelle version identifier as well.
  1338 This simplifies use of multiple Isabelle installations.
  1339 
  1340 * isabelle-process: option -S (secure mode) disables some critical
  1341 operations, notably runtime compilation and evaluation of ML source
  1342 code.
  1343 
  1344 
  1345 New in Isabelle2005 (October 2005)
  1346 ----------------------------------
  1347 
  1348 *** General ***
  1349 
  1350 * Theory headers: the new header syntax for Isar theories is
  1351 
  1352   theory <name>
  1353   imports <theory1> ... <theoryN>
  1354   uses <file1> ... <fileM>
  1355   begin
  1356 
  1357 where the 'uses' part is optional.  The previous syntax
  1358 
  1359   theory <name> = <theory1> + ... + <theoryN>:
  1360 
  1361 will disappear in the next release.  Use isatool fixheaders to convert
  1362 existing theory files.  Note that there is no change in ancient
  1363 non-Isar theories now, but these will disappear soon.
  1364 
  1365 * Theory loader: parent theories can now also be referred to via
  1366 relative and absolute paths.
  1367 
  1368 * Command 'find_theorems' searches for a list of criteria instead of a
  1369 list of constants. Known criteria are: intro, elim, dest, name:string,
  1370 simp:term, and any term. Criteria can be preceded by '-' to select
  1371 theorems that do not match. Intro, elim, dest select theorems that
  1372 match the current goal, name:s selects theorems whose fully qualified
  1373 name contain s, and simp:term selects all simplification rules whose
  1374 lhs match term.  Any other term is interpreted as pattern and selects
  1375 all theorems matching the pattern. Available in ProofGeneral under
  1376 'ProofGeneral -> Find Theorems' or C-c C-f.  Example:
  1377 
  1378   C-c C-f (100) "(_::nat) + _ + _" intro -name: "HOL."
  1379 
  1380 prints the last 100 theorems matching the pattern "(_::nat) + _ + _",
  1381 matching the current goal as introduction rule and not having "HOL."
  1382 in their name (i.e. not being defined in theory HOL).
  1383 
  1384 * Command 'thms_containing' has been discontinued in favour of
  1385 'find_theorems'; INCOMPATIBILITY.
  1386 
  1387 * Communication with Proof General is now 8bit clean, which means that
  1388 Unicode text in UTF-8 encoding may be used within theory texts (both
  1389 formal and informal parts).  Cf. option -U of the Isabelle Proof
  1390 General interface.  Here are some simple examples (cf. src/HOL/ex):
  1391 
  1392   http://isabelle.in.tum.de/library/HOL/ex/Hebrew.html
  1393   http://isabelle.in.tum.de/library/HOL/ex/Chinese.html
  1394 
  1395 * Improved efficiency of the Simplifier and, to a lesser degree, the
  1396 Classical Reasoner.  Typical big applications run around 2 times
  1397 faster.
  1398 
  1399 
  1400 *** Document preparation ***
  1401 
  1402 * Commands 'display_drafts' and 'print_drafts' perform simple output
  1403 of raw sources.  Only those symbols that do not require additional
  1404 LaTeX packages (depending on comments in isabellesym.sty) are
  1405 displayed properly, everything else is left verbatim.  isatool display
  1406 and isatool print are used as front ends (these are subject to the
  1407 DVI/PDF_VIEWER and PRINT_COMMAND settings, respectively).
  1408 
  1409 * Command tags control specific markup of certain regions of text,
  1410 notably folding and hiding.  Predefined tags include "theory" (for
  1411 theory begin and end), "proof" for proof commands, and "ML" for
  1412 commands involving ML code; the additional tags "visible" and
  1413 "invisible" are unused by default.  Users may give explicit tag
  1414 specifications in the text, e.g. ''by %invisible (auto)''.  The
  1415 interpretation of tags is determined by the LaTeX job during document
  1416 preparation: see option -V of isatool usedir, or options -n and -t of
  1417 isatool document, or even the LaTeX macros \isakeeptag, \isafoldtag,
  1418 \isadroptag.
  1419 
  1420 Several document versions may be produced at the same time via isatool
  1421 usedir (the generated index.html will link all of them).  Typical
  1422 specifications include ''-V document=theory,proof,ML'' to present
  1423 theory/proof/ML parts faithfully, ''-V outline=/proof,/ML'' to fold
  1424 proof and ML commands, and ''-V mutilated=-theory,-proof,-ML'' to omit
  1425 these parts without any formal replacement text.  The Isabelle site
  1426 default settings produce ''document'' and ''outline'' versions as
  1427 specified above.
  1428 
  1429 * Several new antiquotations:
  1430 
  1431   @{term_type term} prints a term with its type annotated;
  1432 
  1433   @{typeof term} prints the type of a term;
  1434 
  1435   @{const const} is the same as @{term const}, but checks that the
  1436   argument is a known logical constant;
  1437 
  1438   @{term_style style term} and @{thm_style style thm} print a term or
  1439   theorem applying a "style" to it
  1440 
  1441   @{ML text}
  1442 
  1443 Predefined styles are 'lhs' and 'rhs' printing the lhs/rhs of
  1444 definitions, equations, inequations etc., 'concl' printing only the
  1445 conclusion of a meta-logical statement theorem, and 'prem1' .. 'prem19'
  1446 to print the specified premise.  TermStyle.add_style provides an ML
  1447 interface for introducing further styles.  See also the "LaTeX Sugar"
  1448 document practical applications.  The ML antiquotation prints
  1449 type-checked ML expressions verbatim.
  1450 
  1451 * Markup commands 'chapter', 'section', 'subsection', 'subsubsection',
  1452 and 'text' support optional locale specification '(in loc)', which
  1453 specifies the default context for interpreting antiquotations.  For
  1454 example: 'text (in lattice) {* @{thm inf_assoc}*}'.
  1455 
  1456 * Option 'locale=NAME' of antiquotations specifies an alternative
  1457 context interpreting the subsequent argument.  For example: @{thm
  1458 [locale=lattice] inf_assoc}.
  1459 
  1460 * Proper output of proof terms (@{prf ...} and @{full_prf ...}) within
  1461 a proof context.
  1462 
  1463 * Proper output of antiquotations for theory commands involving a
  1464 proof context (such as 'locale' or 'theorem (in loc) ...').
  1465 
  1466 * Delimiters of outer tokens (string etc.) now produce separate LaTeX
  1467 macros (\isachardoublequoteopen, isachardoublequoteclose etc.).
  1468 
  1469 * isatool usedir: new option -C (default true) controls whether option
  1470 -D should include a copy of the original document directory; -C false
  1471 prevents unwanted effects such as copying of administrative CVS data.
  1472 
  1473 
  1474 *** Pure ***
  1475 
  1476 * Considerably improved version of 'constdefs' command.  Now performs
  1477 automatic type-inference of declared constants; additional support for
  1478 local structure declarations (cf. locales and HOL records), see also
  1479 isar-ref manual.  Potential INCOMPATIBILITY: need to observe strictly
  1480 sequential dependencies of definitions within a single 'constdefs'
  1481 section; moreover, the declared name needs to be an identifier.  If
  1482 all fails, consider to fall back on 'consts' and 'defs' separately.
  1483 
  1484 * Improved indexed syntax and implicit structures.  First of all,
  1485 indexed syntax provides a notational device for subscripted
  1486 application, using the new syntax \<^bsub>term\<^esub> for arbitrary
  1487 expressions.  Secondly, in a local context with structure
  1488 declarations, number indexes \<^sub>n or the empty index (default
  1489 number 1) refer to a certain fixed variable implicitly; option
  1490 show_structs controls printing of implicit structures.  Typical
  1491 applications of these concepts involve record types and locales.
  1492 
  1493 * New command 'no_syntax' removes grammar declarations (and
  1494 translations) resulting from the given syntax specification, which is
  1495 interpreted in the same manner as for the 'syntax' command.
  1496 
  1497 * 'Advanced' translation functions (parse_translation etc.) may depend
  1498 on the signature of the theory context being presently used for
  1499 parsing/printing, see also isar-ref manual.
  1500 
  1501 * Improved 'oracle' command provides a type-safe interface to turn an
  1502 ML expression of type theory -> T -> term into a primitive rule of
  1503 type theory -> T -> thm (i.e. the functionality of Thm.invoke_oracle
  1504 is already included here); see also FOL/ex/IffExample.thy;
  1505 INCOMPATIBILITY.
  1506 
  1507 * axclass: name space prefix for class "c" is now "c_class" (was "c"
  1508 before); "cI" is no longer bound, use "c.intro" instead.
  1509 INCOMPATIBILITY.  This change avoids clashes of fact bindings for
  1510 axclasses vs. locales.
  1511 
  1512 * Improved internal renaming of symbolic identifiers -- attach primes
  1513 instead of base 26 numbers.
  1514 
  1515 * New flag show_question_marks controls printing of leading question
  1516 marks in schematic variable names.
  1517 
  1518 * In schematic variable names, *any* symbol following \<^isub> or
  1519 \<^isup> is now treated as part of the base name.  For example, the
  1520 following works without printing of awkward ".0" indexes:
  1521 
  1522   lemma "x\<^isub>1 = x\<^isub>2 ==> x\<^isub>2 = x\<^isub>1"
  1523     by simp
  1524 
  1525 * Inner syntax includes (*(*nested*) comments*).
  1526 
  1527 * Pretty printer now supports unbreakable blocks, specified in mixfix
  1528 annotations as "(00...)".
  1529 
  1530 * Clear separation of logical types and nonterminals, where the latter
  1531 may only occur in 'syntax' specifications or type abbreviations.
  1532 Before that distinction was only partially implemented via type class
  1533 "logic" vs. "{}".  Potential INCOMPATIBILITY in rare cases of improper
  1534 use of 'types'/'consts' instead of 'nonterminals'/'syntax'.  Some very
  1535 exotic syntax specifications may require further adaption
  1536 (e.g. Cube/Cube.thy).
  1537 
  1538 * Removed obsolete type class "logic", use the top sort {} instead.
  1539 Note that non-logical types should be declared as 'nonterminals'
  1540 rather than 'types'.  INCOMPATIBILITY for new object-logic
  1541 specifications.
  1542 
  1543 * Attributes 'induct' and 'cases': type or set names may now be
  1544 locally fixed variables as well.
  1545 
  1546 * Simplifier: can now control the depth to which conditional rewriting
  1547 is traced via the PG menu Isabelle -> Settings -> Trace Simp Depth
  1548 Limit.
  1549 
  1550 * Simplifier: simplification procedures may now take the current
  1551 simpset into account (cf. Simplifier.simproc(_i) / mk_simproc
  1552 interface), which is very useful for calling the Simplifier
  1553 recursively.  Minor INCOMPATIBILITY: the 'prems' argument of simprocs
  1554 is gone -- use prems_of_ss on the simpset instead.  Moreover, the
  1555 low-level mk_simproc no longer applies Logic.varify internally, to
  1556 allow for use in a context of fixed variables.
  1557 
  1558 * thin_tac now works even if the assumption being deleted contains !!
  1559 or ==>.  More generally, erule now works even if the major premise of
  1560 the elimination rule contains !! or ==>.
  1561 
  1562 * Method 'rules' has been renamed to 'iprover'. INCOMPATIBILITY.
  1563 
  1564 * Reorganized bootstrapping of the Pure theories; CPure is now derived
  1565 from Pure, which contains all common declarations already.  Both
  1566 theories are defined via plain Isabelle/Isar .thy files.
  1567 INCOMPATIBILITY: elements of CPure (such as the CPure.intro /
  1568 CPure.elim / CPure.dest attributes) now appear in the Pure name space;
  1569 use isatool fixcpure to adapt your theory and ML sources.
  1570 
  1571 * New syntax 'name(i-j, i-, i, ...)' for referring to specific
  1572 selections of theorems in named facts via index ranges.
  1573 
  1574 * 'print_theorems': in theory mode, really print the difference
  1575 wrt. the last state (works for interactive theory development only),
  1576 in proof mode print all local facts (cf. 'print_facts');
  1577 
  1578 * 'hide': option '(open)' hides only base names.
  1579 
  1580 * More efficient treatment of intermediate checkpoints in interactive
  1581 theory development.
  1582 
  1583 * Code generator is now invoked via code_module (incremental code
  1584 generation) and code_library (modular code generation, ML structures
  1585 for each theory).  INCOMPATIBILITY: new keywords 'file' and 'contains'
  1586 must be quoted when used as identifiers.
  1587 
  1588 * New 'value' command for reading, evaluating and printing terms using
  1589 the code generator.  INCOMPATIBILITY: command keyword 'value' must be
  1590 quoted when used as identifier.
  1591 
  1592 
  1593 *** Locales ***
  1594 
  1595 * New commands for the interpretation of locale expressions in
  1596 theories (1), locales (2) and proof contexts (3).  These generate
  1597 proof obligations from the expression specification.  After the
  1598 obligations have been discharged, theorems of the expression are added
  1599 to the theory, target locale or proof context.  The synopsis of the
  1600 commands is a follows:
  1601 
  1602   (1) interpretation expr inst
  1603   (2) interpretation target < expr
  1604   (3) interpret expr inst
  1605 
  1606 Interpretation in theories and proof contexts require a parameter
  1607 instantiation of terms from the current context.  This is applied to
  1608 specifications and theorems of the interpreted expression.
  1609 Interpretation in locales only permits parameter renaming through the
  1610 locale expression.  Interpretation is smart in that interpretations
  1611 that are active already do not occur in proof obligations, neither are
  1612 instantiated theorems stored in duplicate.  Use 'print_interps' to
  1613 inspect active interpretations of a particular locale.  For details,
  1614 see the Isar Reference manual.  Examples can be found in
  1615 HOL/Finite_Set.thy and HOL/Algebra/UnivPoly.thy.
  1616 
  1617 INCOMPATIBILITY: former 'instantiate' has been withdrawn, use
  1618 'interpret' instead.
  1619 
  1620 * New context element 'constrains' for adding type constraints to
  1621 parameters.
  1622 
  1623 * Context expressions: renaming of parameters with syntax
  1624 redeclaration.
  1625 
  1626 * Locale declaration: 'includes' disallowed.
  1627 
  1628 * Proper static binding of attribute syntax -- i.e. types / terms /
  1629 facts mentioned as arguments are always those of the locale definition
  1630 context, independently of the context of later invocations.  Moreover,
  1631 locale operations (renaming and type / term instantiation) are applied
  1632 to attribute arguments as expected.
  1633 
  1634 INCOMPATIBILITY of the ML interface: always pass Attrib.src instead of
  1635 actual attributes; rare situations may require Attrib.attribute to
  1636 embed those attributes into Attrib.src that lack concrete syntax.
  1637 Attribute implementations need to cooperate properly with the static
  1638 binding mechanism.  Basic parsers Args.XXX_typ/term/prop and
  1639 Attrib.XXX_thm etc. already do the right thing without further
  1640 intervention.  Only unusual applications -- such as "where" or "of"
  1641 (cf. src/Pure/Isar/attrib.ML), which process arguments depending both
  1642 on the context and the facts involved -- may have to assign parsed
  1643 values to argument tokens explicitly.
  1644 
  1645 * Changed parameter management in theorem generation for long goal
  1646 statements with 'includes'.  INCOMPATIBILITY: produces a different
  1647 theorem statement in rare situations.
  1648 
  1649 * Locale inspection command 'print_locale' omits notes elements.  Use
  1650 'print_locale!' to have them included in the output.
  1651 
  1652 
  1653 *** Provers ***
  1654 
  1655 * Provers/hypsubst.ML: improved version of the subst method, for
  1656 single-step rewriting: it now works in bound variable contexts. New is
  1657 'subst (asm)', for rewriting an assumption.  INCOMPATIBILITY: may
  1658 rewrite a different subterm than the original subst method, which is
  1659 still available as 'simplesubst'.
  1660 
  1661 * Provers/quasi.ML: new transitivity reasoners for transitivity only
  1662 and quasi orders.
  1663 
  1664 * Provers/trancl.ML: new transitivity reasoner for transitive and
  1665 reflexive-transitive closure of relations.
  1666 
  1667 * Provers/blast.ML: new reference depth_limit to make blast's depth
  1668 limit (previously hard-coded with a value of 20) user-definable.
  1669 
  1670 * Provers/simplifier.ML has been moved to Pure, where Simplifier.setup
  1671 is peformed already.  Object-logics merely need to finish their
  1672 initial simpset configuration as before.  INCOMPATIBILITY.
  1673 
  1674 
  1675 *** HOL ***
  1676 
  1677 * Symbolic syntax of Hilbert Choice Operator is now as follows:
  1678 
  1679   syntax (epsilon)
  1680     "_Eps" :: "[pttrn, bool] => 'a"    ("(3\<some>_./ _)" [0, 10] 10)
  1681 
  1682 The symbol \<some> is displayed as the alternative epsilon of LaTeX
  1683 and x-symbol; use option '-m epsilon' to get it actually printed.
  1684 Moreover, the mathematically important symbolic identifier \<epsilon>
  1685 becomes available as variable, constant etc.  INCOMPATIBILITY,
  1686 
  1687 * "x > y" abbreviates "y < x" and "x >= y" abbreviates "y <= x".
  1688 Similarly for all quantifiers: "ALL x > y" etc.  The x-symbol for >=
  1689 is \<ge>. New transitivity rules have been added to HOL/Orderings.thy to
  1690 support corresponding Isar calculations.
  1691 
  1692 * "{x:A. P}" abbreviates "{x. x:A & P}", and similarly for "\<in>"
  1693 instead of ":".
  1694 
  1695 * theory SetInterval: changed the syntax for open intervals:
  1696 
  1697   Old       New
  1698   {..n(}    {..<n}
  1699   {)n..}    {n<..}
  1700   {m..n(}   {m..<n}
  1701   {)m..n}   {m<..n}
  1702   {)m..n(}  {m<..<n}
  1703 
  1704 The old syntax is still supported but will disappear in the next
  1705 release.  For conversion use the following Emacs search and replace
  1706 patterns (these are not perfect but work quite well):
  1707 
  1708   {)\([^\.]*\)\.\.  ->  {\1<\.\.}
  1709   \.\.\([^(}]*\)(}  ->  \.\.<\1}
  1710 
  1711 * Theory Commutative_Ring (in Library): method comm_ring for proving
  1712 equalities in commutative rings; method 'algebra' provides a generic
  1713 interface.
  1714 
  1715 * Theory Finite_Set: changed the syntax for 'setsum', summation over
  1716 finite sets: "setsum (%x. e) A", which used to be "\<Sum>x:A. e", is
  1717 now either "SUM x:A. e" or "\<Sum>x \<in> A. e". The bound variable can
  1718 be a tuple pattern.
  1719 
  1720 Some new syntax forms are available:
  1721 
  1722   "\<Sum>x | P. e"      for     "setsum (%x. e) {x. P}"
  1723   "\<Sum>x = a..b. e"   for     "setsum (%x. e) {a..b}"
  1724   "\<Sum>x = a..<b. e"  for     "setsum (%x. e) {a..<b}"
  1725   "\<Sum>x < k. e"      for     "setsum (%x. e) {..<k}"
  1726 
  1727 The latter form "\<Sum>x < k. e" used to be based on a separate
  1728 function "Summation", which has been discontinued.
  1729 
  1730 * theory Finite_Set: in structured induction proofs, the insert case
  1731 is now 'case (insert x F)' instead of the old counterintuitive 'case
  1732 (insert F x)'.
  1733 
  1734 * The 'refute' command has been extended to support a much larger
  1735 fragment of HOL, including axiomatic type classes, constdefs and
  1736 typedefs, inductive datatypes and recursion.
  1737 
  1738 * New tactics 'sat' and 'satx' to prove propositional tautologies.
  1739 Requires zChaff with proof generation to be installed.  See
  1740 HOL/ex/SAT_Examples.thy for examples.
  1741 
  1742 * Datatype induction via method 'induct' now preserves the name of the
  1743 induction variable. For example, when proving P(xs::'a list) by
  1744 induction on xs, the induction step is now P(xs) ==> P(a#xs) rather
  1745 than P(list) ==> P(a#list) as previously.  Potential INCOMPATIBILITY
  1746 in unstructured proof scripts.
  1747 
  1748 * Reworked implementation of records.  Improved scalability for
  1749 records with many fields, avoiding performance problems for type
  1750 inference. Records are no longer composed of nested field types, but
  1751 of nested extension types. Therefore the record type only grows linear
  1752 in the number of extensions and not in the number of fields.  The
  1753 top-level (users) view on records is preserved.  Potential
  1754 INCOMPATIBILITY only in strange cases, where the theory depends on the
  1755 old record representation. The type generated for a record is called
  1756 <record_name>_ext_type.
