NEWS
author wenzelm
Sat Apr 08 22:51:06 2006 +0200 (2006-04-08)
changeset 19363 667b5ea637dd
parent 19279 48b527d0331b
child 19377 1f717bd6b7ea
permissions -rw-r--r--
refined 'abbreviation';
     1 Isabelle NEWS -- history user-relevant changes
     2 ==============================================
     3 
     4 New in this Isabelle release
     5 ----------------------------
     6 
     7 *** General ***
     8 
     9 * Theory syntax: the header format ``theory A = B + C:'' has been
    10 discontinued in favour of ``theory A imports B C begin''.  Use isatool
    11 fixheaders to convert existing theory files.  INCOMPATIBILITY.
    12 
    13 * Theory syntax: the old non-Isar theory file format has been
    14 discontinued altogether.  Note that ML proof scripts may still be used
    15 with Isar theories; migration is usually quite simple with the ML
    16 function use_legacy_bindings.  INCOMPATIBILITY.
    17 
    18 * Legacy goal package: reduced interface to the bare minimum required
    19 to keep existing proof scripts running.  Most other user-level
    20 functions are now part of the OldGoals structure, which is *not* open
    21 by default (consider isatool expandshort before open OldGoals).
    22 Removed top_sg, prin, printyp, pprint_term/typ altogether, because
    23 these tend to cause confusion about the actual goal (!) context being
    24 used here, which is not necessarily the same as the_context().
    25 
    26 * Command 'find_theorems': support "*" wildcard in "name:" criterion.
    27 
    28 
    29 *** Document preparation ***
    30 
    31 * Added antiquotations @{ML_type text} and @{ML_struct text} which
    32 check the given source text as ML type/structure, printing verbatim.
    33 
    34 
    35 *** Pure ***
    36 
    37 * Command 'no_translations' removes translation rules from theory
    38 syntax.
    39 
    40 * Isar: improper proof element 'guess' is like 'obtain', but derives
    41 the obtained context from the course of reasoning!  For example:
    42 
    43   assume "EX x y. A x & B y"   -- "any previous fact"
    44   then guess x and y by clarify
    45 
    46 This technique is potentially adventurous, depending on the facts and
    47 proof tools being involved here.
    48 
    49 * Isar: known facts from the proof context may be specified as literal
    50 propositions, using ASCII back-quote syntax.  This works wherever
    51 named facts used to be allowed so far, in proof commands, proof
    52 methods, attributes etc.  Literal facts are retrieved from the context
    53 according to unification of type and term parameters.  For example,
    54 provided that "A" and "A ==> B" and "!!x. P x ==> Q x" are known
    55 theorems in the current context, then these are valid literal facts:
    56 `A` and `A ==> B` and `!!x. P x ==> Q x" as well as `P a ==> Q a` etc.
    57 
    58 There is also a proof method "fact" which does the same composition
    59 for explicit goal states, e.g. the following proof texts coincide with
    60 certain special cases of literal facts:
    61 
    62   have "A" by fact                 ==  note `A`
    63   have "A ==> B" by fact           ==  note `A ==> B`
    64   have "!!x. P x ==> Q x" by fact  ==  note `!!x. P x ==> Q x`
    65   have "P a ==> Q a" by fact       ==  note `P a ==> Q a`
    66 
    67 * Isar: 'def' now admits simultaneous definitions, e.g.:
    68 
    69   def x == "t" and y == "u"
    70 
    71 * Isar: added command 'unfolding', which is structurally similar to
    72 'using', but affects both the goal state and facts by unfolding given
    73 rewrite rules.  Thus many occurrences of the 'unfold' method or
    74 'unfolded' attribute may be replaced by first-class proof text.
    75 
    76 * Isar: methods 'unfold' / 'fold', attributes 'unfolded' / 'folded',
    77 and command 'unfolding' now all support object-level equalities
    78 (potentially conditional).  The underlying notion of rewrite rule is
    79 analogous to the 'rule_format' attribute, but *not* that of the
    80 Simplifier (which is usually more generous).
    81 
    82 * Isar: the goal restriction operator [N] (default N = 1) evaluates a
    83 method expression within a sandbox consisting of the first N
    84 sub-goals, which need to exist.  For example, ``simp_all [3]''
    85 simplifies the first three sub-goals, while (rule foo, simp_all)[]
    86 simplifies all new goals that emerge from applying rule foo to the
    87 originally first one.