  1757 
  1758 Flag record_quick_and_dirty_sensitive can be enabled to skip the
  1759 proofs triggered by a record definition or a simproc (if
  1760 quick_and_dirty is enabled).  Definitions of large records can take
  1761 quite long.
  1762 
  1763 New simproc record_upd_simproc for simplification of multiple record
  1764 updates enabled by default.  Moreover, trivial updates are also
  1765 removed: r(|x := x r|) = r.  INCOMPATIBILITY: old proofs break
  1766 occasionally, since simplification is more powerful by default.
  1767 
  1768 * typedef: proper support for polymorphic sets, which contain extra
  1769 type-variables in the term.
  1770 
  1771 * Simplifier: automatically reasons about transitivity chains
  1772 involving "trancl" (r^+) and "rtrancl" (r^*) by setting up tactics
  1773 provided by Provers/trancl.ML as additional solvers.  INCOMPATIBILITY:
  1774 old proofs break occasionally as simplification may now solve more
  1775 goals than previously.
  1776 
  1777 * Simplifier: converts x <= y into x = y if assumption y <= x is
  1778 present.  Works for all partial orders (class "order"), in particular
  1779 numbers and sets.  For linear orders (e.g. numbers) it treats ~ x < y
  1780 just like y <= x.
  1781 
  1782 * Simplifier: new simproc for "let x = a in f x".  If a is a free or
  1783 bound variable or a constant then the let is unfolded.  Otherwise
  1784 first a is simplified to b, and then f b is simplified to g. If
  1785 possible we abstract b from g arriving at "let x = b in h x",
  1786 otherwise we unfold the let and arrive at g.  The simproc can be
  1787 enabled/disabled by the reference use_let_simproc.  Potential
  1788 INCOMPATIBILITY since simplification is more powerful by default.
  1789 
  1790 * Classical reasoning: the meson method now accepts theorems as arguments.
  1791 
  1792 * Prover support: pre-release of the Isabelle-ATP linkup, which runs background
  1793 jobs to provide advice on the provability of subgoals.
  1794 
  1795 * Theory OrderedGroup and Ring_and_Field: various additions and
  1796 improvements to faciliate calculations involving equalities and
  1797 inequalities.
  1798 
  1799 The following theorems have been eliminated or modified
  1800 (INCOMPATIBILITY):
  1801 
  1802   abs_eq             now named abs_of_nonneg
  1803   abs_of_ge_0        now named abs_of_nonneg
  1804   abs_minus_eq       now named abs_of_nonpos
  1805   imp_abs_id         now named abs_of_nonneg
  1806   imp_abs_neg_id     now named abs_of_nonpos
  1807   mult_pos           now named mult_pos_pos
  1808   mult_pos_le        now named mult_nonneg_nonneg
  1809   mult_pos_neg_le    now named mult_nonneg_nonpos
  1810   mult_pos_neg2_le   now named mult_nonneg_nonpos2
  1811   mult_neg           now named mult_neg_neg
  1812   mult_neg_le        now named mult_nonpos_nonpos
  1813 
  1814 * The following lemmas in Ring_and_Field have been added to the simplifier:
  1815      
  1816      zero_le_square
  1817      not_square_less_zero 
  1818 
  1819   The following lemmas have been deleted from Real/RealPow:
  1820   
  1821      realpow_zero_zero
  1822      realpow_two
  1823      realpow_less
  1824      zero_le_power
  1825      realpow_two_le
  1826      abs_realpow_two
  1827      realpow_two_abs     
  1828 
  1829 * Theory Parity: added rules for simplifying exponents.
  1830 
  1831 * Theory List:
  1832 
  1833 The following theorems have been eliminated or modified
  1834 (INCOMPATIBILITY):
  1835 
  1836   list_all_Nil       now named list_all.simps(1)
  1837   list_all_Cons      now named list_all.simps(2)
  1838   list_all_conv      now named list_all_iff
  1839   set_mem_eq         now named mem_iff
  1840 
  1841 * Theories SetsAndFunctions and BigO (see HOL/Library) support
  1842 asymptotic "big O" calculations.  See the notes in BigO.thy.
  1843 
  1844 
  1845 *** HOL-Complex ***
  1846 
  1847 * Theory RealDef: better support for embedding natural numbers and
  1848 integers in the reals.
  1849 
  1850 The following theorems have been eliminated or modified
  1851 (INCOMPATIBILITY):
  1852 
  1853   exp_ge_add_one_self  now requires no hypotheses
  1854   real_of_int_add      reversed direction of equality (use [symmetric])
  1855   real_of_int_minus    reversed direction of equality (use [symmetric])
  1856   real_of_int_diff     reversed direction of equality (use [symmetric])
  1857   real_of_int_mult     reversed direction of equality (use [symmetric])
  1858 
  1859 * Theory RComplete: expanded support for floor and ceiling functions.
  1860 
  1861 * Theory Ln is new, with properties of the natural logarithm
  1862 
  1863 * Hyperreal: There is a new type constructor "star" for making
  1864 nonstandard types.  The old type names are now type synonyms:
  1865 
  1866   hypreal = real star
  1867   hypnat = nat star
  1868   hcomplex = complex star
  1869 
  1870 * Hyperreal: Many groups of similarly-defined constants have been
  1871 replaced by polymorphic versions (INCOMPATIBILITY):
  1872 
  1873   star_of <-- hypreal_of_real, hypnat_of_nat, hcomplex_of_complex
  1874 
  1875   starset      <-- starsetNat, starsetC
  1876   *s*          <-- *sNat*, *sc*
  1877   starset_n    <-- starsetNat_n, starsetC_n
  1878   *sn*         <-- *sNatn*, *scn*
  1879   InternalSets <-- InternalNatSets, InternalCSets
  1880 
  1881   starfun      <-- starfun{Nat,Nat2,C,RC,CR}
  1882   *f*          <-- *fNat*, *fNat2*, *fc*, *fRc*, *fcR*
  1883   starfun_n    <-- starfun{Nat,Nat2,C,RC,CR}_n
  1884   *fn*         <-- *fNatn*, *fNat2n*, *fcn*, *fRcn*, *fcRn*
  1885   InternalFuns <-- InternalNatFuns, InternalNatFuns2, Internal{C,RC,CR}Funs
  1886 
  1887 * Hyperreal: Many type-specific theorems have been removed in favor of
  1888 theorems specific to various axiomatic type classes (INCOMPATIBILITY):
  1889 
  1890   add_commute <-- {hypreal,hypnat,hcomplex}_add_commute
  1891   add_assoc   <-- {hypreal,hypnat,hcomplex}_add_assocs
  1892   OrderedGroup.add_0 <-- {hypreal,hypnat,hcomplex}_add_zero_left
  1893   OrderedGroup.add_0_right <-- {hypreal,hcomplex}_add_zero_right
  1894   right_minus <-- hypreal_add_minus
  1895   left_minus <-- {hypreal,hcomplex}_add_minus_left
  1896   mult_commute <-- {hypreal,hypnat,hcomplex}_mult_commute
  1897   mult_assoc <-- {hypreal,hypnat,hcomplex}_mult_assoc
  1898   mult_1_left <-- {hypreal,hypnat}_mult_1, hcomplex_mult_one_left
  1899   mult_1_right <-- hcomplex_mult_one_right
  1900   mult_zero_left <-- hcomplex_mult_zero_left
  1901   left_distrib <-- {hypreal,hypnat,hcomplex}_add_mult_distrib
  1902   right_distrib <-- hypnat_add_mult_distrib2
  1903   zero_neq_one <-- {hypreal,hypnat,hcomplex}_zero_not_eq_one
  1904   right_inverse <-- hypreal_mult_inverse
  1905   left_inverse <-- hypreal_mult_inverse_left, hcomplex_mult_inv_left
  1906   order_refl <-- {hypreal,hypnat}_le_refl
  1907   order_trans <-- {hypreal,hypnat}_le_trans
  1908   order_antisym <-- {hypreal,hypnat}_le_anti_sym
  1909   order_less_le <-- {hypreal,hypnat}_less_le
  1910   linorder_linear <-- {hypreal,hypnat}_le_linear
  1911   add_left_mono <-- {hypreal,hypnat}_add_left_mono
  1912   mult_strict_left_mono <-- {hypreal,hypnat}_mult_less_mono2
  1913   add_nonneg_nonneg <-- hypreal_le_add_order
  1914 
  1915 * Hyperreal: Separate theorems having to do with type-specific
  1916 versions of constants have been merged into theorems that apply to the
  1917 new polymorphic constants (INCOMPATIBILITY):
  1918 
  1919   STAR_UNIV_set <-- {STAR_real,NatStar_real,STARC_complex}_set
  1920   STAR_empty_set <-- {STAR,NatStar,STARC}_empty_set
  1921   STAR_Un <-- {STAR,NatStar,STARC}_Un
  1922   STAR_Int <-- {STAR,NatStar,STARC}_Int
  1923   STAR_Compl <-- {STAR,NatStar,STARC}_Compl
  1924   STAR_subset <-- {STAR,NatStar,STARC}_subset
  1925   STAR_mem <-- {STAR,NatStar,STARC}_mem
  1926   STAR_mem_Compl <-- {STAR,STARC}_mem_Compl
  1927   STAR_diff <-- {STAR,STARC}_diff
  1928   STAR_star_of_image_subset <-- {STAR_hypreal_of_real, NatStar_hypreal_of_real,
  1929     STARC_hcomplex_of_complex}_image_subset
  1930   starset_n_Un <-- starset{Nat,C}_n_Un
  1931   starset_n_Int <-- starset{Nat,C}_n_Int
  1932   starset_n_Compl <-- starset{Nat,C}_n_Compl
  1933   starset_n_diff <-- starset{Nat,C}_n_diff
  1934   InternalSets_Un <-- Internal{Nat,C}Sets_Un
  1935   InternalSets_Int <-- Internal{Nat,C}Sets_Int
  1936   InternalSets_Compl <-- Internal{Nat,C}Sets_Compl
  1937   InternalSets_diff <-- Internal{Nat,C}Sets_diff
  1938   InternalSets_UNIV_diff <-- Internal{Nat,C}Sets_UNIV_diff
  1939   InternalSets_starset_n <-- Internal{Nat,C}Sets_starset{Nat,C}_n
  1940   starset_starset_n_eq <-- starset{Nat,C}_starset{Nat,C}_n_eq
  1941   starset_n_starset <-- starset{Nat,C}_n_starset{Nat,C}
  1942   starfun_n_starfun <-- starfun{Nat,Nat2,C,RC,CR}_n_starfun{Nat,Nat2,C,RC,CR}
  1943   starfun <-- starfun{Nat,Nat2,C,RC,CR}
  1944   starfun_mult <-- starfun{Nat,Nat2,C,RC,CR}_mult
  1945   starfun_add <-- starfun{Nat,Nat2,C,RC,CR}_add
  1946   starfun_minus <-- starfun{Nat,Nat2,C,RC,CR}_minus
  1947   starfun_diff <-- starfun{C,RC,CR}_diff
  1948   starfun_o <-- starfun{NatNat2,Nat2,_stafunNat,C,C_starfunRC,_starfunCR}_o
  1949   starfun_o2 <-- starfun{NatNat2,_stafunNat,C,C_starfunRC,_starfunCR}_o2
  1950   starfun_const_fun <-- starfun{Nat,Nat2,C,RC,CR}_const_fun
  1951   starfun_inverse <-- starfun{Nat,C,RC,CR}_inverse
  1952   starfun_eq <-- starfun{Nat,Nat2,C,RC,CR}_eq
  1953   starfun_eq_iff <-- starfun{C,RC,CR}_eq_iff
  1954   starfun_Id <-- starfunC_Id
  1955   starfun_approx <-- starfun{Nat,CR}_approx
  1956   starfun_capprox <-- starfun{C,RC}_capprox
  1957   starfun_abs <-- starfunNat_rabs
  1958   starfun_lambda_cancel <-- starfun{C,CR,RC}_lambda_cancel
  1959   starfun_lambda_cancel2 <-- starfun{C,CR,RC}_lambda_cancel2
  1960   starfun_mult_HFinite_approx <-- starfunCR_mult_HFinite_capprox
  1961   starfun_mult_CFinite_capprox <-- starfun{C,RC}_mult_CFinite_capprox
  1962   starfun_add_capprox <-- starfun{C,RC}_add_capprox
  1963   starfun_add_approx <-- starfunCR_add_approx
  1964   starfun_inverse_inverse <-- starfunC_inverse_inverse
  1965   starfun_divide <-- starfun{C,CR,RC}_divide
  1966   starfun_n <-- starfun{Nat,C}_n
  1967   starfun_n_mult <-- starfun{Nat,C}_n_mult
  1968   starfun_n_add <-- starfun{Nat,C}_n_add
  1969   starfun_n_add_minus <-- starfunNat_n_add_minus
  1970   starfun_n_const_fun <-- starfun{Nat,C}_n_const_fun
  1971   starfun_n_minus <-- starfun{Nat,C}_n_minus
  1972   starfun_n_eq <-- starfun{Nat,C}_n_eq
  1973 
  1974   star_n_add <-- {hypreal,hypnat,hcomplex}_add
  1975   star_n_minus <-- {hypreal,hcomplex}_minus
  1976   star_n_diff <-- {hypreal,hcomplex}_diff
  1977   star_n_mult <-- {hypreal,hcomplex}_mult
  1978   star_n_inverse <-- {hypreal,hcomplex}_inverse
  1979   star_n_le <-- {hypreal,hypnat}_le
  1980   star_n_less <-- {hypreal,hypnat}_less
  1981   star_n_zero_num <-- {hypreal,hypnat,hcomplex}_zero_num
  1982   star_n_one_num <-- {hypreal,hypnat,hcomplex}_one_num
  1983   star_n_abs <-- hypreal_hrabs
  1984   star_n_divide <-- hcomplex_divide
  1985 
  1986   star_of_add <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_add
  1987   star_of_minus <-- {hypreal_of_real,hcomplex_of_complex}_minus
  1988   star_of_diff <-- hypreal_of_real_diff
  1989   star_of_mult <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_mult
  1990   star_of_one <-- {hypreal_of_real,hcomplex_of_complex}_one
  1991   star_of_zero <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_zero
  1992   star_of_le <-- {hypreal_of_real,hypnat_of_nat}_le_iff
  1993   star_of_less <-- {hypreal_of_real,hypnat_of_nat}_less_iff
  1994   star_of_eq <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_eq_iff
  1995   star_of_inverse <-- {hypreal_of_real,hcomplex_of_complex}_inverse
  1996   star_of_divide <-- {hypreal_of_real,hcomplex_of_complex}_divide
  1997   star_of_of_nat <-- {hypreal_of_real,hcomplex_of_complex}_of_nat
  1998   star_of_of_int <-- {hypreal_of_real,hcomplex_of_complex}_of_int
  1999   star_of_number_of <-- {hypreal,hcomplex}_number_of
  2000   star_of_number_less <-- number_of_less_hypreal_of_real_iff
  2001   star_of_number_le <-- number_of_le_hypreal_of_real_iff
  2002   star_of_eq_number <-- hypreal_of_real_eq_number_of_iff
  2003   star_of_less_number <-- hypreal_of_real_less_number_of_iff
  2004   star_of_le_number <-- hypreal_of_real_le_number_of_iff
  2005   star_of_power <-- hypreal_of_real_power
  2006   star_of_eq_0 <-- hcomplex_of_complex_zero_iff
  2007 
  2008 * Hyperreal: new method "transfer" that implements the transfer
  2009 principle of nonstandard analysis. With a subgoal that mentions
  2010 nonstandard types like "'a star", the command "apply transfer"
  2011 replaces it with an equivalent one that mentions only standard types.
  2012 To be successful, all free variables must have standard types; non-
  2013 standard variables must have explicit universal quantifiers.
  2014 
  2015 * Hyperreal: A theory of Taylor series.
  2016 
  2017 
  2018 *** HOLCF ***
  2019 
  2020 * Discontinued special version of 'constdefs' (which used to support
  2021 continuous functions) in favor of the general Pure one with full
  2022 type-inference.
  2023 
  2024 * New simplification procedure for solving continuity conditions; it
  2025 is much faster on terms with many nested lambda abstractions (cubic
  2026 instead of exponential time).
  2027 
  2028 * New syntax for domain package: selector names are now optional.
  2029 Parentheses should be omitted unless argument is lazy, for example:
  2030 
  2031   domain 'a stream = cons "'a" (lazy "'a stream")
  2032 
  2033 * New command 'fixrec' for defining recursive functions with pattern
  2034 matching; defining multiple functions with mutual recursion is also
  2035 supported.  Patterns may include the constants cpair, spair, up, sinl,
  2036 sinr, or any data constructor defined by the domain package. The given
  2037 equations are proven as rewrite rules. See HOLCF/ex/Fixrec_ex.thy for
  2038 syntax and examples.
  2039 
  2040 * New commands 'cpodef' and 'pcpodef' for defining predicate subtypes
  2041 of cpo and pcpo types. Syntax is exactly like the 'typedef' command,
  2042 but the proof obligation additionally includes an admissibility
  2043 requirement. The packages generate instances of class cpo or pcpo,
  2044 with continuity and strictness theorems for Rep and Abs.
  2045 
  2046 * HOLCF: Many theorems have been renamed according to a more standard naming
  2047 scheme (INCOMPATIBILITY):
  2048 
  2049   foo_inject:  "foo$x = foo$y ==> x = y"
  2050   foo_eq:      "(foo$x = foo$y) = (x = y)"
  2051   foo_less:    "(foo$x << foo$y) = (x << y)"
  2052   foo_strict:  "foo$UU = UU"
  2053   foo_defined: "... ==> foo$x ~= UU"
  2054   foo_defined_iff: "(foo$x = UU) = (x = UU)"
  2055 
  2056 
  2057 *** ZF ***
  2058 
  2059 * ZF/ex: theories Group and Ring provide examples in abstract algebra,
  2060 including the First Isomorphism Theorem (on quotienting by the kernel
  2061 of a homomorphism).
  2062 
  2063 * ZF/Simplifier: install second copy of type solver that actually
  2064 makes use of TC rules declared to Isar proof contexts (or locales);
  2065 the old version is still required for ML proof scripts.
  2066 
  2067 
  2068 *** Cube ***
  2069 
  2070 * Converted to Isar theory format; use locales instead of axiomatic
  2071 theories.
  2072 
  2073 
  2074 *** ML ***
  2075 
  2076 * Pure/library.ML: added ##>, ##>>, #>> -- higher-order counterparts
  2077 for ||>, ||>>, |>>,
  2078 
  2079 * Pure/library.ML no longer defines its own option datatype, but uses
  2080 that of the SML basis, which has constructors NONE and SOME instead of
  2081 None and Some, as well as exception Option.Option instead of OPTION.
  2082 The functions the, if_none, is_some, is_none have been adapted
  2083 accordingly, while Option.map replaces apsome.
  2084 
  2085 * Pure/library.ML: the exception LIST has been given up in favour of
  2086 the standard exceptions Empty and Subscript, as well as
  2087 Library.UnequalLengths.  Function like Library.hd and Library.tl are
  2088 superceded by the standard hd and tl functions etc.
  2089 
  2090 A number of basic list functions are no longer exported to the ML
  2091 toplevel, as they are variants of predefined functions.  The following
  2092 suggests how one can translate existing code:
  2093 
  2094     rev_append xs ys = List.revAppend (xs, ys)
  2095     nth_elem (i, xs) = List.nth (xs, i)
  2096     last_elem xs = List.last xs
  2097     flat xss = List.concat xss
  2098     seq fs = List.app fs
  2099     partition P xs = List.partition P xs
  2100     mapfilter f xs = List.mapPartial f xs
  2101 
  2102 * Pure/library.ML: several combinators for linear functional
  2103 transformations, notably reverse application and composition:
  2104 
  2105   x |> f                f #> g
  2106   (x, y) |-> f          f #-> g
  2107 
  2108 * Pure/library.ML: introduced/changed precedence of infix operators:
  2109 
  2110   infix 1 |> |-> ||> ||>> |>> |>>> #> #->;
  2111   infix 2 ?;
  2112   infix 3 o oo ooo oooo;
  2113   infix 4 ~~ upto downto;
  2114 
  2115 Maybe INCOMPATIBILITY when any of those is used in conjunction with other
  2116 infix operators.
  2117 
  2118 * Pure/library.ML: natural list combinators fold, fold_rev, and
  2119 fold_map support linear functional transformations and nesting.  For
  2120 example:
  2121 
  2122   fold f [x1, ..., xN] y =
  2123     y |> f x1 |> ... |> f xN
  2124 
  2125   (fold o fold) f [xs1, ..., xsN] y =
  2126     y |> fold f xs1 |> ... |> fold f xsN
  2127 
  2128   fold f [x1, ..., xN] =
  2129     f x1 #> ... #> f xN
  2130 
  2131   (fold o fold) f [xs1, ..., xsN] =
  2132     fold f xs1 #> ... #> fold f xsN
  2133 
  2134 * Pure/library.ML: the following selectors on type 'a option are
  2135 available:
  2136 
  2137   the:               'a option -> 'a  (*partial*)
  2138   these:             'a option -> 'a  where 'a = 'b list
  2139   the_default: 'a -> 'a option -> 'a
  2140   the_list:          'a option -> 'a list
  2141 
  2142 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  2143 basic operations for association lists, following natural argument
  2144 order; moreover the explicit equality predicate passed here avoids
  2145 potentially expensive polymorphic runtime equality checks.
  2146 The old functions may be expressed as follows:
  2147 
  2148   assoc = uncurry (AList.lookup (op =))
  2149   assocs = these oo AList.lookup (op =)
  2150   overwrite = uncurry (AList.update (op =)) o swap
  2151 
  2152 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  2153 
  2154   val make: ('a -> 'b) -> 'a list -> ('a * 'b) list
  2155   val find: ('a * 'b -> bool) -> ('c * 'b) list -> 'a -> 'c list
  2156 
  2157 replacing make_keylist and keyfilter (occassionally used)
  2158 Naive rewrites:
  2159 
  2160   make_keylist = AList.make
  2161   keyfilter = AList.find (op =)
  2162 
  2163 * eq_fst and eq_snd now take explicit equality parameter, thus
  2164   avoiding eqtypes. Naive rewrites:
  2165 
  2166     eq_fst = eq_fst (op =)
  2167     eq_snd = eq_snd (op =)
  2168 
  2169 * Removed deprecated apl and apr (rarely used).
  2170   Naive rewrites:
  2171 
  2172     apl (n, op) =>>= curry op n
  2173     apr (op, m) =>>= fn n => op (n, m)
  2174 
  2175 * Pure/General: structure OrdList (cf. Pure/General/ord_list.ML)
  2176 provides a reasonably efficient light-weight implementation of sets as
  2177 lists.
  2178 
  2179 * Pure/General: generic tables (cf. Pure/General/table.ML) provide a
  2180 few new operations; existing lookup and update are now curried to
  2181 follow natural argument order (for use with fold etc.);
  2182 INCOMPATIBILITY, use (uncurry Symtab.lookup) etc. as last resort.
  2183 
  2184 * Pure/General: output via the Isabelle channels of
  2185 writeln/warning/error etc. is now passed through Output.output, with a
  2186 hook for arbitrary transformations depending on the print_mode
  2187 (cf. Output.add_mode -- the first active mode that provides a output
  2188 function wins).  Already formatted output may be embedded into further
  2189 text via Output.raw; the result of Pretty.string_of/str_of and derived
  2190 functions (string_of_term/cterm/thm etc.) is already marked raw to
  2191 accommodate easy composition of diagnostic messages etc.  Programmers
  2192 rarely need to care about Output.output or Output.raw at all, with
  2193 some notable exceptions: Output.output is required when bypassing the
  2194 standard channels (writeln etc.), or in token translations to produce
  2195 properly formatted results; Output.raw is required when capturing
  2196 already output material that will eventually be presented to the user
  2197 a second time.  For the default print mode, both Output.output and
  2198 Output.raw have no effect.
  2199 
  2200 * Pure/General: Output.time_accumulator NAME creates an operator ('a
  2201 -> 'b) -> 'a -> 'b to measure runtime and count invocations; the
  2202 cumulative results are displayed at the end of a batch session.
  2203 
  2204 * Pure/General: File.sysify_path and File.quote_sysify path have been
  2205 replaced by File.platform_path and File.shell_path (with appropriate
  2206 hooks).  This provides a clean interface for unusual systems where the
  2207 internal and external process view of file names are different.
  2208 
  2209 * Pure: more efficient orders for basic syntactic entities: added
  2210 fast_string_ord, fast_indexname_ord, fast_term_ord; changed sort_ord
  2211 and typ_ord to use fast_string_ord and fast_indexname_ord (term_ord is
  2212 NOT affected); structures Symtab, Vartab, Typtab, Termtab use the fast
  2213 orders now -- potential INCOMPATIBILITY for code that depends on a
  2214 particular order for Symtab.keys, Symtab.dest, etc. (consider using
  2215 Library.sort_strings on result).
  2216 
  2217 * Pure/term.ML: combinators fold_atyps, fold_aterms, fold_term_types,
  2218 fold_types traverse types/terms from left to right, observing natural
  2219 argument order.  Supercedes previous foldl_XXX versions, add_frees,
  2220 add_vars etc. have been adapted as well: INCOMPATIBILITY.
  2221 
  2222 * Pure: name spaces have been refined, with significant changes of the
  2223 internal interfaces -- INCOMPATIBILITY.  Renamed cond_extern(_table)
  2224 to extern(_table).  The plain name entry path is superceded by a
  2225 general 'naming' context, which also includes the 'policy' to produce
  2226 a fully qualified name and external accesses of a fully qualified
  2227 name; NameSpace.extend is superceded by context dependent
  2228 Sign.declare_name.  Several theory and proof context operations modify
  2229 the naming context.  Especially note Theory.restore_naming and
  2230 ProofContext.restore_naming to get back to a sane state; note that
  2231 Theory.add_path is no longer sufficient to recover from
  2232 Theory.absolute_path in particular.
  2233 
  2234 * Pure: new flags short_names (default false) and unique_names
  2235 (default true) for controlling output of qualified names.  If
  2236 short_names is set, names are printed unqualified.  If unique_names is
  2237 reset, the name prefix is reduced to the minimum required to achieve
  2238 the original result when interning again, even if there is an overlap
  2239 with earlier declarations.
  2240 
  2241 * Pure/TheoryDataFun: change of the argument structure; 'prep_ext' is
  2242 now 'extend', and 'merge' gets an additional Pretty.pp argument
  2243 (useful for printing error messages).  INCOMPATIBILITY.
  2244 
  2245 * Pure: major reorganization of the theory context.  Type Sign.sg and
  2246 Theory.theory are now identified, referring to the universal
  2247 Context.theory (see Pure/context.ML).  Actual signature and theory
  2248 content is managed as theory data.  The old code and interfaces were
  2249 spread over many files and structures; the new arrangement introduces
  2250 considerable INCOMPATIBILITY to gain more clarity:
  2251 
  2252   Context -- theory management operations (name, identity, inclusion,
  2253     parents, ancestors, merge, etc.), plus generic theory data;
  2254 
  2255   Sign -- logical signature and syntax operations (declaring consts,
  2256     types, etc.), plus certify/read for common entities;
  2257 
  2258   Theory -- logical theory operations (stating axioms, definitions,
  2259     oracles), plus a copy of logical signature operations (consts,
  2260     types, etc.); also a few basic management operations (Theory.copy,
  2261     Theory.merge, etc.)
  2262 
  2263 The most basic sign_of operations (Theory.sign_of, Thm.sign_of_thm
  2264 etc.) as well as the sign field in Thm.rep_thm etc. have been retained
  2265 for convenience -- they merely return the theory.
  2266 
  2267 * Pure: type Type.tsig is superceded by theory in most interfaces.
  2268 
  2269 * Pure: the Isar proof context type is already defined early in Pure
  2270 as Context.proof (note that ProofContext.context and Proof.context are
  2271 aliases, where the latter is the preferred name).  This enables other
  2272 Isabelle components to refer to that type even before Isar is present.
  2273 
  2274 * Pure/sign/theory: discontinued named name spaces (i.e. classK,
  2275 typeK, constK, axiomK, oracleK), but provide explicit operations for
  2276 any of these kinds.  For example, Sign.intern typeK is now
  2277 Sign.intern_type, Theory.hide_space Sign.typeK is now
  2278 Theory.hide_types.  Also note that former
  2279 Theory.hide_classes/types/consts are now
  2280 Theory.hide_classes_i/types_i/consts_i, while the non '_i' versions
  2281 internalize their arguments!  INCOMPATIBILITY.
  2282 
  2283 * Pure: get_thm interface (of PureThy and ProofContext) expects
  2284 datatype thmref (with constructors Name and NameSelection) instead of
  2285 plain string -- INCOMPATIBILITY;
  2286 
  2287 * Pure: cases produced by proof methods specify options, where NONE
  2288 means to remove case bindings -- INCOMPATIBILITY in
  2289 (RAW_)METHOD_CASES.
  2290 
  2291 * Pure: the following operations retrieve axioms or theorems from a
  2292 theory node or theory hierarchy, respectively:
  2293 
  2294   Theory.axioms_of: theory -> (string * term) list
  2295   Theory.all_axioms_of: theory -> (string * term) list
  2296   PureThy.thms_of: theory -> (string * thm) list
  2297   PureThy.all_thms_of: theory -> (string * thm) list
  2298 
  2299 * Pure: print_tac now outputs the goal through the trace channel.
  2300 
  2301 * Isar toplevel: improved diagnostics, mostly for Poly/ML only.
  2302 Reference Toplevel.debug (default false) controls detailed printing
  2303 and tracing of low-level exceptions; Toplevel.profiling (default 0)
  2304 controls execution profiling -- set to 1 for time and 2 for space
  2305 (both increase the runtime).
  2306 
  2307 * Isar session: The initial use of ROOT.ML is now always timed,
  2308 i.e. the log will show the actual process times, in contrast to the
  2309 elapsed wall-clock time that the outer shell wrapper produces.
  2310 
  2311 * Simplifier: improved handling of bound variables (nameless
  2312 representation, avoid allocating new strings).  Simprocs that invoke
  2313 the Simplifier recursively should use Simplifier.inherit_bounds to
  2314 avoid local name clashes.  Failure to do so produces warnings
  2315 "Simplifier: renamed bound variable ..."; set Simplifier.debug_bounds
  2316 for further details.
  2317 
  2318 * ML functions legacy_bindings and use_legacy_bindings produce ML fact
  2319 bindings for all theorems stored within a given theory; this may help
  2320 in porting non-Isar theories to Isar ones, while keeping ML proof
  2321 scripts for the time being.
  2322 
  2323 * ML operator HTML.with_charset specifies the charset begin used for
  2324 generated HTML files.  For example:
  2325 
  2326   HTML.with_charset "utf-8" use_thy "Hebrew";
  2327   HTML.with_charset "utf-8" use_thy "Chinese";
  2328 
  2329 
  2330 *** System ***
  2331 
  2332 * Allow symlinks to all proper Isabelle executables (Isabelle,
  2333 isabelle, isatool etc.).
  2334 
  2335 * ISABELLE_DOC_FORMAT setting specifies preferred document format (for
  2336 isatool doc, isatool mkdir, display_drafts etc.).
  2337 
  2338 * isatool usedir: option -f allows specification of the ML file to be
  2339 used by Isabelle; default is ROOT.ML.
  2340 
  2341 * New isatool version outputs the version identifier of the Isabelle
  2342 distribution being used.
  2343 
  2344 * HOL: new isatool dimacs2hol converts files in DIMACS CNF format
  2345 (containing Boolean satisfiability problems) into Isabelle/HOL
  2346 theories.
  2347 
  2348 
  2349 
  2350 New in Isabelle2004 (April 2004)
  2351 --------------------------------
  2352 
  2353 *** General ***
  2354 
  2355 * Provers/order.ML:  new efficient reasoner for partial and linear orders.
  2356   Replaces linorder.ML.
  2357 
  2358 * Pure: Greek letters (except small lambda, \<lambda>), as well as Gothic
  2359   (\<aa>...\<zz>\<AA>...\<ZZ>), calligraphic (\<A>...\<Z>), and Euler
  2360   (\<a>...\<z>), are now considered normal letters, and can therefore
  2361   be used anywhere where an ASCII letter (a...zA...Z) has until
  2362   now. COMPATIBILITY: This obviously changes the parsing of some
  2363   terms, especially where a symbol has been used as a binder, say
  2364   '\<Pi>x. ...', which is now a type error since \<Pi>x will be parsed
  2365   as an identifier.  Fix it by inserting a space around former
  2366   symbols.  Call 'isatool fixgreek' to try to fix parsing errors in
  2367   existing theory and ML files.
  2368 
  2369 * Pure: Macintosh and Windows line-breaks are now allowed in theory files.
  2370 
  2371 * Pure: single letter sub/superscripts (\<^isub> and \<^isup>) are now
  2372   allowed in identifiers. Similar to Greek letters \<^isub> is now considered
  2373   a normal (but invisible) letter. For multiple letter subscripts repeat
  2374   \<^isub> like this: x\<^isub>1\<^isub>2.
  2375 
  2376 * Pure: There are now sub-/superscripts that can span more than one
  2377   character. Text between \<^bsub> and \<^esub> is set in subscript in
  2378   ProofGeneral and LaTeX, text between \<^bsup> and \<^esup> in
  2379   superscript. The new control characters are not identifier parts.
  2380 
  2381 * Pure: Control-symbols of the form \<^raw:...> will literally print the
  2382   content of "..." to the latex file instead of \isacntrl... . The "..."
  2383   may consist of any printable characters excluding the end bracket >.
  2384 
  2385 * Pure: Using new Isar command "finalconsts" (or the ML functions
  2386   Theory.add_finals or Theory.add_finals_i) it is now possible to
  2387   declare constants "final", which prevents their being given a definition
  2388   later.  It is useful for constants whose behaviour is fixed axiomatically
  2389   rather than definitionally, such as the meta-logic connectives.
  2390 
  2391 * Pure: 'instance' now handles general arities with general sorts
  2392   (i.e. intersections of classes),
  2393 
  2394 * Presentation: generated HTML now uses a CSS style sheet to make layout
  2395   (somewhat) independent of content. It is copied from lib/html/isabelle.css.
  2396   It can be changed to alter the colors/layout of generated pages.
  2397 
  2398 
  2399 *** Isar ***
  2400 
  2401 * Tactic emulation methods rule_tac, erule_tac, drule_tac, frule_tac,
  2402   cut_tac, subgoal_tac and thin_tac:
  2403   - Now understand static (Isar) contexts.  As a consequence, users of Isar
  2404     locales are no longer forced to write Isar proof scripts.
  2405     For details see Isar Reference Manual, paragraph 4.3.2: Further tactic
  2406     emulations.
  2407   - INCOMPATIBILITY: names of variables to be instantiated may no
  2408     longer be enclosed in quotes.  Instead, precede variable name with `?'.
  2409     This is consistent with the instantiation attribute "where".
  2410 
  2411 * Attributes "where" and "of":
  2412   - Now take type variables of instantiated theorem into account when reading
  2413     the instantiation string.  This fixes a bug that caused instantiated
  2414     theorems to have too special types in some circumstances.
  2415   - "where" permits explicit instantiations of type variables.
  2416 
  2417 * Calculation commands "moreover" and "also" no longer interfere with
  2418   current facts ("this"), admitting arbitrary combinations with "then"
  2419   and derived forms.
  2420 
  2421 * Locales:
  2422   - Goal statements involving the context element "includes" no longer
  2423     generate theorems with internal delta predicates (those ending on
  2424     "_axioms") in the premise.
  2425     Resolve particular premise with <locale>.intro to obtain old form.
  2426   - Fixed bug in type inference ("unify_frozen") that prevented mix of target
  2427     specification and "includes" elements in goal statement.
  2428   - Rule sets <locale>.intro and <locale>.axioms no longer declared as
  2429     [intro?] and [elim?] (respectively) by default.
  2430   - Experimental command for instantiation of locales in proof contexts:
  2431         instantiate <label>[<attrs>]: <loc>
  2432     Instantiates locale <loc> and adds all its theorems to the current context
  2433     taking into account their attributes.  Label and attrs are optional
  2434     modifiers, like in theorem declarations.  If present, names of
  2435     instantiated theorems are qualified with <label>, and the attributes
  2436     <attrs> are applied after any attributes these theorems might have already.
  2437       If the locale has assumptions, a chained fact of the form
  2438     "<loc> t1 ... tn" is expected from which instantiations of the parameters
  2439     are derived.  The command does not support old-style locales declared
  2440     with "locale (open)".
  2441       A few (very simple) examples can be found in FOL/ex/LocaleInst.thy.
  2442 
  2443 * HOL: Tactic emulation methods induct_tac and case_tac understand static
  2444   (Isar) contexts.
  2445 
  2446 
  2447 *** HOL ***
  2448 
  2449 * Proof import: new image HOL4 contains the imported library from
  2450   the HOL4 system with about 2500 theorems. It is imported by
  2451   replaying proof terms produced by HOL4 in Isabelle. The HOL4 image
  2452   can be used like any other Isabelle image.  See
  2453   HOL/Import/HOL/README for more information.
  2454 
  2455 * Simplifier:
  2456   - Much improved handling of linear and partial orders.
  2457     Reasoners for linear and partial orders are set up for type classes
  2458     "linorder" and "order" respectively, and are added to the default simpset
  2459     as solvers.  This means that the simplifier can build transitivity chains
  2460     to solve goals from the assumptions.
  2461   - INCOMPATIBILITY: old proofs break occasionally.  Typically, applications
  2462     of blast or auto after simplification become unnecessary because the goal
  2463     is solved by simplification already.
  2464 
  2465 * Numerics: new theory Ring_and_Field contains over 250 basic numerical laws,
  2466     all proved in axiomatic type classes for semirings, rings and fields.
  2467 
  2468 * Numerics:
  2469   - Numeric types (nat, int, and in HOL-Complex rat, real, complex, etc.) are
  2470     now formalized using the Ring_and_Field theory mentioned above.
  2471   - INCOMPATIBILITY: simplification and arithmetic behaves somewhat differently
  2472     than before, because now they are set up once in a generic manner.
  2473   - INCOMPATIBILITY: many type-specific arithmetic laws have gone.
  2474     Look for the general versions in Ring_and_Field (and Power if they concern
  2475     exponentiation).
  2476 
  2477 * Type "rat" of the rational numbers is now available in HOL-Complex.
  2478 
  2479 * Records:
  2480   - Record types are now by default printed with their type abbreviation
  2481     instead of the list of all field types. This can be configured via
  2482     the reference "print_record_type_abbr".
  2483   - Simproc "record_upd_simproc" for simplification of multiple updates added
  2484     (not enabled by default).
  2485   - Simproc "record_ex_sel_eq_simproc" to simplify EX x. sel r = x resp.
  2486     EX x. x = sel r to True (not enabled by default).
  2487   - Tactic "record_split_simp_tac" to split and simplify records added.
  2488 
  2489 * 'specification' command added, allowing for definition by
  2490   specification.  There is also an 'ax_specification' command that
  2491   introduces the new constants axiomatically.
  2492 
  2493 * arith(_tac) is now able to generate counterexamples for reals as well.
  2494 
  2495 * HOL-Algebra: new locale "ring" for non-commutative rings.
  2496 
  2497 * HOL-ex: InductiveInvariant_examples illustrates advanced recursive function
  2498   definitions, thanks to Sava Krsti\'{c} and John Matthews.
  2499 
  2500 * HOL-Matrix: a first theory for matrices in HOL with an application of
  2501   matrix theory to linear programming.
  2502 
  2503 * Unions and Intersections:
  2504   The latex output syntax of UN and INT has been changed
  2505   from "\Union x \in A. B" to "\Union_{x \in A} B"
  2506   i.e. the index formulae has become a subscript.
  2507   Similarly for "\Union x. B", and for \Inter instead of \Union.
  2508 
  2509 * Unions and Intersections over Intervals:
  2510   There is new short syntax "UN i<=n. A" for "UN i:{0..n}. A". There is
  2511   also an x-symbol version with subscripts "\<Union>\<^bsub>i <= n\<^esub>. A"
  2512   like in normal math, and corresponding versions for < and for intersection.
  2513 
  2514 * HOL/List: Ordering "lexico" is renamed "lenlex" and the standard
  2515   lexicographic dictonary ordering has been added as "lexord".
  2516 
  2517 * ML: the legacy theory structures Int and List have been removed. They had
  2518   conflicted with ML Basis Library structures having the same names.