    88 
    89 * Isar: the conclusion of a long theorem statement is now either
    90 'shows' (a simultaneous conjunction, as before), or 'obtains'
    91 (essentially a disjunction of cases with local parameters and
    92 assumptions).  The latter allows to express general elimination rules
    93 adequately; in this notation common elimination rules look like this:
    94 
    95   lemma exE:    -- "EX x. P x ==> (!!x. P x ==> thesis) ==> thesis"
    96     assumes "EX x. P x"
    97     obtains x where "P x"
    98 
    99   lemma conjE:  -- "A & B ==> (A ==> B ==> thesis) ==> thesis"
   100     assumes "A & B"
   101     obtains A and B
   102 
   103   lemma disjE:  -- "A | B ==> (A ==> thesis) ==> (B ==> thesis) ==> thesis"
   104     assumes "A | B"
   105     obtains
   106       A
   107     | B
   108 
   109 The subsequent classical rules even refer to the formal "thesis"
   110 explicitly:
   111 
   112   lemma classical:     -- "(~ thesis ==> thesis) ==> thesis"
   113     obtains "~ thesis"
   114 
   115   lemma Peirce's_Law:  -- "((thesis ==> something) ==> thesis) ==> thesis"
   116     obtains "thesis ==> something"
   117 
   118 The actual proof of an 'obtains' statement is analogous to that of the
   119 Isar proof element 'obtain', only that there may be several cases.
   120 Optional case names may be specified in parentheses; these will be
   121 available both in the present proof and as annotations in the
   122 resulting rule, for later use with the 'cases' method (cf. attribute
   123 case_names).
   124 
   125 * Isar: 'print_statement' prints theorems from the current theory or
   126 proof context in long statement form, according to the syntax of a
   127 top-level lemma.
   128 
   129 * Isar: 'obtain' takes an optional case name for the local context
   130 introduction rule (default "that").
   131 
   132 * Isar/locales: new derived specification elements 'definition',
   133 'abbreviation', 'axiomatization', which support type-inference, admit
   134 object-level specifications (equality, equivalence).  See also the
   135 isar-ref manual.  Examples:
   136 
   137   definition
   138     "f x y = x + y + 1"
   139     "g x = f x x"
   140 
   141   thm f_def g_def
   142 
   143   axiomatization
   144     eq  (infix "===" 50)
   145     where eq_refl: "x === x" and eq_subst: "x === y ==> P x ==> P y"
   146 
   147   abbreviation
   148     neq  (infix "=!=" 50)
   149     "x =!= y == ~ (x === y)"
   150 
   151 These specifications may be also used in a locale context.  Then the
   152 constants being introduced depend on certain fixed parameters, and the
   153 constant name is qualified by the locale base name.  An internal
   154 abbreviation takes care for convenient input and output, making the
   155 parameters implicit and using the original short name.  See also
   156 HOL/ex/Abstract_NAT.thy for an example of deriving polymorphic
   157 entities from a monomorphic theory.
   158 
   159 Presently, abbreviations are only available 'in' a target locale, but
   160 not inherited by general import expressions.  Also note that
   161 'abbreviation' may be used as a type-safe replacement for 'syntax' +
   162 'translations' in common applications.
   163 
   164 * Provers/induct: improved internal context management to support
   165 local fixes and defines on-the-fly.  Thus explicit meta-level
   166 connectives !! and ==> are rarely required anymore in inductive goals
   167 (using object-logic connectives for this purpose has been long
   168 obsolete anyway).  The subsequent proof patterns illustrate advanced
   169 techniques of natural induction; general datatypes and inductive sets
   170 work analogously (see also src/HOL/Lambda for realistic examples).
   171 
   172 (1) This is how to ``strengthen'' an inductive goal wrt. certain
   173 parameters:
   174 
   175   lemma
   176     fixes n :: nat and x :: 'a
   177     assumes a: "A n x"
   178     shows "P n x"
   179     using a                     -- {* make induct insert fact a *}
   180   proof (induct n fixing: x)    -- {* generalize goal to "!!x. A n x ==> P n x" *}
   181     case 0
   182     show ?case sorry
   183   next
   184     case (Suc n)
   185     note `!!x. A n x ==> P n x` -- {* induction hypothesis, according to induction rule *}
   186     note `A (Suc n) x`          -- {* induction premise, stemming from fact a *}
   187     show ?case sorry
   188   qed
   189 
   190 (2) This is how to perform induction over ``expressions of a certain
   191 form'', using a locally defined inductive parameter n == "a x"
   192 together with strengthening (the latter is usually required to get
   193 sufficiently flexible induction hypotheses):
   194 
   195   lemma
   196     fixes a :: "'a => nat"
   197     assumes a: "A (a x)"
   198     shows "P (a x)"
   199     using a
   200   proof (induct n == "a x" fixing: x)
   201     ...