  2519 
  2520 * 'refute' command added to search for (finite) countermodels.  Only works
  2521   for a fragment of HOL.  The installation of an external SAT solver is
  2522   highly recommended.  See "HOL/Refute.thy" for details.
  2523 
  2524 * 'quickcheck' command: Allows to find counterexamples by evaluating
  2525   formulae under an assignment of free variables to random values.
  2526   In contrast to 'refute', it can deal with inductive datatypes,
  2527   but cannot handle quantifiers. See "HOL/ex/Quickcheck_Examples.thy"
  2528   for examples.
  2529 
  2530 
  2531 *** HOLCF ***
  2532 
  2533 * Streams now come with concatenation and are part of the HOLCF image
  2534 
  2535 
  2536 
  2537 New in Isabelle2003 (May 2003)
  2538 ------------------------------
  2539 
  2540 *** General ***
  2541 
  2542 * Provers/simplifier:
  2543 
  2544   - Completely reimplemented method simp (ML: Asm_full_simp_tac):
  2545     Assumptions are now subject to complete mutual simplification,
  2546     not just from left to right. The simplifier now preserves
  2547     the order of assumptions.
  2548 
  2549     Potential INCOMPATIBILITY:
  2550 
  2551     -- simp sometimes diverges where the old version did
  2552        not, e.g. invoking simp on the goal
  2553 
  2554         [| P (f x); y = x; f x = f y |] ==> Q
  2555 
  2556        now gives rise to the infinite reduction sequence
  2557 
  2558         P(f x) --(f x = f y)--> P(f y) --(y = x)--> P(f x) --(f x = f y)--> ...
  2559 
  2560        Using "simp (asm_lr)" (ML: Asm_lr_simp_tac) instead often solves this
  2561        kind of problem.
  2562 
  2563     -- Tactics combining classical reasoner and simplification (such as auto)
  2564        are also affected by this change, because many of them rely on
  2565        simp. They may sometimes diverge as well or yield a different numbers
  2566        of subgoals. Try to use e.g. force, fastsimp, or safe instead of auto
  2567        in case of problems. Sometimes subsequent calls to the classical
  2568        reasoner will fail because a preceeding call to the simplifier too
  2569        eagerly simplified the goal, e.g. deleted redundant premises.
  2570 
  2571   - The simplifier trace now shows the names of the applied rewrite rules
  2572 
  2573   - You can limit the number of recursive invocations of the simplifier
  2574     during conditional rewriting (where the simplifie tries to solve the
  2575     conditions before applying the rewrite rule):
  2576     ML "simp_depth_limit := n"
  2577     where n is an integer. Thus you can force termination where previously
  2578     the simplifier would diverge.
  2579 
  2580   - Accepts free variables as head terms in congruence rules.  Useful in Isar.
  2581 
  2582   - No longer aborts on failed congruence proof.  Instead, the
  2583     congruence is ignored.
  2584 
  2585 * Pure: New generic framework for extracting programs from constructive
  2586   proofs. See HOL/Extraction.thy for an example instantiation, as well
  2587   as HOL/Extraction for some case studies.
  2588 
  2589 * Pure: The main goal of the proof state is no longer shown by default, only
  2590 the subgoals. This behaviour is controlled by a new flag.
  2591    PG menu: Isabelle/Isar -> Settings -> Show Main Goal
  2592 (ML: Proof.show_main_goal).
  2593 
  2594 * Pure: You can find all matching introduction rules for subgoal 1, i.e. all
  2595 rules whose conclusion matches subgoal 1:
  2596       PG menu: Isabelle/Isar -> Show me -> matching rules
  2597 The rules are ordered by how closely they match the subgoal.
  2598 In particular, rules that solve a subgoal outright are displayed first
  2599 (or rather last, the way they are printed).
  2600 (ML: ProofGeneral.print_intros())
  2601 
  2602 * Pure: New flag trace_unify_fail causes unification to print
  2603 diagnostic information (PG: in trace buffer) when it fails. This is
  2604 useful for figuring out why single step proofs like rule, erule or
  2605 assumption failed.
  2606 
  2607 * Pure: Locale specifications now produce predicate definitions
  2608 according to the body of text (covering assumptions modulo local
  2609 definitions); predicate "loc_axioms" covers newly introduced text,
  2610 while "loc" is cumulative wrt. all included locale expressions; the
  2611 latter view is presented only on export into the global theory
  2612 context; potential INCOMPATIBILITY, use "(open)" option to fall back
  2613 on the old view without predicates;
  2614 
  2615 * Pure: predefined locales "var" and "struct" are useful for sharing
  2616 parameters (as in CASL, for example); just specify something like
  2617 ``var x + var y + struct M'' as import;
  2618 
  2619 * Pure: improved thms_containing: proper indexing of facts instead of
  2620 raw theorems; check validity of results wrt. current name space;
  2621 include local facts of proof configuration (also covers active
  2622 locales), cover fixed variables in index; may use "_" in term
  2623 specification; an optional limit for the number of printed facts may
  2624 be given (the default is 40);
  2625 
  2626 * Pure: disallow duplicate fact bindings within new-style theory files
  2627 (batch-mode only);
  2628 
  2629 * Provers: improved induct method: assumptions introduced by case
  2630 "foo" are split into "foo.hyps" (from the rule) and "foo.prems" (from
  2631 the goal statement); "foo" still refers to all facts collectively;
  2632 
  2633 * Provers: the function blast.overloaded has been removed: all constants
  2634 are regarded as potentially overloaded, which improves robustness in exchange
  2635 for slight decrease in efficiency;
  2636 
  2637 * Provers/linorder: New generic prover for transitivity reasoning over
  2638 linear orders.  Note: this prover is not efficient!
  2639 
  2640 * Isar: preview of problems to finish 'show' now produce an error
  2641 rather than just a warning (in interactive mode);
  2642 
  2643 
  2644 *** HOL ***
  2645 
  2646 * arith(_tac)
  2647 
  2648  - Produces a counter example if it cannot prove a goal.
  2649    Note that the counter example may be spurious if the goal is not a formula
  2650    of quantifier-free linear arithmetic.
  2651    In ProofGeneral the counter example appears in the trace buffer.
  2652 
  2653  - Knows about div k and mod k where k is a numeral of type nat or int.
  2654 
  2655  - Calls full Presburger arithmetic (by Amine Chaieb) if quantifier-free
  2656    linear arithmetic fails. This takes account of quantifiers and divisibility.
  2657    Presburger arithmetic can also be called explicitly via presburger(_tac).
  2658 
  2659 * simp's arithmetic capabilities have been enhanced a bit: it now
  2660 takes ~= in premises into account (by performing a case split);
  2661 
  2662 * simp reduces "m*(n div m) + n mod m" to n, even if the two summands
  2663 are distributed over a sum of terms;
  2664 
  2665 * New tactic "trans_tac" and method "trans" instantiate
  2666 Provers/linorder.ML for axclasses "order" and "linorder" (predicates
  2667 "<=", "<" and "=").
  2668 
  2669 * function INCOMPATIBILITIES: Pi-sets have been redefined and moved from main
  2670 HOL to Library/FuncSet; constant "Fun.op o" is now called "Fun.comp";
  2671 
  2672 * 'typedef' command has new option "open" to suppress the set
  2673 definition;
  2674 
  2675 * functions Min and Max on finite sets have been introduced (theory
  2676 Finite_Set);
  2677 
  2678 * attribute [symmetric] now works for relations as well; it turns
  2679 (x,y) : R^-1 into (y,x) : R, and vice versa;
  2680 
  2681 * induct over a !!-quantified statement (say !!x1..xn):
  2682   each "case" automatically performs "fix x1 .. xn" with exactly those names.
  2683 
  2684 * Map: `empty' is no longer a constant but a syntactic abbreviation for
  2685 %x. None. Warning: empty_def now refers to the previously hidden definition
  2686 of the empty set.
  2687 
  2688 * Algebra: formalization of classical algebra.  Intended as base for
  2689 any algebraic development in Isabelle.  Currently covers group theory
  2690 (up to Sylow's theorem) and ring theory (Universal Property of
  2691 Univariate Polynomials).  Contributions welcome;
  2692 
  2693 * GroupTheory: deleted, since its material has been moved to Algebra;
  2694 
  2695 * Complex: new directory of the complex numbers with numeric constants,
  2696 nonstandard complex numbers, and some complex analysis, standard and
  2697 nonstandard (Jacques Fleuriot);
  2698 
  2699 * HOL-Complex: new image for analysis, replacing HOL-Real and HOL-Hyperreal;
  2700 
  2701 * Hyperreal: introduced Gauge integration and hyperreal logarithms (Jacques
  2702 Fleuriot);
  2703 
  2704 * Real/HahnBanach: updated and adapted to locales;
  2705 
  2706 * NumberTheory: added Gauss's law of quadratic reciprocity (by Avigad,
  2707 Gray and Kramer);
  2708 
  2709 * UNITY: added the Meier-Sanders theory of progress sets;
  2710 
  2711 * MicroJava: bytecode verifier and lightweight bytecode verifier
  2712 as abstract algorithms, instantiated to the JVM;
  2713 
  2714 * Bali: Java source language formalization. Type system, operational
  2715 semantics, axiomatic semantics. Supported language features:
  2716 classes, interfaces, objects,virtual methods, static methods,
  2717 static/instance fields, arrays, access modifiers, definite
  2718 assignment, exceptions.
  2719 
  2720 
  2721 *** ZF ***
  2722 
  2723 * ZF/Constructible: consistency proof for AC (Gdel's constructible
  2724 universe, etc.);
  2725 
  2726 * Main ZF: virtually all theories converted to new-style format;
  2727 
  2728 
  2729 *** ML ***
  2730 
  2731 * Pure: Tactic.prove provides sane interface for internal proofs;
  2732 omits the infamous "standard" operation, so this is more appropriate
  2733 than prove_goalw_cterm in many situations (e.g. in simprocs);
  2734 
  2735 * Pure: improved error reporting of simprocs;
  2736 
  2737 * Provers: Simplifier.simproc(_i) provides sane interface for setting
  2738 up simprocs;
  2739 
  2740 
  2741 *** Document preparation ***
  2742 
  2743 * uses \par instead of \\ for line breaks in theory text. This may
  2744 shift some page breaks in large documents. To get the old behaviour
  2745 use \renewcommand{\isanewline}{\mbox{}\\\mbox{}} in root.tex.
  2746 
  2747 * minimized dependencies of isabelle.sty and isabellesym.sty on
  2748 other packages
  2749 
  2750 * \<euro> now needs package babel/greek instead of marvosym (which
  2751 broke \Rightarrow)
  2752 
  2753 * normal size for \<zero>...\<nine> (uses \mathbf instead of
  2754 textcomp package)
  2755 
  2756 
  2757 
  2758 New in Isabelle2002 (March 2002)
  2759 --------------------------------
  2760 
  2761 *** Document preparation ***
  2762 
  2763 * greatly simplified document preparation setup, including more
  2764 graceful interpretation of isatool usedir -i/-d/-D options, and more
  2765 instructive isatool mkdir; users should basically be able to get
  2766 started with "isatool mkdir HOL Test && isatool make"; alternatively,
  2767 users may run a separate document processing stage manually like this:
  2768 "isatool usedir -D output HOL Test && isatool document Test/output";
  2769 
  2770 * theory dependency graph may now be incorporated into documents;
  2771 isatool usedir -g true will produce session_graph.eps/.pdf for use
  2772 with \includegraphics of LaTeX;
  2773 
  2774 * proper spacing of consecutive markup elements, especially text
  2775 blocks after section headings;
  2776 
  2777 * support bold style (for single symbols only), input syntax is like
  2778 this: "\<^bold>\<alpha>" or "\<^bold>A";
  2779 
  2780 * \<bullet> is now output as bold \cdot by default, which looks much
  2781 better in printed text;
  2782 
  2783 * added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
  2784 note that these symbols are currently unavailable in Proof General /
  2785 X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
  2786 
  2787 * isatool latex no longer depends on changed TEXINPUTS, instead
  2788 isatool document copies the Isabelle style files to the target
  2789 location;
  2790 
  2791 
  2792 *** Isar ***
  2793 
  2794 * Pure/Provers: improved proof by cases and induction;
  2795   - 'case' command admits impromptu naming of parameters (such as
  2796     "case (Suc n)");
  2797   - 'induct' method divinates rule instantiation from the inductive
  2798     claim; no longer requires excessive ?P bindings for proper
  2799     instantiation of cases;
  2800   - 'induct' method properly enumerates all possibilities of set/type
  2801     rules; as a consequence facts may be also passed through *type*
  2802     rules without further ado;
  2803   - 'induct' method now derives symbolic cases from the *rulified*
  2804     rule (before it used to rulify cases stemming from the internal
  2805     atomized version); this means that the context of a non-atomic
  2806     statement becomes is included in the hypothesis, avoiding the
  2807     slightly cumbersome show "PROP ?case" form;
  2808   - 'induct' may now use elim-style induction rules without chaining
  2809     facts, using ``missing'' premises from the goal state; this allows
  2810     rules stemming from inductive sets to be applied in unstructured
  2811     scripts, while still benefitting from proper handling of non-atomic
  2812     statements; NB: major inductive premises need to be put first, all
  2813     the rest of the goal is passed through the induction;
  2814   - 'induct' proper support for mutual induction involving non-atomic
  2815     rule statements (uses the new concept of simultaneous goals, see
  2816     below);
  2817   - append all possible rule selections, but only use the first
  2818     success (no backtracking);
  2819   - removed obsolete "(simplified)" and "(stripped)" options of methods;
  2820   - undeclared rule case names default to numbers 1, 2, 3, ...;
  2821   - added 'print_induct_rules' (covered by help item in recent Proof
  2822     General versions);
  2823   - moved induct/cases attributes to Pure, methods to Provers;
  2824   - generic method setup instantiated for FOL and HOL;
  2825 
  2826 * Pure: support multiple simultaneous goal statements, for example
  2827 "have a: A and b: B" (same for 'theorem' etc.); being a pure
  2828 meta-level mechanism, this acts as if several individual goals had
  2829 been stated separately; in particular common proof methods need to be
  2830 repeated in order to cover all claims; note that a single elimination
  2831 step is *not* sufficient to establish the two conjunctions, so this
  2832 fails:
  2833 
  2834   assume "A & B" then have A and B ..   (*".." fails*)
  2835 
  2836 better use "obtain" in situations as above; alternative refer to
  2837 multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
  2838 
  2839 * Pure: proper integration with ``locales''; unlike the original
  2840 version by Florian Kammller, Isar locales package high-level proof
  2841 contexts rather than raw logical ones (e.g. we admit to include
  2842 attributes everywhere); operations on locales include merge and
  2843 rename; support for implicit arguments (``structures''); simultaneous
  2844 type-inference over imports and text; see also HOL/ex/Locales.thy for
  2845 some examples;
  2846 
  2847 * Pure: the following commands have been ``localized'', supporting a
  2848 target locale specification "(in name)": 'lemma', 'theorem',
  2849 'corollary', 'lemmas', 'theorems', 'declare'; the results will be
  2850 stored both within the locale and at the theory level (exported and
  2851 qualified by the locale name);
  2852 
  2853 * Pure: theory goals may now be specified in ``long'' form, with
  2854 ad-hoc contexts consisting of arbitrary locale elements. for example
  2855 ``lemma foo: fixes x assumes "A x" shows "B x"'' (local syntax and
  2856 definitions may be given, too); the result is a meta-level rule with
  2857 the context elements being discharged in the obvious way;
  2858 
  2859 * Pure: new proof command 'using' allows to augment currently used
  2860 facts after a goal statement ('using' is syntactically analogous to
  2861 'apply', but acts on the goal's facts only); this allows chained facts
  2862 to be separated into parts given before and after a claim, as in
  2863 ``from a and b have C using d and e <proof>'';
  2864 
  2865 * Pure: renamed "antecedent" case to "rule_context";
  2866 
  2867 * Pure: new 'judgment' command records explicit information about the
  2868 object-logic embedding (used by several tools internally); no longer
  2869 use hard-wired "Trueprop";
  2870 
  2871 * Pure: added 'corollary' command;
  2872 
  2873 * Pure: fixed 'token_translation' command;
  2874 
  2875 * Pure: removed obsolete 'exported' attribute;
  2876 
  2877 * Pure: dummy pattern "_" in is/let is now automatically lifted over
  2878 bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
  2879 supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
  2880 
  2881 * Pure: method 'atomize' presents local goal premises as object-level
  2882 statements (atomic meta-level propositions); setup controlled via
  2883 rewrite rules declarations of 'atomize' attribute; example
  2884 application: 'induct' method with proper rule statements in improper
  2885 proof *scripts*;
  2886 
  2887 * Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
  2888 now consider the syntactic context of assumptions, giving a better
  2889 chance to get type-inference of the arguments right (this is
  2890 especially important for locales);
  2891 
  2892 * Pure: "sorry" no longer requires quick_and_dirty in interactive
  2893 mode;
  2894 
  2895 * Pure/obtain: the formal conclusion "thesis", being marked as
  2896 ``internal'', may no longer be reference directly in the text;
  2897 potential INCOMPATIBILITY, may need to use "?thesis" in rare
  2898 situations;
  2899 
  2900 * Pure: generic 'sym' attribute which declares a rule both as pure
  2901 'elim?' and for the 'symmetric' operation;
  2902 
  2903 * Pure: marginal comments ``--'' may now occur just anywhere in the
  2904 text; the fixed correlation with particular command syntax has been
  2905 discontinued;
  2906 
  2907 * Pure: new method 'rules' is particularly well-suited for proof
  2908 search in intuitionistic logic; a bit slower than 'blast' or 'fast',
  2909 but often produces more compact proof terms with less detours;
  2910 
  2911 * Pure/Provers/classical: simplified integration with pure rule
  2912 attributes and methods; the classical "intro?/elim?/dest?"
  2913 declarations coincide with the pure ones; the "rule" method no longer
  2914 includes classically swapped intros; "intro" and "elim" methods no
  2915 longer pick rules from the context; also got rid of ML declarations
  2916 AddXIs/AddXEs/AddXDs; all of this has some potential for
  2917 INCOMPATIBILITY;
  2918 
  2919 * Provers/classical: attribute 'swapped' produces classical inversions
  2920 of introduction rules;
  2921 
  2922 * Provers/simplifier: 'simplified' attribute may refer to explicit
  2923 rules instead of full simplifier context; 'iff' attribute handles
  2924 conditional rules;
  2925 
  2926 * HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
  2927 
  2928 * HOL: 'recdef' now fails on unfinished automated proofs, use
  2929 "(permissive)" option to recover old behavior;
  2930 
  2931 * HOL: 'inductive' no longer features separate (collective) attributes
  2932 for 'intros' (was found too confusing);
  2933 
  2934 * HOL: properly declared induction rules less_induct and
  2935 wf_induct_rule;
  2936 
  2937 
  2938 *** HOL ***
  2939 
  2940 * HOL: moved over to sane numeral syntax; the new policy is as
  2941 follows:
  2942 
  2943   - 0 and 1 are polymorphic constants, which are defined on any
  2944   numeric type (nat, int, real etc.);
  2945 
  2946   - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
  2947   binary representation internally;
  2948 
  2949   - type nat has special constructor Suc, and generally prefers Suc 0
  2950   over 1::nat and Suc (Suc 0) over 2::nat;
  2951 
  2952 This change may cause significant problems of INCOMPATIBILITY; here
  2953 are some hints on converting existing sources:
  2954 
  2955   - due to the new "num" token, "-0" and "-1" etc. are now atomic
  2956   entities, so expressions involving "-" (unary or binary minus) need
  2957   to be spaced properly;
  2958 
  2959   - existing occurrences of "1" may need to be constraint "1::nat" or
  2960   even replaced by Suc 0; similar for old "2";
  2961 
  2962   - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
  2963 
  2964   - remove all special provisions on numerals in proofs;
  2965 
  2966 * HOL: simp rules nat_number expand numerals on nat to Suc/0
  2967 representation (depends on bin_arith_simps in the default context);
  2968 
  2969 * HOL: symbolic syntax for x^2 (numeral 2);
  2970 
  2971 * HOL: the class of all HOL types is now called "type" rather than
  2972 "term"; INCOMPATIBILITY, need to adapt references to this type class
  2973 in axclass/classes, instance/arities, and (usually rare) occurrences
  2974 in typings (of consts etc.); internally the class is called
  2975 "HOL.type", ML programs should refer to HOLogic.typeS;
  2976 
  2977 * HOL/record package improvements:
  2978   - new derived operations "fields" to build a partial record section,
  2979     "extend" to promote a fixed record to a record scheme, and
  2980     "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
  2981     declared as simp by default;
  2982   - shared operations ("more", "fields", etc.) now need to be always
  2983     qualified) --- potential INCOMPATIBILITY;
  2984   - removed "make_scheme" operations (use "make" with "extend") --
  2985     INCOMPATIBILITY;
  2986   - removed "more" class (simply use "term") -- INCOMPATIBILITY;
  2987   - provides cases/induct rules for use with corresponding Isar
  2988     methods (for concrete records, record schemes, concrete more
  2989     parts, and schematic more parts -- in that order);
  2990   - internal definitions directly based on a light-weight abstract
  2991     theory of product types over typedef rather than datatype;
  2992 
  2993 * HOL: generic code generator for generating executable ML code from
  2994 specifications; specific support for HOL constructs such as inductive
  2995 datatypes and sets, as well as recursive functions; can be invoked
  2996 via 'generate_code' theory section;
  2997 
  2998 * HOL: canonical cases/induct rules for n-tuples (n = 3..7);
  2999 
  3000 * HOL: consolidated and renamed several theories.  In particular:
  3001         Ord.thy has been absorbed into HOL.thy
  3002         String.thy has been absorbed into List.thy
  3003 
  3004 * HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
  3005 (beware of argument permutation!);
  3006 
  3007 * HOL: linorder_less_split superseded by linorder_cases;
  3008 
  3009 * HOL/List: "nodups" renamed to "distinct";
  3010 
  3011 * HOL: added "The" definite description operator; move Hilbert's "Eps"
  3012 to peripheral theory "Hilbert_Choice"; some INCOMPATIBILITIES:
  3013   - Ex_def has changed, now need to use some_eq_ex
  3014 
  3015 * HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
  3016 in this (rare) case use:
  3017 
  3018   delSWrapper "split_all_tac"
  3019   addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
  3020 
  3021 * HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
  3022 MAY FAIL;
  3023 
  3024 * HOL: introduced f^n = f o ... o f; warning: due to the limits of
  3025 Isabelle's type classes, ^ on functions and relations has too general
  3026 a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
  3027 necessary to attach explicit type constraints;
  3028 
  3029 * HOL/Relation: the prefix name of the infix "O" has been changed from
  3030 "comp" to "rel_comp"; INCOMPATIBILITY: a few theorems have been
  3031 renamed accordingly (eg "compI" -> "rel_compI").
  3032 
  3033 * HOL: syntax translations now work properly with numerals and records
  3034 expressions;
  3035 
  3036 * HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
  3037 of "lam" -- INCOMPATIBILITY;
  3038 
  3039 * HOL: got rid of some global declarations (potential INCOMPATIBILITY
  3040 for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
  3041 renamed "Product_Type.unit";
  3042 
  3043 * HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
  3044 
  3045 * HOL: removed obsolete theorem "optionE" (use "option.exhaust", or
  3046 the "cases" method);
  3047 
  3048 * HOL/GroupTheory: group theory examples including Sylow's theorem (by
  3049 Florian Kammller);
  3050 
  3051 * HOL/IMP: updated and converted to new-style theory format; several
  3052 parts turned into readable document, with proper Isar proof texts and
  3053 some explanations (by Gerwin Klein);
  3054 
  3055 * HOL-Real: added Complex_Numbers (by Gertrud Bauer);
  3056 
  3057 * HOL-Hyperreal is now a logic image;
  3058 
  3059 
  3060 *** HOLCF ***
  3061 
  3062 * Isar: consts/constdefs supports mixfix syntax for continuous
  3063 operations;
  3064 
  3065 * Isar: domain package adapted to new-style theory format, e.g. see
  3066 HOLCF/ex/Dnat.thy;
  3067 
  3068 * theory Lift: proper use of rep_datatype lift instead of ML hacks --
  3069 potential INCOMPATIBILITY; now use plain induct_tac instead of former
  3070 lift.induct_tac, always use UU instead of Undef;
  3071 
  3072 * HOLCF/IMP: updated and converted to new-style theory;
  3073 
  3074 
  3075 *** ZF ***
  3076 
  3077 * Isar: proper integration of logic-specific tools and packages,
  3078 including theory commands '(co)inductive', '(co)datatype',
  3079 'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
  3080 'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
  3081 
  3082 * theory Main no longer includes AC; for the Axiom of Choice, base
  3083 your theory on Main_ZFC;
  3084 
  3085 * the integer library now covers quotients and remainders, with many
  3086 laws relating division to addition, multiplication, etc.;
  3087 
  3088 * ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
  3089 typeless version of the formalism;
  3090 
  3091 * ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory
  3092 format;
  3093 
  3094 * ZF/Induct: new directory for examples of inductive definitions,
  3095 including theory Multiset for multiset orderings; converted to
  3096 new-style theory format;
  3097 
  3098 * ZF: many new theorems about lists, ordinals, etc.;
  3099 
  3100 
  3101 *** General ***
  3102 
  3103 * Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
  3104 variable proof controls level of detail: 0 = no proofs (only oracle
  3105 dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
  3106 also ref manual for further ML interfaces;
  3107 
  3108 * Pure/axclass: removed obsolete ML interface
  3109 goal_subclass/goal_arity;
  3110 
  3111 * Pure/syntax: new token syntax "num" for plain numerals (without "#"
  3112 of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
  3113 separate tokens, so expressions involving minus need to be spaced
  3114 properly;
  3115 
  3116 * Pure/syntax: support non-oriented infixes, using keyword "infix"
  3117 rather than "infixl" or "infixr";
  3118 
  3119 * Pure/syntax: concrete syntax for dummy type variables admits genuine
  3120 sort constraint specifications in type inference; e.g. "x::_::foo"
  3121 ensures that the type of "x" is of sort "foo" (but not necessarily a
  3122 type variable);
  3123 
  3124 * Pure/syntax: print modes "type_brackets" and "no_type_brackets"
  3125 control output of nested => (types); the default behavior is
  3126 "type_brackets";
  3127 
  3128 * Pure/syntax: builtin parse translation for "_constify" turns valued
  3129 tokens into AST constants;
  3130 
  3131 * Pure/syntax: prefer later declarations of translations and print
  3132 translation functions; potential INCOMPATIBILITY: need to reverse
  3133 multiple declarations for same syntax element constant;
  3134 
  3135 * Pure/show_hyps reset by default (in accordance to existing Isar
  3136 practice);
  3137 
  3138 * Provers/classical: renamed addaltern to addafter, addSaltern to
  3139 addSafter;
  3140 
  3141 * Provers/clasimp: ``iff'' declarations now handle conditional rules
  3142 as well;
  3143 
  3144 * system: tested support for MacOS X; should be able to get Isabelle +
  3145 Proof General to work in a plain Terminal after installing Poly/ML
  3146 (e.g. from the Isabelle distribution area) and GNU bash alone
  3147 (e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
  3148 support requires further installations, e.g. from
  3149 http://fink.sourceforge.net/);
  3150 
  3151 * system: support Poly/ML 4.1.1 (able to manage larger heaps);
  3152 
  3153 * system: reduced base memory usage by Poly/ML (approx. 20 MB instead
  3154 of 40 MB), cf. ML_OPTIONS;
  3155 
  3156 * system: Proof General keywords specification is now part of the
  3157 Isabelle distribution (see etc/isar-keywords.el);
  3158 
  3159 * system: support for persistent Proof General sessions (refrain from
  3160 outdating all loaded theories on startup); user may create writable
  3161 logic images like this: ``isabelle -q HOL Test'';
  3162 
  3163 * system: smart selection of Isabelle process versus Isabelle
  3164 interface, accommodates case-insensitive file systems (e.g. HFS+); may
  3165 run both "isabelle" and "Isabelle" even if file names are badly
  3166 damaged (executable inspects the case of the first letter of its own
  3167 name); added separate "isabelle-process" and "isabelle-interface";
  3168 
  3169 * system: refrain from any attempt at filtering input streams; no
  3170 longer support ``8bit'' encoding of old isabelle font, instead proper
  3171 iso-latin characters may now be used; the related isatools
  3172 "symbolinput" and "nonascii" have disappeared as well;
  3173 
  3174 * system: removed old "xterm" interface (the print modes "xterm" and
  3175 "xterm_color" are still available for direct use in a suitable
  3176 terminal);
  3177 
  3178 
  3179 
  3180 New in Isabelle99-2 (February 2001)
  3181 -----------------------------------
  3182 
  3183 *** Overview of INCOMPATIBILITIES ***
  3184 
  3185 * HOL: please note that theories in the Library and elsewhere often use the
  3186 new-style (Isar) format; to refer to their theorems in an ML script you must
  3187 bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
  3188 
  3189 * HOL: inductive package no longer splits induction rule aggressively,
  3190 but only as far as specified by the introductions given; the old
  3191 format may be recovered via ML function complete_split_rule or attribute
  3192 'split_rule (complete)';
  3193 
  3194 * HOL: induct renamed to lfp_induct, lfp_Tarski to lfp_unfold,
  3195 gfp_Tarski to gfp_unfold;
  3196 
  3197 * HOL: contrapos, contrapos2 renamed to contrapos_nn, contrapos_pp;
  3198 
  3199 * HOL: infix "dvd" now has priority 50 rather than 70 (because it is a
  3200 relation); infix "^^" has been renamed "``"; infix "``" has been
  3201 renamed "`"; "univalent" has been renamed "single_valued";
  3202 
  3203 * HOL/Real: "rinv" and "hrinv" replaced by overloaded "inverse"
  3204 operation;
  3205 
  3206 * HOLCF: infix "`" has been renamed "$"; the symbol syntax is \<cdot>;
  3207 
  3208 * Isar: 'obtain' no longer declares "that" fact as simp/intro;
  3209 
  3210 * Isar/HOL: method 'induct' now handles non-atomic goals; as a
  3211 consequence, it is no longer monotonic wrt. the local goal context
  3212 (which is now passed through the inductive cases);
  3213 
  3214 * Document preparation: renamed standard symbols \<ll> to \<lless> and
  3215 \<gg> to \<ggreater>;
  3216 
  3217 
  3218 *** Document preparation ***
  3219 
  3220 * \isabellestyle{NAME} selects version of Isabelle output (currently
  3221 available: are "it" for near math-mode best-style output, "sl" for
  3222 slanted text style, and "tt" for plain type-writer; if no
  3223 \isabellestyle command is given, output is according to slanted
  3224 type-writer);
  3225 
  3226 * support sub/super scripts (for single symbols only), input syntax is
  3227 like this: "A\<^sup>*" or "A\<^sup>\<star>";
  3228 
  3229 * some more standard symbols; see Appendix A of the system manual for
  3230 the complete list of symbols defined in isabellesym.sty;
  3231 
  3232 * improved isabelle style files; more abstract symbol implementation
  3233 (should now use \isamath{...} and \isatext{...} in custom symbol
  3234 definitions);
  3235 
  3236 * antiquotation @{goals} and @{subgoals} for output of *dynamic* goals
  3237 state; Note that presentation of goal states does not conform to
  3238 actual human-readable proof documents.  Please do not include goal
  3239 states into document output unless you really know what you are doing!
  3240 
  3241 * proper indentation of antiquoted output with proportional LaTeX
  3242 fonts;
  3243 
  3244 * no_document ML operator temporarily disables LaTeX document
  3245 generation;
  3246 
  3247 * isatool unsymbolize tunes sources for plain ASCII communication;
  3248 
  3249 
  3250 *** Isar ***
  3251 
  3252 * Pure: Isar now suffers initial goal statements to contain unbound
  3253 schematic variables (this does not conform to actual readable proof
  3254 documents, due to unpredictable outcome and non-compositional proof
  3255 checking); users who know what they are doing may use schematic goals
  3256 for Prolog-style synthesis of proven results;
  3257 
  3258 * Pure: assumption method (an implicit finishing) now handles actual
  3259 rules as well;
  3260 
  3261 * Pure: improved 'obtain' --- moved to Pure, insert "that" into
  3262 initial goal, declare "that" only as Pure intro (only for single
  3263 steps); the "that" rule assumption may now be involved in implicit
  3264 finishing, thus ".." becomes a feasible for trivial obtains;
  3265 
  3266 * Pure: default proof step now includes 'intro_classes'; thus trivial
  3267 instance proofs may be performed by "..";
  3268 
  3269 * Pure: ?thesis / ?this / "..." now work for pure meta-level
  3270 statements as well;
  3271 
  3272 * Pure: more robust selection of calculational rules;
  3273 
  3274 * Pure: the builtin notion of 'finished' goal now includes the ==-refl
  3275 rule (as well as the assumption rule);
  3276 
  3277 * Pure: 'thm_deps' command visualizes dependencies of theorems and
  3278 lemmas, using the graph browser tool;
  3279 
  3280 * Pure: predict failure of "show" in interactive mode;
  3281 
  3282 * Pure: 'thms_containing' now takes actual terms as arguments;
  3283 
  3284 * HOL: improved method 'induct' --- now handles non-atomic goals
  3285 (potential INCOMPATIBILITY); tuned error handling;
  3286 
  3287 * HOL: cases and induct rules now provide explicit hints about the
  3288 number of facts to be consumed (0 for "type" and 1 for "set" rules);
  3289 any remaining facts are inserted into the goal verbatim;
  3290 
  3291 * HOL: local contexts (aka cases) may now contain term bindings as
  3292 well; the 'cases' and 'induct' methods new provide a ?case binding for
  3293 the result to be shown in each case;
  3294 
  3295 * HOL: added 'recdef_tc' command;
  3296 
  3297 * isatool convert assists in eliminating legacy ML scripts;
  3298 
  3299 
  3300 *** HOL ***
  3301 
  3302 * HOL/Library: a collection of generic theories to be used together
  3303 with main HOL; the theory loader path already includes this directory
  3304 by default; the following existing theories have been moved here:
  3305 HOL/Induct/Multiset, HOL/Induct/Acc (as Accessible_Part), HOL/While
  3306 (as While_Combinator), HOL/Lex/Prefix (as List_Prefix);
  3307 
  3308 * HOL/Unix: "Some aspects of Unix file-system security", a typical
  3309 modelling and verification task performed in Isabelle/HOL +
  3310 Isabelle/Isar + Isabelle document preparation (by Markus Wenzel).
  3311 
  3312 * HOL/Algebra: special summation operator SUM no longer exists, it has
  3313 been replaced by setsum; infix 'assoc' now has priority 50 (like
  3314 'dvd'); axiom 'one_not_zero' has been moved from axclass 'ring' to
  3315 'domain', this makes the theory consistent with mathematical
  3316 literature;
  3317 
  3318 * HOL basics: added overloaded operations "inverse" and "divide"
  3319 (infix "/"), syntax for generic "abs" operation, generic summation
  3320 operator \<Sum>;
  3321 
  3322 * HOL/typedef: simplified package, provide more useful rules (see also
  3323 HOL/subset.thy);
  3324 
  3325 * HOL/datatype: induction rule for arbitrarily branching datatypes is
  3326 now expressed as a proper nested rule (old-style tactic scripts may
  3327 require atomize_strip_tac to cope with non-atomic premises);
  3328 
  3329 * HOL: renamed theory "Prod" to "Product_Type", renamed "split" rule
  3330 to "split_conv" (old name still available for compatibility);
  3331 
  3332 * HOL: improved concrete syntax for strings (e.g. allows translation
  3333 rules with string literals);
  3334 
  3335 * HOL-Real-Hyperreal: this extends HOL-Real with the hyperreals
  3336  and Fleuriot's mechanization of analysis, including the transcendental
  3337  functions for the reals;
  3338 
  3339 * HOL/Real, HOL/Hyperreal: improved arithmetic simplification;
  3340 
  3341 
  3342 *** CTT ***
  3343 
  3344 * CTT: x-symbol support for Pi, Sigma, -->, : (membership); note that
  3345 "lam" is displayed as TWO lambda-symbols
  3346 
  3347 * CTT: theory Main now available, containing everything (that is, Bool
  3348 and Arith);
  3349 
  3350 
  3351 *** General ***
  3352 
  3353 * Pure: the Simplifier has been implemented properly as a derived rule
  3354 outside of the actual kernel (at last!); the overall performance
  3355 penalty in practical applications is about 50%, while reliability of
  3356 the Isabelle inference kernel has been greatly improved;
  3357 
  3358 * print modes "brackets" and "no_brackets" control output of nested =>
  3359 (types) and ==> (props); the default behaviour is "brackets";
  3360 
  3361 * Provers: fast_tac (and friends) now handle actual object-logic rules
  3362 as assumptions as well;
  3363 
  3364 * system: support Poly/ML 4.0;
  3365 
  3366 * system: isatool install handles KDE version 1 or 2;
  3367 
  3368 
  3369 
  3370 New in Isabelle99-1 (October 2000)
  3371 ----------------------------------
  3372 
  3373 *** Overview of INCOMPATIBILITIES ***
  3374 
  3375 * HOL: simplification of natural numbers is much changed; to partly
  3376 recover the old behaviour (e.g. to prevent n+n rewriting to #2*n)
  3377 issue the following ML commands:
  3378 
  3379   Delsimprocs Nat_Numeral_Simprocs.cancel_numerals;
  3380   Delsimprocs [Nat_Numeral_Simprocs.combine_numerals];
  3381 
  3382 * HOL: simplification no longer dives into case-expressions; this is
  3383 controlled by "t.weak_case_cong" for each datatype t;
  3384 
  3385 * HOL: nat_less_induct renamed to less_induct;
  3386 
  3387 * HOL: systematic renaming of the SOME (Eps) rules, may use isatool
  3388 fixsome to patch .thy and .ML sources automatically;
  3389 
  3390   select_equality  -> some_equality
  3391   select_eq_Ex     -> some_eq_ex
  3392   selectI2EX       -> someI2_ex
  3393   selectI2         -> someI2
  3394   selectI          -> someI
  3395   select1_equality -> some1_equality
  3396   Eps_sym_eq       -> some_sym_eq_trivial
  3397   Eps_eq           -> some_eq_trivial
  3398 
  3399 * HOL: exhaust_tac on datatypes superceded by new generic case_tac;
  3400 
  3401 * HOL: removed obsolete theorem binding expand_if (refer to split_if
  3402 instead);
  3403 
  3404 * HOL: the recursion equations generated by 'recdef' are now called
  3405 f.simps instead of f.rules;
  3406 
  3407 * HOL: qed_spec_mp now also handles bounded ALL as well;
  3408 
  3409 * HOL: 0 is now overloaded, so the type constraint ":: nat" may
  3410 sometimes be needed;
  3411 
  3412 * HOL: the constant for "f``x" is now "image" rather than "op ``";
  3413 
  3414 * HOL: the constant for "f-``x" is now "vimage" rather than "op -``";
  3415 
  3416 * HOL: the disjoint sum is now "<+>" instead of "Plus"; the cartesian
  3417 product is now "<*>" instead of "Times"; the lexicographic product is
  3418 now "<*lex*>" instead of "**";
  3419 
  3420 * HOL: theory Sexp is now in HOL/Induct examples (it used to be part
  3421 of main HOL, but was unused); better use HOL's datatype package;
  3422 
  3423 * HOL: removed "symbols" syntax for constant "override" of theory Map;
  3424 the old syntax may be recovered as follows:
  3425 
  3426   syntax (symbols)
  3427     override  :: "('a ~=> 'b) => ('a ~=> 'b) => ('a ~=> 'b)"
  3428       (infixl "\\<oplus>" 100)
  3429 
  3430 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3431 
  3432 * HOL/ML: even fewer consts are declared as global (see theories Ord,
  3433 Lfp, Gfp, WF); this only affects ML packages that refer to const names
  3434 internally;
  3435 
  3436 * HOL and ZF: syntax for quotienting wrt an equivalence relation
  3437 changed from A/r to A//r;
  3438 
  3439 * ZF: new treatment of arithmetic (nat & int) may break some old
  3440 proofs;
  3441 
  3442 * Isar: renamed some attributes (RS -> THEN, simplify -> simplified,
  3443 rulify -> rule_format, elimify -> elim_format, ...);
  3444 
  3445 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3446 intro/intro!/intro!! flags to intro!/intro/intro? (in most cases, one
  3447 should have to change intro!! to intro? only); replaced "delrule" by
  3448 "rule del";
  3449 
  3450 * Isar/HOL: renamed "intrs" to "intros" in inductive definitions;
  3451 
  3452 * Provers: strengthened force_tac by using new first_best_tac;
  3453 
  3454 * LaTeX document preparation: several changes of isabelle.sty (see
  3455 lib/texinputs);
  3456 
  3457 
  3458 *** Document preparation ***
  3459 
  3460 * formal comments (text blocks etc.) in new-style theories may now
  3461 contain antiquotations of thm/prop/term/typ/text to be presented
  3462 according to latex print mode; concrete syntax is like this:
  3463 @{term[show_types] "f(x) = a + x"};
  3464 
  3465 * isatool mkdir provides easy setup of Isabelle session directories,
  3466 including proper document sources;
  3467 
  3468 * generated LaTeX sources are now deleted after successful run
  3469 (isatool document -c); may retain a copy somewhere else via -D option
  3470 of isatool usedir;
  3471 
  3472 * isatool usedir -D now lets isatool latex -o sty update the Isabelle
  3473 style files, achieving self-contained LaTeX sources and simplifying
  3474 LaTeX debugging;
  3475 
  3476 * old-style theories now produce (crude) LaTeX output as well;
  3477 
  3478 * browser info session directories are now self-contained (may be put
  3479 on WWW server seperately); improved graphs of nested sessions; removed
  3480 graph for 'all sessions';
  3481 
  3482 * several improvements in isabelle style files; \isabellestyle{it}
  3483 produces fake math mode output; \isamarkupheader is now \section by
  3484 default; see lib/texinputs/isabelle.sty etc.;
  3485 
  3486 
  3487 *** Isar ***
  3488 
  3489 * Isar/Pure: local results and corresponding term bindings are now
  3490 subject to Hindley-Milner polymorphism (similar to ML); this
  3491 accommodates incremental type-inference very nicely;
  3492 
  3493 * Isar/Pure: new derived language element 'obtain' supports
  3494 generalized existence reasoning;
  3495 
  3496 * Isar/Pure: new calculational elements 'moreover' and 'ultimately'
  3497 support accumulation of results, without applying any rules yet;
  3498 useful to collect intermediate results without explicit name
  3499 references, and for use with transitivity rules with more than 2
  3500 premises;
  3501 
  3502 * Isar/Pure: scalable support for case-analysis type proofs: new
  3503 'case' language element refers to local contexts symbolically, as
  3504 produced by certain proof methods; internally, case names are attached
  3505 to theorems as "tags";
  3506 
  3507 * Isar/Pure: theory command 'hide' removes declarations from
  3508 class/type/const name spaces;
  3509 
  3510 * Isar/Pure: theory command 'defs' supports option "(overloaded)" to
  3511 indicate potential overloading;
  3512 
  3513 * Isar/Pure: changed syntax of local blocks from {{ }} to { };
  3514 
  3515 * Isar/Pure: syntax of sorts made 'inner', i.e. have to write
  3516 "{a,b,c}" instead of {a,b,c};
  3517 
  3518 * Isar/Pure now provides its own version of intro/elim/dest
  3519 attributes; useful for building new logics, but beware of confusion
  3520 with the version in Provers/classical;
  3521 
  3522 * Isar/Pure: the local context of (non-atomic) goals is provided via
  3523 case name 'antecedent';
  3524 
  3525 * Isar/Pure: removed obsolete 'transfer' attribute (transfer of thms
  3526 to the current context is now done automatically);
  3527 
  3528 * Isar/Pure: theory command 'method_setup' provides a simple interface
  3529 for definining proof methods in ML;
  3530 
  3531 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3532 intro/intro!/intro!! flags to intro!/intro/intro? (INCOMPATIBILITY, in
  3533 most cases, one should have to change intro!! to intro? only);
  3534 replaced "delrule" by "rule del";
  3535 
  3536 * Isar/Provers: new 'hypsubst' method, plain 'subst' method and
  3537 'symmetric' attribute (the latter supercedes [RS sym]);
  3538 
  3539 * Isar/Provers: splitter support (via 'split' attribute and 'simp'
  3540 method modifier); 'simp' method: 'only:' modifier removes loopers as
  3541 well (including splits);
  3542 
  3543 * Isar/Provers: Simplifier and Classical methods now support all kind
  3544 of modifiers used in the past, including 'cong', 'iff', etc.
  3545 
  3546 * Isar/Provers: added 'fastsimp' and 'clarsimp' methods (combination
  3547 of Simplifier and Classical reasoner);
  3548 
  3549 * Isar/HOL: new proof method 'cases' and improved version of 'induct'
  3550 now support named cases; major packages (inductive, datatype, primrec,
  3551 recdef) support case names and properly name parameters;
  3552 
  3553 * Isar/HOL: new transitivity rules for substitution in inequalities --
  3554 monotonicity conditions are extracted to be proven at end of
  3555 calculations;
  3556 
  3557 * Isar/HOL: removed 'case_split' thm binding, should use 'cases' proof
  3558 method anyway;
  3559 
  3560 * Isar/HOL: removed old expand_if = split_if; theorems if_splits =
  3561 split_if split_if_asm; datatype package provides theorems foo.splits =
  3562 foo.split foo.split_asm for each datatype;
  3563 
  3564 * Isar/HOL: tuned inductive package, rename "intrs" to "intros"
  3565 (potential INCOMPATIBILITY), emulation of mk_cases feature for proof
  3566 scripts: new 'inductive_cases' command and 'ind_cases' method; (Note:
  3567 use "(cases (simplified))" method in proper proof texts);
  3568 
  3569 * Isar/HOL: added global 'arith_split' attribute for 'arith' method;
  3570 
  3571 * Isar: names of theorems etc. may be natural numbers as well;
  3572 
  3573 * Isar: 'pr' command: optional arguments for goals_limit and
  3574 ProofContext.prems_limit; no longer prints theory contexts, but only
  3575 proof states;
  3576 
  3577 * Isar: diagnostic commands 'pr', 'thm', 'prop', 'term', 'typ' admit
  3578 additional print modes to be specified; e.g. "pr(latex)" will print
  3579 proof state according to the Isabelle LaTeX style;
  3580 
  3581 * Isar: improved support for emulating tactic scripts, including proof
  3582 methods 'rule_tac' etc., 'cut_tac', 'thin_tac', 'subgoal_tac',
  3583 'rename_tac', 'rotate_tac', 'tactic', and 'case_tac' / 'induct_tac'
  3584 (for HOL datatypes);
  3585 
  3586 * Isar: simplified (more robust) goal selection of proof methods: 1st
  3587 goal, all goals, or explicit goal specifier (tactic emulation); thus
  3588 'proof method scripts' have to be in depth-first order;
  3589 
  3590 * Isar: tuned 'let' syntax: replaced 'as' keyword by 'and';
  3591 
  3592 * Isar: removed 'help' command, which hasn't been too helpful anyway;
  3593 should instead use individual commands for printing items
  3594 (print_commands, print_methods etc.);
  3595 
  3596 * Isar: added 'nothing' --- the empty list of theorems;
  3597 
  3598 
  3599 *** HOL ***
  3600 
  3601 * HOL/MicroJava: formalization of a fragment of Java, together with a
  3602 corresponding virtual machine and a specification of its bytecode
  3603 verifier and a lightweight bytecode verifier, including proofs of
  3604 type-safety; by Gerwin Klein, Tobias Nipkow, David von Oheimb, and
  3605 Cornelia Pusch (see also the homepage of project Bali at
  3606 http://isabelle.in.tum.de/Bali/);
  3607 
  3608 * HOL/Algebra: new theory of rings and univariate polynomials, by
  3609 Clemens Ballarin;
  3610 
  3611 * HOL/NumberTheory: fundamental Theorem of Arithmetic, Chinese
  3612 Remainder Theorem, Fermat/Euler Theorem, Wilson's Theorem, by Thomas M
  3613 Rasmussen;
  3614 
  3615 * HOL/Lattice: fundamental concepts of lattice theory and order
  3616 structures, including duals, properties of bounds versus algebraic
  3617 laws, lattice operations versus set-theoretic ones, the Knaster-Tarski
  3618 Theorem for complete lattices etc.; may also serve as a demonstration
  3619 for abstract algebraic reasoning using axiomatic type classes, and
  3620 mathematics-style proof in Isabelle/Isar; by Markus Wenzel;
  3621 
  3622 * HOL/Prolog: a (bare-bones) implementation of Lambda-Prolog, by David
  3623 von Oheimb;
  3624 
  3625 * HOL/IMPP: extension of IMP with local variables and mutually
  3626 recursive procedures, by David von Oheimb;
  3627 
  3628 * HOL/Lambda: converted into new-style theory and document;
  3629 
  3630 * HOL/ex/Multiquote: example of multiple nested quotations and
  3631 anti-quotations -- basically a generalized version of de-Bruijn
  3632 representation; very useful in avoiding lifting of operations;
  3633 
  3634 * HOL/record: added general record equality rule to simpset; fixed
  3635 select-update simplification procedure to handle extended records as
  3636 well; admit "r" as field name;
  3637 
  3638 * HOL: 0 is now overloaded over the new sort "zero", allowing its use with
  3639 other numeric types and also as the identity of groups, rings, etc.;
  3640 
  3641 * HOL: new axclass plus_ac0 for addition with the AC-laws and 0 as identity.
  3642 Types nat and int belong to this axclass;
  3643 
  3644 * HOL: greatly improved simplification involving numerals of type nat, int, real:
  3645    (i + #8 + j) = Suc k simplifies to  #7 + (i + j) = k
  3646    i*j + k + j*#3*i     simplifies to  #4*(i*j) + k
  3647   two terms #m*u and #n*u are replaced by #(m+n)*u
  3648     (where #m, #n and u can implicitly be 1; this is simproc combine_numerals)
  3649   and the term/formula #m*u+x ~~ #n*u+y simplifies simplifies to #(m-n)+x ~~ y
  3650     or x ~~ #(n-m)+y, where ~~ is one of = < <= or - (simproc cancel_numerals);
  3651 
  3652 * HOL: meson_tac is available (previously in ex/meson.ML); it is a
  3653 powerful prover for predicate logic but knows nothing of clasets; see
  3654 ex/mesontest.ML and ex/mesontest2.ML for example applications;
  3655 
  3656 * HOL: new version of "case_tac" subsumes both boolean case split and
  3657 "exhaust_tac" on datatypes; INCOMPATIBILITY: exhaust_tac no longer
  3658 exists, may define val exhaust_tac = case_tac for ad-hoc portability;
  3659 
  3660 * HOL: simplification no longer dives into case-expressions: only the
  3661 selector expression is simplified, but not the remaining arms; to
  3662 enable full simplification of case-expressions for datatype t, you may
  3663 remove t.weak_case_cong from the simpset, either globally (Delcongs
  3664 [thm"t.weak_case_cong"];) or locally (delcongs [...]).
  3665 
  3666 * HOL/recdef: the recursion equations generated by 'recdef' for
  3667 function 'f' are now called f.simps instead of f.rules; if all
  3668 termination conditions are proved automatically, these simplification
  3669 rules are added to the simpset, as in primrec; rules may be named
  3670 individually as well, resulting in a separate list of theorems for
  3671 each equation;
  3672 
  3673 * HOL/While is a new theory that provides a while-combinator. It
  3674 permits the definition of tail-recursive functions without the
  3675 provision of a termination measure. The latter is necessary once the
  3676 invariant proof rule for while is applied.
  3677 
  3678 * HOL: new (overloaded) notation for the set of elements below/above
  3679 some element: {..u}, {..u(}, {l..}, {)l..}. See theory SetInterval.
  3680 
  3681 * HOL: theorems impI, allI, ballI bound as "strip";
  3682 
  3683 * HOL: new tactic induct_thm_tac: thm -> string -> int -> tactic
  3684 induct_tac th "x1 ... xn" expects th to have a conclusion of the form
  3685 P v1 ... vn and abbreviates res_inst_tac [("v1","x1"),...,("vn","xn")] th;
  3686 
  3687 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3688 
  3689 * HOL: theory Sexp now in HOL/Induct examples (it used to be part of
  3690 main HOL, but was unused);
  3691 
  3692 * HOL: fewer consts declared as global (e.g. have to refer to
  3693 "Lfp.lfp" instead of "lfp" internally; affects ML packages only);
  3694 
  3695 * HOL: tuned AST representation of nested pairs, avoiding bogus output
  3696 in case of overlap with user translations (e.g. judgements over
  3697 tuples); (note that the underlying logical represenation is still
  3698 bogus);
  3699 
  3700 
  3701 *** ZF ***
  3702 
  3703 * ZF: simplification automatically cancels common terms in arithmetic
  3704 expressions over nat and int;
  3705 
  3706 * ZF: new treatment of nat to minimize type-checking: all operators
  3707 coerce their operands to a natural number using the function natify,
  3708 making the algebraic laws unconditional;
  3709 
  3710 * ZF: as above, for int: operators coerce their operands to an integer
  3711 using the function intify;
  3712 
  3713 * ZF: the integer library now contains many of the usual laws for the
  3714 orderings, including $<=, and monotonicity laws for $+ and $*;
  3715 
  3716 * ZF: new example ZF/ex/NatSum to demonstrate integer arithmetic
  3717 simplification;
  3718 
  3719 * FOL and ZF: AddIffs now available, giving theorems of the form P<->Q
  3720 to the simplifier and classical reasoner simultaneously;
  3721 
  3722 
  3723 *** General ***
  3724 
  3725 * Provers: blast_tac now handles actual object-logic rules as
  3726 assumptions; note that auto_tac uses blast_tac internally as well;
  3727 
  3728 * Provers: new functions rulify/rulify_no_asm: thm -> thm for turning
  3729 outer -->/All/Ball into ==>/!!; qed_spec_mp now uses rulify_no_asm;
  3730 
  3731 * Provers: delrules now handles destruct rules as well (no longer need
  3732 explicit make_elim);
  3733 
  3734 * Provers: Blast_tac now warns of and ignores "weak elimination rules" e.g.
  3735   [| inj ?f;          ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3736 use instead the strong form,
  3737   [| inj ?f; ~ ?W ==> ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3738 in HOL, FOL and ZF the function cla_make_elim will create such rules
  3739 from destruct-rules;
  3740 
  3741 * Provers: Simplifier.easy_setup provides a fast path to basic
  3742 Simplifier setup for new object-logics;
  3743 
  3744 * Pure: AST translation rules no longer require constant head on LHS;
  3745 
  3746 * Pure: improved name spaces: ambiguous output is qualified; support
  3747 for hiding of names;
  3748 
  3749 * system: smart setup of canonical ML_HOME, ISABELLE_INTERFACE, and
  3750 XSYMBOL_HOME; no longer need to do manual configuration in most
  3751 situations;
  3752 
  3753 * system: compression of ML heaps images may now be controlled via -c
  3754 option of isabelle and isatool usedir (currently only observed by
  3755 Poly/ML);
  3756 
  3757 * system: isatool installfonts may handle X-Symbol fonts as well (very
  3758 useful for remote X11);
  3759 
  3760 * system: provide TAGS file for Isabelle sources;
  3761 
  3762 * ML: infix 'OF' is a version of 'MRS' with more appropriate argument
  3763 order;
  3764 
  3765 * ML: renamed flags Syntax.trace_norm_ast to Syntax.trace_ast; global
  3766 timing flag supersedes proof_timing and Toplevel.trace;
  3767 
  3768 * ML: new combinators |>> and |>>> for incremental transformations
  3769 with secondary results (e.g. certain theory extensions):
  3770 
  3771 * ML: PureThy.add_defs gets additional argument to indicate potential
  3772 overloading (usually false);
  3773 
  3774 * ML: PureThy.add_thms/add_axioms/add_defs now return theorems as
  3775 results;
  3776 
  3777 
  3778 
  3779 New in Isabelle99 (October 1999)
  3780 --------------------------------
  3781 
  3782 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3783 
  3784 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)
  3785 are no longer simplified.  (This allows the simplifier to unfold recursive
  3786 functional programs.)  To restore the old behaviour, declare
  3787 
  3788     Delcongs [if_weak_cong];
  3789 
  3790 * HOL: Removed the obsolete syntax "Compl A"; use -A for set
  3791 complement;
  3792 
  3793 * HOL: the predicate "inj" is now defined by translation to "inj_on";
  3794 
  3795 * HOL/datatype: mutual_induct_tac no longer exists --
  3796   use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]
  3797 
  3798 * HOL/typedef: fixed type inference for representing set; type
  3799 arguments now have to occur explicitly on the rhs as type constraints;
  3800 
  3801 * ZF: The con_defs part of an inductive definition may no longer refer
  3802 to constants declared in the same theory;
  3803 
  3804 * HOL, ZF: the function mk_cases, generated by the inductive
  3805 definition package, has lost an argument.  To simplify its result, it
  3806 uses the default simpset instead of a supplied list of theorems.
  3807 
  3808 * HOL/List: the constructors of type list are now Nil and Cons;
  3809 
  3810 * Simplifier: the type of the infix ML functions
  3811         setSSolver addSSolver setSolver addSolver
  3812 is now  simpset * solver -> simpset  where `solver' is a new abstract type
  3813 for packaging solvers. A solver is created via
  3814         mk_solver: string -> (thm list -> int -> tactic) -> solver
  3815 where the string argument is only a comment.
  3816 
  3817 
  3818 *** Proof tools ***
  3819 
  3820 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a
  3821 decision procedure for linear arithmetic. Currently it is used for
  3822 types `nat', `int', and `real' in HOL (see below); it can, should and
  3823 will be instantiated for other types and logics as well.
  3824 
  3825 * The simplifier now accepts rewrite rules with flexible heads, eg
  3826      hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y
  3827   They are applied like any rule with a non-pattern lhs, i.e. by first-order
  3828   matching.
  3829 
  3830 
  3831 *** General ***
  3832 
  3833 * New Isabelle/Isar subsystem provides an alternative to traditional
  3834 tactical theorem proving; together with the ProofGeneral/isar user
  3835 interface it offers an interactive environment for developing human
  3836 readable proof documents (Isar == Intelligible semi-automated
  3837 reasoning); for further information see isatool doc isar-ref,
  3838 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/
  3839 
  3840 * improved and simplified presentation of theories: better HTML markup
  3841 (including colors), graph views in several sizes; isatool usedir now
  3842 provides a proper interface for user theories (via -P option); actual
  3843 document preparation based on (PDF)LaTeX is available as well (for
  3844 new-style theories only); see isatool doc system for more information;
  3845 
  3846 * native support for Proof General, both for classic Isabelle and
  3847 Isabelle/Isar;
  3848 
  3849 * ML function thm_deps visualizes dependencies of theorems and lemmas,
  3850 using the graph browser tool;
  3851 
  3852 * Isabelle manuals now also available as PDF;
  3853 
  3854 * theory loader rewritten from scratch (may not be fully
  3855 bug-compatible); old loadpath variable has been replaced by show_path,
  3856 add_path, del_path, reset_path functions; new operations such as
  3857 update_thy, touch_thy, remove_thy, use/update_thy_only (see also
  3858 isatool doc ref);
  3859 
  3860 * improved isatool install: option -k creates KDE application icon,
  3861 option -p DIR installs standalone binaries;
  3862 
  3863 * added ML_PLATFORM setting (useful for cross-platform installations);
  3864 more robust handling of platform specific ML images for SML/NJ;
  3865 
  3866 * the settings environment is now statically scoped, i.e. it is never
  3867 created again in sub-processes invoked from isabelle, isatool, or
  3868 Isabelle;
  3869 
  3870 * path element specification '~~' refers to '$ISABELLE_HOME';
  3871 
  3872 * in locales, the "assumes" and "defines" parts may be omitted if
  3873 empty;
  3874 
  3875 * new print_mode "xsymbols" for extended symbol support (e.g. genuine
  3876 long arrows);
  3877 
  3878 * new print_mode "HTML";
  3879 
  3880 * new flag show_tags controls display of tags of theorems (which are
  3881 basically just comments that may be attached by some tools);
  3882 
  3883 * Isamode 2.6 requires patch to accomodate change of Isabelle font
  3884 mode and goal output format:
  3885 
  3886 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el
  3887 244c244
  3888 <       (list (isa-getenv "ISABELLE") "-msymbols" logic-name)
  3889 ---
  3890 >       (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)
  3891 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el
  3892 181c181
  3893 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"
  3894 ---
  3895 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"
  3896 
  3897 * function bind_thms stores lists of theorems (cf. bind_thm);
  3898 
  3899 * new shorthand tactics ftac, eatac, datac, fatac;
  3900 
  3901 * qed (and friends) now accept "" as result name; in that case the
  3902 theorem is not stored, but proper checks and presentation of the
  3903 result still apply;
  3904 
  3905 * theorem database now also indexes constants "Trueprop", "all",
  3906 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;
  3907 
  3908 
  3909 *** HOL ***
  3910 
  3911 ** HOL arithmetic **
  3912 
  3913 * There are now decision procedures for linear arithmetic over nat and
  3914 int:
  3915 
  3916 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',
  3917 `+', `-', `Suc', `min', `max' and numerical constants; other subterms
  3918 are treated as atomic; subformulae not involving type `nat' or `int'
  3919 are ignored; quantified subformulae are ignored unless they are
  3920 positive universal or negative existential. The tactic has to be
  3921 invoked by hand and can be a little bit slow. In particular, the
  3922 running time is exponential in the number of occurrences of `min' and
  3923 `max', and `-' on `nat'.
  3924 
  3925 2. fast_arith_tac is a cut-down version of arith_tac: it only takes
  3926 (negated) (in)equalities among the premises and the conclusion into
  3927 account (i.e. no compound formulae) and does not know about `min' and
  3928 `max', and `-' on `nat'. It is fast and is used automatically by the
  3929 simplifier.
  3930 
  3931 NB: At the moment, these decision procedures do not cope with mixed
  3932 nat/int formulae where the two parts interact, such as `m < n ==>
  3933 int(m) < int(n)'.
  3934 
  3935 * HOL/Numeral provides a generic theory of numerals (encoded
  3936 efficiently as bit strings); setup for types nat/int/real is in place;
  3937 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than
  3938 int, existing theories and proof scripts may require a few additional
  3939 type constraints;
  3940 
  3941 * integer division and remainder can now be performed on constant
  3942 arguments;
  3943 
  3944 * many properties of integer multiplication, division and remainder
  3945 are now available;
  3946 
  3947 * An interface to the Stanford Validity Checker (SVC) is available through the
  3948 tactic svc_tac.  Propositional tautologies and theorems of linear arithmetic
  3949 are proved automatically.  SVC must be installed separately, and its results
  3950 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any
  3951 invocation of the underlying oracle).  For SVC see
  3952   http://verify.stanford.edu/SVC
  3953 
  3954 * IsaMakefile: the HOL-Real target now builds an actual image;
  3955 
  3956 
  3957 ** HOL misc **
  3958 
  3959 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces
  3960 (in Isabelle/Isar) -- by Gertrud Bauer;
  3961 
  3962 * HOL/BCV: generic model of bytecode verification, i.e. data-flow
  3963 analysis for assembly languages with subtypes;
  3964 
  3965 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization
  3966 -- avoids syntactic ambiguities and treats state, transition, and
  3967 temporal levels more uniformly; introduces INCOMPATIBILITIES due to
  3968 changed syntax and (many) tactics;
  3969 
  3970 * HOL/inductive: Now also handles more general introduction rules such
  3971   as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity
  3972   theorems are now maintained within the theory (maintained via the
  3973   "mono" attribute);
  3974 
  3975 * HOL/datatype: Now also handles arbitrarily branching datatypes
  3976   (using function types) such as
  3977 
  3978   datatype 'a tree = Atom 'a | Branch "nat => 'a tree"
  3979 
  3980 * HOL/record: record_simproc (part of the default simpset) takes care
  3981 of selectors applied to updated records; record_split_tac is no longer
  3982 part of the default claset; update_defs may now be removed from the
  3983 simpset in many cases; COMPATIBILITY: old behavior achieved by
  3984 
  3985   claset_ref () := claset() addSWrapper record_split_wrapper;
  3986   Delsimprocs [record_simproc]
  3987 
  3988 * HOL/typedef: fixed type inference for representing set; type
  3989 arguments now have to occur explicitly on the rhs as type constraints;
  3990 
  3991 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem
  3992 names rather than an ML expression;
  3993 
  3994 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be
  3995 supplied later.  Program schemes can be defined, such as
  3996     "While B C s = (if B s then While B C (C s) else s)"
  3997 where the well-founded relation can be chosen after B and C have been given.
  3998 
  3999 * HOL/List: the constructors of type list are now Nil and Cons;
  4000 INCOMPATIBILITY: while [] and infix # syntax is still there, of
  4001 course, ML tools referring to List.list.op # etc. have to be adapted;
  4002 
  4003 * HOL_quantifiers flag superseded by "HOL" print mode, which is
  4004 disabled by default; run isabelle with option -m HOL to get back to
  4005 the original Gordon/HOL-style output;
  4006 
  4007 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,
  4008 ALL x<=y. P, EX x<y. P, EX x<=y. P;
  4009 
  4010 * HOL basic syntax simplified (more orthogonal): all variants of
  4011 All/Ex now support plain / symbolic / HOL notation; plain syntax for
  4012 Eps operator is provided as well: "SOME x. P[x]";
  4013 
  4014 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;
  4015 
  4016 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made
  4017 thus available for user theories;
  4018 
  4019 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with
  4020 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the
  4021 time;
  4022 
  4023 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec
  4024 several times and then mp;
  4025 
  4026 
  4027 *** LK ***
  4028 
  4029 * the notation <<...>> is now available as a notation for sequences of
  4030 formulas;
  4031 
  4032 * the simplifier is now installed
  4033 
  4034 * the axiom system has been generalized (thanks to Soren Heilmann)
  4035 
  4036 * the classical reasoner now has a default rule database
  4037 
  4038 
  4039 *** ZF ***
  4040 
  4041 * new primrec section allows primitive recursive functions to be given
  4042 directly (as in HOL) over datatypes and the natural numbers;
  4043 
  4044 * new tactics induct_tac and exhaust_tac for induction (or case
  4045 analysis) over datatypes and the natural numbers;
  4046 
  4047 * the datatype declaration of type T now defines the recursor T_rec;
  4048 
  4049 * simplification automatically does freeness reasoning for datatype
  4050 constructors;
  4051 
  4052 * automatic type-inference, with AddTCs command to insert new
  4053 type-checking rules;
  4054 
  4055 * datatype introduction rules are now added as Safe Introduction rules
  4056 to the claset;
  4057 
  4058 * the syntax "if P then x else y" is now available in addition to
  4059 if(P,x,y);
  4060 
  4061 
  4062 *** Internal programming interfaces ***
  4063 
  4064 * tuned simplifier trace output; new flag debug_simp;
  4065 
  4066 * structures Vartab / Termtab (instances of TableFun) offer efficient
  4067 tables indexed by indexname_ord / term_ord (compatible with aconv);
  4068 
  4069 * AxClass.axclass_tac lost the theory argument;
  4070 
  4071 * tuned current_goals_markers semantics: begin / end goal avoids
  4072 printing empty lines;
  4073 
  4074 * removed prs and prs_fn hook, which was broken because it did not
  4075 include \n in its semantics, forcing writeln to add one
  4076 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:
  4077 string -> unit if you really want to output text without newline;
  4078 
  4079 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to
  4080 plain output, interface builders may have to enable 'isabelle_font'
  4081 mode to get Isabelle font glyphs as before;
  4082 
  4083 * refined token_translation interface; INCOMPATIBILITY: output length
  4084 now of type real instead of int;
  4085 
  4086 * theory loader actions may be traced via new ThyInfo.add_hook
  4087 interface (see src/Pure/Thy/thy_info.ML); example application: keep
  4088 your own database of information attached to *whole* theories -- as
  4089 opposed to intra-theory data slots offered via TheoryDataFun;
  4090 
  4091 * proper handling of dangling sort hypotheses (at last!);
  4092 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing
  4093 extra sort hypotheses that can be witnessed from the type signature;
  4094 the force_strip_shyps flag is gone, any remaining shyps are simply
  4095 left in the theorem (with a warning issued by strip_shyps_warning);
  4096 
  4097 
  4098 
  4099 New in Isabelle98-1 (October 1998)
  4100 ----------------------------------
  4101 
  4102 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  4103 
  4104 * several changes of automated proof tools;
  4105 
  4106 * HOL: major changes to the inductive and datatype packages, including
  4107 some minor incompatibilities of theory syntax;
  4108 
  4109 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now
  4110 called `inj_on';
  4111 
  4112 * HOL: removed duplicate thms in Arith:
  4113   less_imp_add_less  should be replaced by  trans_less_add1
  4114   le_imp_add_le      should be replaced by  trans_le_add1
  4115 
  4116 * HOL: unary minus is now overloaded (new type constraints may be
  4117 required);
  4118 
  4119 * HOL and ZF: unary minus for integers is now #- instead of #~.  In
  4120 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is
  4121 now taken as an integer constant.
  4122 
  4123 * Pure: ML function 'theory_of' renamed to 'theory';
  4124 
  4125 
  4126 *** Proof tools ***
  4127 
  4128 * Simplifier:
  4129   1. Asm_full_simp_tac is now more aggressive.
  4130      1. It will sometimes reorient premises if that increases their power to
  4131         simplify.
  4132      2. It does no longer proceed strictly from left to right but may also
  4133         rotate premises to achieve further simplification.
  4134      For compatibility reasons there is now Asm_lr_simp_tac which is like the
  4135      old Asm_full_simp_tac in that it does not rotate premises.
  4136   2. The simplifier now knows a little bit about nat-arithmetic.
  4137 
  4138 * Classical reasoner: wrapper mechanism for the classical reasoner now
  4139 allows for selected deletion of wrappers, by introduction of names for
  4140 wrapper functionals.  This implies that addbefore, addSbefore,
  4141 addaltern, and addSaltern now take a pair (name, tactic) as argument,
  4142 and that adding two tactics with the same name overwrites the first
  4143 one (emitting a warning).
  4144   type wrapper = (int -> tactic) -> (int -> tactic)
  4145   setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by
  4146   addWrapper, addSWrapper: claset * (string * wrapper) -> claset
  4147   delWrapper, delSWrapper: claset *  string            -> claset
  4148   getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;
  4149 
  4150 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE
  4151 semantics; addbefore now affects only the unsafe part of step_tac
  4152 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY
  4153 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac
  4154 by Force_tac;
  4155 
  4156 * Classical reasoner: setwrapper to setWrapper and compwrapper to
  4157 compWrapper; added safe wrapper (and access functions for it);
  4158 
  4159 * HOL/split_all_tac is now much faster and fails if there is nothing
  4160 to split.  Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order
  4161 and the names of the automatically generated variables have changed.
  4162 split_all_tac has moved within claset() from unsafe wrappers to safe
  4163 wrappers, which means that !!-bound variables are split much more
  4164 aggressively, and safe_tac and clarify_tac now split such variables.
  4165 If this splitting is not appropriate, use delSWrapper "split_all_tac".
  4166 Note: the same holds for record_split_tac, which does the job of
  4167 split_all_tac for record fields.
  4168 
  4169 * HOL/Simplifier: Rewrite rules for case distinctions can now be added
  4170 permanently to the default simpset using Addsplits just like
  4171 Addsimps. They can be removed via Delsplits just like
  4172 Delsimps. Lower-case versions are also available.
  4173 
  4174 * HOL/Simplifier: The rule split_if is now part of the default
  4175 simpset. This means that the simplifier will eliminate all occurrences
  4176 of if-then-else in the conclusion of a goal. To prevent this, you can
  4177 either remove split_if completely from the default simpset by
  4178 `Delsplits [split_if]' or remove it in a specific call of the
  4179 simplifier using `... delsplits [split_if]'.  You can also add/delete
  4180 other case splitting rules to/from the default simpset: every datatype
  4181 generates suitable rules `split_t_case' and `split_t_case_asm' (where
  4182 t is the name of the datatype).
  4183 
  4184 * Classical reasoner / Simplifier combination: new force_tac (and
  4185 derivatives Force_tac, force) combines rewriting and classical
  4186 reasoning (and whatever other tools) similarly to auto_tac, but is
  4187 aimed to solve the given subgoal completely.
  4188 
  4189 
  4190 *** General ***
  4191 
  4192 * new top-level commands `Goal' and `Goalw' that improve upon `goal'
  4193 and `goalw': the theory is no longer needed as an explicit argument -
  4194 the current theory context is used; assumptions are no longer returned
  4195 at the ML-level unless one of them starts with ==> or !!; it is
  4196 recommended to convert to these new commands using isatool fixgoal
  4197 (backup your sources first!);
  4198 
  4199 * new top-level commands 'thm' and 'thms' for retrieving theorems from
  4200 the current theory context, and 'theory' to lookup stored theories;
  4201 
  4202 * new theory section 'locale' for declaring constants, assumptions and
  4203 definitions that have local scope;
  4204 
  4205 * new theory section 'nonterminals' for purely syntactic types;
  4206 
  4207 * new theory section 'setup' for generic ML setup functions
  4208 (e.g. package initialization);
  4209 
  4210 * the distribution now includes Isabelle icons: see
  4211 lib/logo/isabelle-{small,tiny}.xpm;
  4212 
  4213 * isatool install - install binaries with absolute references to
  4214 ISABELLE_HOME/bin;
  4215 
  4216 * isatool logo -- create instances of the Isabelle logo (as EPS);
  4217 
  4218 * print mode 'emacs' reserved for Isamode;
  4219 
  4220 * support multiple print (ast) translations per constant name;
  4221 
  4222 * theorems involving oracles are now printed with a suffixed [!];
  4223 
  4224 
  4225 *** HOL ***
  4226 
  4227 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');
  4228 
  4229 * HOL/inductive package reorganized and improved: now supports mutual
  4230 definitions such as
  4231 
  4232   inductive EVEN ODD
  4233     intrs
  4234       null "0 : EVEN"
  4235       oddI "n : EVEN ==> Suc n : ODD"
  4236       evenI "n : ODD ==> Suc n : EVEN"
  4237 
  4238 new theorem list "elims" contains an elimination rule for each of the
  4239 recursive sets; inductive definitions now handle disjunctive premises
  4240 correctly (also ZF);
  4241 
  4242 INCOMPATIBILITIES: requires Inductive as an ancestor; component
  4243 "mutual_induct" no longer exists - the induction rule is always
  4244 contained in "induct";
  4245 
  4246 
  4247 * HOL/datatype package re-implemented and greatly improved: now
  4248 supports mutually recursive datatypes such as
  4249 
  4250   datatype
  4251     'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)
  4252             | SUM ('a aexp) ('a aexp)
  4253             | DIFF ('a aexp) ('a aexp)
  4254             | NUM 'a
  4255   and
  4256     'a bexp = LESS ('a aexp) ('a aexp)
  4257             | AND ('a bexp) ('a bexp)
  4258             | OR ('a bexp) ('a bexp)
  4259 
  4260 as well as indirectly recursive datatypes such as
  4261 
  4262   datatype
  4263     ('a, 'b) term = Var 'a
  4264                   | App 'b ((('a, 'b) term) list)
  4265 
  4266 The new tactic  mutual_induct_tac [<var_1>, ..., <var_n>] i  performs
  4267 induction on mutually / indirectly recursive datatypes.
  4268 
  4269 Primrec equations are now stored in theory and can be accessed via
  4270 <function_name>.simps.
  4271 
  4272 INCOMPATIBILITIES:
  4273 
  4274   - Theories using datatypes must now have theory Datatype as an
  4275     ancestor.
  4276   - The specific <typename>.induct_tac no longer exists - use the
  4277     generic induct_tac instead.
  4278   - natE has been renamed to nat.exhaust - use exhaust_tac
  4279     instead of res_inst_tac ... natE. Note that the variable
  4280     names in nat.exhaust differ from the names in natE, this
  4281     may cause some "fragile" proofs to fail.
  4282   - The theorems split_<typename>_case and split_<typename>_case_asm
  4283     have been renamed to <typename>.split and <typename>.split_asm.
  4284   - Since default sorts of type variables are now handled correctly,
  4285     some datatype definitions may have to be annotated with explicit
  4286     sort constraints.
  4287   - Primrec definitions no longer require function name and type
  4288     of recursive argument.
  4289 
  4290 Consider using isatool fixdatatype to adapt your theories and proof
  4291 scripts to the new package (backup your sources first!).
  4292 
  4293 
  4294 * HOL/record package: considerably improved implementation; now
  4295 includes concrete syntax for record types, terms, updates; theorems
  4296 for surjective pairing and splitting !!-bound record variables; proof
  4297 support is as follows:
  4298 
  4299   1) standard conversions (selectors or updates applied to record
  4300 constructor terms) are part of the standard simpset;
  4301 
  4302   2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are
  4303 made part of standard simpset and claset via addIffs;
  4304 
  4305   3) a tactic for record field splitting (record_split_tac) is part of
  4306 the standard claset (addSWrapper);
  4307 
  4308 To get a better idea about these rules you may retrieve them via
  4309 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is
  4310 the name of your record type.
  4311 
  4312 The split tactic 3) conceptually simplifies by the following rule:
  4313 
  4314   "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"
  4315 
  4316 Thus any record variable that is bound by meta-all will automatically
  4317 blow up into some record constructor term, consequently the
  4318 simplifications of 1), 2) apply.  Thus force_tac, auto_tac etc. shall
  4319 solve record problems automatically.
  4320 
  4321 
  4322 * reorganized the main HOL image: HOL/Integ and String loaded by
  4323 default; theory Main includes everything;
  4324 
  4325 * automatic simplification of integer sums and comparisons, using cancellation;
  4326 
  4327 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;
  4328 
  4329 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;
  4330 
  4331 * many new identities for unions, intersections, set difference, etc.;
  4332 
  4333 * expand_if, expand_split, expand_sum_case and expand_nat_case are now
  4334 called split_if, split_split, split_sum_case and split_nat_case (to go
  4335 with add/delsplits);
  4336 
  4337 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting
  4338 (?x::unit) = (); this is made part of the default simpset, which COULD
  4339 MAKE EXISTING PROOFS FAIL under rare circumstances (consider
  4340 'Delsimprocs [unit_eq_proc];' as last resort); also note that
  4341 unit_abs_eta_conv is added in order to counter the effect of
  4342 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by
  4343 %u.f();
  4344 
  4345 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which
  4346 makes more sense);
  4347 
  4348 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4349   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4350   disjointness reasoning but breaking a few old proofs.
  4351 
  4352 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1
  4353 to 'converse' from 'inverse' (for compatibility with ZF and some
  4354 literature);
  4355 
  4356 * HOL/recdef can now declare non-recursive functions, with {} supplied as
  4357 the well-founded relation;
  4358 
  4359 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of
  4360     Compl A.  The "Compl" syntax remains available as input syntax for this
  4361     release ONLY.
  4362 
  4363 * HOL/Update: new theory of function updates:
  4364     f(a:=b) == %x. if x=a then b else f x
  4365 may also be iterated as in f(a:=b,c:=d,...);
  4366 
  4367 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;
  4368 
  4369 * HOL/List:
  4370   - new function list_update written xs[i:=v] that updates the i-th
  4371     list position. May also be iterated as in xs[i:=a,j:=b,...].
  4372   - new function `upt' written [i..j(] which generates the list
  4373     [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper
  4374     bound write [i..j], which is a shorthand for [i..j+1(].
  4375   - new lexicographic orderings and corresponding wellfoundedness theorems.
  4376 
  4377 * HOL/Arith:
  4378   - removed 'pred' (predecessor) function;
  4379   - generalized some theorems about n-1;
  4380   - many new laws about "div" and "mod";
  4381   - new laws about greatest common divisors (see theory ex/Primes);
  4382 
  4383 * HOL/Relation: renamed the relational operator r^-1 "converse"
  4384 instead of "inverse";
  4385 
  4386 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness
  4387   of the multiset ordering;
  4388 
  4389 * directory HOL/Real: a construction of the reals using Dedekind cuts
  4390   (not included by default);
  4391 
  4392 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;
  4393 
  4394 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted
  4395   programs, i.e. different program variables may have different types.
  4396 
  4397 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4398   [previously, this check worked only if the rewrite rule was unconditional]
  4399   Now rew can involve either definitions or equalities (either == or =).
  4400 
  4401 
  4402 *** ZF ***
  4403 
  4404 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains
  4405   only the theorems proved on ZF.ML;
  4406 
  4407 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4408   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4409   disjointness reasoning but breaking a few old proofs.
  4410 
  4411 * ZF/Update: new theory of function updates
  4412     with default rewrite rule  f(x:=y) ` z = if(z=x, y, f`z)
  4413   may also be iterated as in f(a:=b,c:=d,...);
  4414 
  4415 * in  let x=t in u(x), neither t nor u(x) has to be an FOL term.
  4416 
  4417 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4418   [previously, this check worked only if the rewrite rule was unconditional]
  4419   Now rew can involve either definitions or equalities (either == or =).
  4420 
  4421 * case_tac provided for compatibility with HOL
  4422     (like the old excluded_middle_tac, but with subgoals swapped)
  4423 
  4424 
  4425 *** Internal programming interfaces ***
  4426 
  4427 * Pure: several new basic modules made available for general use, see
  4428 also src/Pure/README;
  4429 
  4430 * improved the theory data mechanism to support encapsulation (data
  4431 kind name replaced by private Object.kind, acting as authorization
  4432 key); new type-safe user interface via functor TheoryDataFun; generic
  4433 print_data function becomes basically useless;
  4434 
  4435 * removed global_names compatibility flag -- all theory declarations
  4436 are qualified by default;
  4437 
  4438 * module Pure/Syntax now offers quote / antiquote translation
  4439 functions (useful for Hoare logic etc. with implicit dependencies);
  4440 see HOL/ex/Antiquote for an example use;
  4441 
  4442 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->
  4443 cterm -> thm;
  4444 
  4445 * new tactical CHANGED_GOAL for checking that a tactic modifies a
  4446 subgoal;
  4447 
  4448 * Display.print_goals function moved to Locale.print_goals;
  4449 
  4450 * standard print function for goals supports current_goals_markers
  4451 variable for marking begin of proof, end of proof, start of goal; the
  4452 default is ("", "", ""); setting current_goals_markers := ("<proof>",
  4453 "</proof>", "<goal>") causes SGML like tagged proof state printing,
  4454 for example;
  4455 
  4456 
  4457 
  4458 New in Isabelle98 (January 1998)
  4459 --------------------------------
  4460 
  4461 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  4462 
  4463 * changed lexical syntax of terms / types: dots made part of long
  4464 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";
  4465 
  4466 * simpset (and claset) reference variable replaced by functions
  4467 simpset / simpset_ref;
  4468 
  4469 * no longer supports theory aliases (via merge) and non-trivial
  4470 implicit merge of thms' signatures;
  4471 
  4472 * most internal names of constants changed due to qualified names;
  4473 
  4474 * changed Pure/Sequence interface (see Pure/seq.ML);
  4475 
  4476 
  4477 *** General Changes ***
  4478 
  4479 * hierachically structured name spaces (for consts, types, axms, thms
  4480 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of
  4481 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:
  4482 isatool fixdots ensures space after dots (e.g. "%x. x"); set
  4483 long_names for fully qualified output names; NOTE: ML programs
  4484 (special tactics, packages etc.) referring to internal names may have
  4485 to be adapted to cope with fully qualified names; in case of severe
  4486 backward campatibility problems try setting 'global_names' at compile
  4487 time to have enrything declared within a flat name space; one may also
  4488 fine tune name declarations in theories via the 'global' and 'local'
  4489 section;
  4490 
  4491 * reimplemented the implicit simpset and claset using the new anytype
  4492 data filed in signatures; references simpset:simpset ref etc. are
  4493 replaced by functions simpset:unit->simpset and
  4494 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp
  4495 to patch your ML files accordingly;
  4496 
  4497 * HTML output now includes theory graph data for display with Java
  4498 applet or isatool browser; data generated automatically via isatool
  4499 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);
  4500 
  4501 * defs may now be conditional; improved rewrite_goals_tac to handle
  4502 conditional equations;
  4503 
  4504 * defs now admits additional type arguments, using TYPE('a) syntax;
  4505 
  4506 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always
  4507 creates a new theory node; implicit merge of thms' signatures is
  4508 restricted to 'trivial' ones; COMPATIBILITY: one may have to use
  4509 transfer:theory->thm->thm in (rare) cases;
  4510 
  4511 * improved handling of draft signatures / theories; draft thms (and
  4512 ctyps, cterms) are automatically promoted to real ones;
  4513 
  4514 * slightly changed interfaces for oracles: admit many per theory, named
  4515 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;
  4516 
  4517 * print_goals: optional output of const types (set show_consts and
  4518 show_types);
  4519 
  4520 * improved output of warnings (###) and errors (***);
  4521 
  4522 * subgoal_tac displays a warning if the new subgoal has type variables;
  4523 
  4524 * removed old README and Makefiles;
  4525 
  4526 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;
  4527 
  4528 * removed obsolete init_pps and init_database;
  4529 
  4530 * deleted the obsolete tactical STATE, which was declared by
  4531     fun STATE tacfun st = tacfun st st;
  4532 
  4533 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~
  4534 (which abbreviates $HOME);
  4535 
  4536 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:
  4537 use isatool fixseq to adapt your ML programs (this works for fully
  4538 qualified references to the Sequence structure only!);
  4539 
  4540 * use_thy no longer requires writable current directory; it always
  4541 reloads .ML *and* .thy file, if either one is out of date;
  4542 
  4543 
  4544 *** Classical Reasoner ***
  4545 
  4546 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new
  4547 tactics that use classical reasoning to simplify a subgoal without
  4548 splitting it into several subgoals;
  4549 
  4550 * Safe_tac: like safe_tac but uses the default claset;
  4551 
  4552 
  4553 *** Simplifier ***
  4554 
  4555 * added simplification meta rules:
  4556     (asm_)(full_)simplify: simpset -> thm -> thm;
  4557 
  4558 * simplifier.ML no longer part of Pure -- has to be loaded by object
  4559 logics (again);
  4560 
  4561 * added prems argument to simplification procedures;
  4562 
  4563 * HOL, FOL, ZF: added infix function `addsplits':
  4564   instead of `<simpset> setloop (split_tac <thms>)'
  4565   you can simply write `<simpset> addsplits <thms>'
  4566 
  4567 
  4568 *** Syntax ***
  4569 
  4570 * TYPE('a) syntax for type reflection terms;
  4571 
  4572 * no longer handles consts with name "" -- declare as 'syntax' instead;
  4573 
  4574 * pretty printer: changed order of mixfix annotation preference (again!);
  4575 
  4576 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;
  4577 
  4578 
  4579 *** HOL ***
  4580 
  4581 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4582   with `addloop' of the simplifier to faciliate case splitting in premises.
  4583 
  4584 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;
  4585 
  4586 * HOL/Auth: new protocol proofs including some for the Internet
  4587   protocol TLS;
  4588 
  4589 * HOL/Map: new theory of `maps' a la VDM;
  4590 
  4591 * HOL/simplifier: simplification procedures nat_cancel_sums for
  4592 cancelling out common nat summands from =, <, <= (in)equalities, or
  4593 differences; simplification procedures nat_cancel_factor for
  4594 cancelling common factor from =, <, <= (in)equalities over natural
  4595 sums; nat_cancel contains both kinds of procedures, it is installed by
  4596 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;
  4597 
  4598 * HOL/simplifier: terms of the form
  4599   `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)'  (or t=x)
  4600   are rewritten to
  4601   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',
  4602   and those of the form
  4603   `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)'  (or t=x)
  4604   are rewritten to
  4605   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',
  4606 
  4607 * HOL/datatype
  4608   Each datatype `t' now comes with a theorem `split_t_case' of the form
  4609 
  4610   P(t_case f1 ... fn x) =
  4611      ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &
  4612         ...
  4613        (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))
  4614      )
  4615 
  4616   and a theorem `split_t_case_asm' of the form
  4617 
  4618   P(t_case f1 ... fn x) =
  4619     ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |
  4620         ...
  4621        (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))
  4622      )
  4623   which can be added to a simpset via `addsplits'. The existing theorems
  4624   expand_list_case and expand_option_case have been renamed to
  4625   split_list_case and split_option_case.
  4626 
  4627 * HOL/Arithmetic:
  4628   - `pred n' is automatically converted to `n-1'.
  4629     Users are strongly encouraged not to use `pred' any longer,
  4630     because it will disappear altogether at some point.
  4631   - Users are strongly encouraged to write "0 < n" rather than
  4632     "n ~= 0". Theorems and proof tools have been modified towards this
  4633     `standard'.
  4634 
  4635 * HOL/Lists:
  4636   the function "set_of_list" has been renamed "set" (and its theorems too);
  4637   the function "nth" now takes its arguments in the reverse order and
  4638   has acquired the infix notation "!" as in "xs!n".
  4639 
  4640 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};
  4641 
  4642 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its
  4643   specialist theorems (like UN1_I) are gone.  Similarly for (INT x.B x);
  4644 
  4645 * HOL/record: extensible records with schematic structural subtyping
  4646 (single inheritance); EXPERIMENTAL version demonstrating the encoding,
  4647 still lacks various theorems and concrete record syntax;
  4648 
  4649 
  4650 *** HOLCF ***
  4651 
  4652 * removed "axioms" and "generated by" sections;
  4653 
  4654 * replaced "ops" section by extended "consts" section, which is capable of
  4655   handling the continuous function space "->" directly;
  4656 
  4657 * domain package:
  4658   . proves theorems immediately and stores them in the theory,
  4659   . creates hierachical name space,
  4660   . now uses normal mixfix annotations (instead of cinfix...),
  4661   . minor changes to some names and values (for consistency),
  4662   . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,
  4663   . separator between mutual domain defs: changed "," to "and",
  4664   . improved handling of sort constraints;  now they have to
  4665     appear on the left-hand side of the equations only;
  4666 
  4667 * fixed LAM <x,y,zs>.b syntax;
  4668 
  4669 * added extended adm_tac to simplifier in HOLCF -- can now discharge
  4670 adm (%x. P (t x)), where P is chainfinite and t continuous;
  4671 
  4672 
  4673 *** FOL and ZF ***
  4674 
  4675 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4676   with `addloop' of the simplifier to faciliate case splitting in premises.
  4677 
  4678 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as
  4679 in HOL, they strip ALL and --> from proved theorems;
  4680 
  4681 
  4682 
  4683 New in Isabelle94-8 (May 1997)
  4684 ------------------------------
  4685 
  4686 *** General Changes ***
  4687 
  4688 * new utilities to build / run / maintain Isabelle etc. (in parts
  4689 still somewhat experimental); old Makefiles etc. still functional;
  4690 
  4691 * new 'Isabelle System Manual';
  4692 
  4693 * INSTALL text, together with ./configure and ./build scripts;
  4694 
  4695 * reimplemented type inference for greater efficiency, better error
  4696 messages and clean internal interface;
  4697 
  4698 * prlim command for dealing with lots of subgoals (an easier way of
  4699 setting goals_limit);
  4700 
  4701 
  4702 *** Syntax ***
  4703 
  4704 * supports alternative (named) syntax tables (parser and pretty
  4705 printer); internal interface is provided by add_modesyntax(_i);
  4706 
  4707 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to
  4708 be used in conjunction with the Isabelle symbol font; uses the
  4709 "symbols" syntax table;
  4710 
  4711 * added token_translation interface (may translate name tokens in
  4712 arbitrary ways, dependent on their type (free, bound, tfree, ...) and
  4713 the current print_mode); IMPORTANT: user print translation functions
  4714 are responsible for marking newly introduced bounds
  4715 (Syntax.mark_boundT);
  4716 
  4717 * token translations for modes "xterm" and "xterm_color" that display
  4718 names in bold, underline etc. or colors (which requires a color
  4719 version of xterm);
  4720 
  4721 * infixes may now be declared with names independent of their syntax;
  4722 
  4723 * added typed_print_translation (like print_translation, but may
  4724 access type of constant);
  4725 
  4726 
  4727 *** Classical Reasoner ***
  4728 
  4729 Blast_tac: a new tactic!  It is often more powerful than fast_tac, but has
  4730 some limitations.  Blast_tac...
  4731   + ignores addss, addbefore, addafter; this restriction is intrinsic
  4732   + ignores elimination rules that don't have the correct format
  4733         (the conclusion MUST be a formula variable)
  4734   + ignores types, which can make HOL proofs fail
  4735   + rules must not require higher-order unification, e.g. apply_type in ZF
  4736     [message "Function Var's argument not a bound variable" relates to this]
  4737   + its proof strategy is more general but can actually be slower
  4738 
  4739 * substitution with equality assumptions no longer permutes other
  4740 assumptions;
  4741 
  4742 * minor changes in semantics of addafter (now called addaltern); renamed
  4743 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper
  4744 (and access functions for it);
  4745 
  4746 * improved combination of classical reasoner and simplifier:
  4747   + functions for handling clasimpsets
  4748   + improvement of addss: now the simplifier is called _after_ the
  4749     safe steps.
  4750   + safe variant of addss called addSss: uses safe simplifications
  4751     _during_ the safe steps. It is more complete as it allows multiple
  4752     instantiations of unknowns (e.g. with slow_tac).
  4753 
  4754 *** Simplifier ***
  4755 
  4756 * added interface for simplification procedures (functions that
  4757 produce *proven* rewrite rules on the fly, depending on current
  4758 redex);
  4759 
  4760 * ordering on terms as parameter (used for ordered rewriting);
  4761 
  4762 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;
  4763 
  4764 * the solver is now split into a safe and an unsafe part.
  4765 This should be invisible for the normal user, except that the
  4766 functions setsolver and addsolver have been renamed to setSolver and
  4767 addSolver; added safe_asm_full_simp_tac;
  4768 
  4769 
  4770 *** HOL ***
  4771 
  4772 * a generic induction tactic `induct_tac' which works for all datatypes and
  4773 also for type `nat';
  4774 
  4775 * a generic case distinction tactic `exhaust_tac' which works for all
  4776 datatypes and also for type `nat';
  4777 
  4778 * each datatype comes with a function `size';
  4779 
  4780 * patterns in case expressions allow tuple patterns as arguments to
  4781 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';
  4782 
  4783 * primrec now also works with type nat;
  4784 
  4785 * recdef: a new declaration form, allows general recursive functions to be
  4786 defined in theory files.  See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.
  4787 
  4788 * the constant for negation has been renamed from "not" to "Not" to
  4789 harmonize with FOL, ZF, LK, etc.;
  4790 
  4791 * HOL/ex/LFilter theory of a corecursive "filter" functional for
  4792 infinite lists;
  4793 
  4794 * HOL/Modelcheck demonstrates invocation of model checker oracle;
  4795 
  4796 * HOL/ex/Ring.thy declares cring_simp, which solves equational
  4797 problems in commutative rings, using axiomatic type classes for + and *;
  4798 
  4799 * more examples in HOL/MiniML and HOL/Auth;
  4800 
  4801 * more default rewrite rules for quantifiers, union/intersection;
  4802 
  4803 * a new constant `arbitrary == @x.False';
  4804 
  4805 * HOLCF/IOA replaces old HOL/IOA;
  4806 
  4807 * HOLCF changes: derived all rules and arities
  4808   + axiomatic type classes instead of classes
  4809   + typedef instead of faking type definitions
  4810   + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.
  4811   + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po
  4812   + eliminated the types void, one, tr
  4813   + use unit lift and bool lift (with translations) instead of one and tr
  4814   + eliminated blift from Lift3.thy (use Def instead of blift)
  4815   all eliminated rules are derived as theorems --> no visible changes ;
  4816 
  4817 
  4818 *** ZF ***
  4819 
  4820 * ZF now has Fast_tac, Simp_tac and Auto_tac.  Union_iff is a now a default
  4821 rewrite rule; this may affect some proofs.  eq_cs is gone but can be put back
  4822 as ZF_cs addSIs [equalityI];
  4823 
  4824 
  4825 
  4826 New in Isabelle94-7 (November 96)
  4827 ---------------------------------
  4828 
  4829 * allowing negative levels (as offsets) in prlev and choplev;
  4830 
  4831 * super-linear speedup for large simplifications;
  4832 
  4833 * FOL, ZF and HOL now use miniscoping: rewriting pushes
  4834 quantifications in as far as possible (COULD MAKE EXISTING PROOFS
  4835 FAIL); can suppress it using the command Delsimps (ex_simps @
  4836 all_simps); De Morgan laws are also now included, by default;
  4837 
  4838 * improved printing of ==>  :  ~:
  4839 
  4840 * new object-logic "Sequents" adds linear logic, while replacing LK
  4841 and Modal (thanks to Sara Kalvala);
  4842 
  4843 * HOL/Auth: correctness proofs for authentication protocols;
  4844 
  4845 * HOL: new auto_tac combines rewriting and classical reasoning (many
  4846 examples on HOL/Auth);
  4847 
  4848 * HOL: new command AddIffs for declaring theorems of the form P=Q to
  4849 the rewriter and classical reasoner simultaneously;
  4850 
  4851 * function uresult no longer returns theorems in "standard" format;
  4852 regain previous version by: val uresult = standard o uresult;
  4853 
  4854 
  4855 
  4856 New in Isabelle94-6
  4857 -------------------
  4858 
  4859 * oracles -- these establish an interface between Isabelle and trusted
  4860 external reasoners, which may deliver results as theorems;
  4861 
  4862 * proof objects (in particular record all uses of oracles);
  4863 
  4864 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;
  4865 
  4866 * "constdefs" section in theory files;
  4867 
  4868 * "primrec" section (HOL) no longer requires names;
  4869 
  4870 * internal type "tactic" now simply "thm -> thm Sequence.seq";
  4871 
  4872 
  4873 
  4874 New in Isabelle94-5
  4875 -------------------
  4876 
  4877 * reduced space requirements;
  4878 
  4879 * automatic HTML generation from theories;
  4880 
  4881 * theory files no longer require "..." (quotes) around most types;
  4882 
  4883 * new examples, including two proofs of the Church-Rosser theorem;
  4884 
  4885 * non-curried (1994) version of HOL is no longer distributed;
  4886 
  4887 
  4888 
  4889 New in Isabelle94-4
  4890 -------------------
  4891 
  4892 * greatly reduced space requirements;
  4893 
  4894 * theory files (.thy) no longer require \...\ escapes at line breaks;
  4895 
  4896 * searchable theorem database (see the section "Retrieving theorems" on
  4897 page 8 of the Reference Manual);
  4898 
  4899 * new examples, including Grabczewski's monumental case study of the
  4900 Axiom of Choice;
  4901 
  4902 * The previous version of HOL renamed to Old_HOL;
  4903 
  4904 * The new version of HOL (previously called CHOL) uses a curried syntax
  4905 for functions.  Application looks like f a b instead of f(a,b);
  4906 
  4907 * Mutually recursive inductive definitions finally work in HOL;
  4908 
  4909 * In ZF, pattern-matching on tuples is now available in all abstractions and
  4910 translates to the operator "split";
  4911 
  4912 
  4913 
  4914 New in Isabelle94-3
  4915 -------------------
  4916 
  4917 * new infix operator, addss, allowing the classical reasoner to
  4918 perform simplification at each step of its search.  Example:
  4919         fast_tac (cs addss ss)
  4920 
  4921 * a new logic, CHOL, the same as HOL, but with a curried syntax
  4922 for functions.  Application looks like f a b instead of f(a,b).  Also pairs
  4923 look like (a,b) instead of <a,b>;
  4924 
  4925 * PLEASE NOTE: CHOL will eventually replace HOL!
  4926 
  4927 * In CHOL, pattern-matching on tuples is now available in all abstractions.
  4928 It translates to the operator "split".  A new theory of integers is available;
  4929 
  4930 * In ZF, integer numerals now denote two's-complement binary integers.
  4931 Arithmetic operations can be performed by rewriting.  See ZF/ex/Bin.ML;
  4932 
  4933 * Many new examples: I/O automata, Church-Rosser theorem, equivalents
  4934 of the Axiom of Choice;
  4935 
  4936 
  4937 
  4938 New in Isabelle94-2
  4939 -------------------
  4940 
  4941 * Significantly faster resolution;
  4942 
  4943 * the different sections in a .thy file can now be mixed and repeated
  4944 freely;
  4945 
  4946 * Database of theorems for FOL, HOL and ZF.  New
  4947 commands including qed, qed_goal and bind_thm store theorems in the database.
  4948 
  4949 * Simple database queries: return a named theorem (get_thm) or all theorems of
  4950 a given theory (thms_of), or find out what theory a theorem was proved in
  4951 (theory_of_thm);
  4952 
  4953 * Bugs fixed in the inductive definition and datatype packages;
  4954 
  4955 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs
  4956 and HOL_dup_cs obsolete;
  4957 
  4958 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1
  4959 have been removed;
  4960 
  4961 * Simpler definition of function space in ZF;
  4962 
  4963 * new results about cardinal and ordinal arithmetic in ZF;
  4964 
  4965 * 'subtype' facility in HOL for introducing new types as subsets of existing
  4966 types;
  4967 
  4968 
  4969 $Id$