   202 
   203 See also HOL/Isar_examples/Puzzle.thy for an application of the this
   204 particular technique.
   205 
   206 (3) This is how to perform existential reasoning ('obtains' or
   207 'obtain') by induction, while avoiding explicit object-logic
   208 encodings:
   209 
   210   lemma
   211     fixes n :: nat
   212     obtains x :: 'a where "P n x" and "Q n x"
   213   proof (induct n fixing: thesis)
   214     case 0
   215     obtain x where "P 0 x" and "Q 0 x" sorry
   216     then show thesis by (rule 0)
   217   next
   218     case (Suc n)
   219     obtain x where "P n x" and "Q n x" by (rule Suc.hyps)
   220     obtain x where "P (Suc n) x" and "Q (Suc n) x" sorry
   221     then show thesis by (rule Suc.prems)
   222   qed
   223 
   224 Here the 'fixing: thesis' specification essentially modifies the scope
   225 of the formal thesis parameter, in order to the get the whole
   226 existence statement through the induction as expected.
   227 
   228 * Provers/induct: mutual induction rules are now specified as a list
   229 of rule sharing the same induction cases.  HOL packages usually
   230 provide foo_bar.inducts for mutually defined items foo and bar
   231 (e.g. inductive sets or datatypes).  INCOMPATIBILITY, users need to
   232 specify mutual induction rules differently, i.e. like this:
   233 
   234   (induct rule: foo_bar.inducts)
   235   (induct set: foo bar)
   236   (induct type: foo bar)
   237 
   238 The ML function ProjectRule.projections turns old-style rules into the
   239 new format.
   240 
   241 * Provers/induct: improved handling of simultaneous goals.  Instead of
   242 introducing object-level conjunction, the statement is now split into
   243 several conclusions, while the corresponding symbolic cases are
   244 nested accordingly.  INCOMPATIBILITY, proofs need to be structured
   245 explicitly.  For example:
   246 
   247   lemma
   248     fixes n :: nat
   249     shows "P n" and "Q n"
   250   proof (induct n)
   251     case 0 case 1
   252     show "P 0" sorry
   253   next
   254     case 0 case 2
   255     show "Q 0" sorry
   256   next
   257     case (Suc n) case 1
   258     note `P n` and `Q n`
   259     show "P (Suc n)" sorry
   260   next
   261     case (Suc n) case 2
   262     note `P n` and `Q n`
   263     show "Q (Suc n)" sorry
   264   qed
   265 
   266 The split into subcases may be deferred as follows -- this is
   267 particularly relevant for goal statements with local premises.
   268 
   269   lemma
   270     fixes n :: nat
   271     shows "A n ==> P n" and "B n ==> Q n"
   272   proof (induct n)
   273     case 0
   274     {
   275       case 1
   276       note `A 0`
   277       show "P 0" sorry
   278     next
   279       case 2
   280       note `B 0`
   281       show "Q 0" sorry
   282     }
   283   next
   284     case (Suc n)
   285     note `A n ==> P n` and `B n ==> Q n`
   286     {
   287       case 1
   288       note `A (Suc n)`
   289       show "P (Suc n)" sorry
   290     next
   291       case 2
   292       note `B (Suc n)`
   293       show "Q (Suc n)" sorry
   294     }
   295   qed
   296 
   297 If simultaneous goals are to be used with mutual rules, the statement
   298 needs to be structured carefully as a two-level conjunction, using
   299 lists of propositions separated by 'and':
   300 
   301   lemma
   302     shows "a : A ==> P1 a"
   303           "a : A ==> P2 a"
   304       and "b : B ==> Q1 b"
   305           "b : B ==> Q2 b"
   306           "b : B ==> Q3 b"
   307   proof (induct set: A B)
   308 
   309 * Provers/induct: support coinduction as well.  See
   310 src/HOL/Library/Coinductive_List.thy for various examples.
   311 
   312 * Simplifier: by default the simplifier trace only shows top level rewrites
   313 now. That is, trace_simp_depth_limit is set to 1 by default. Thus there is
   314 less danger of being flooded by the trace. The trace indicates where parts
   315 have been suppressed.
   316   
   317 * Provers/classical: removed obsolete classical version of elim_format
   318 attribute; classical elim/dest rules are now treated uniformly when
   319 manipulating the claset.
   320 
   321 * Provers/classical: stricter checks to ensure that supplied intro,
   322 dest and elim rules are well-formed; dest and elim rules must have at
   323 least one premise.
   324 
   325 * Provers/classical: attributes dest/elim/intro take an optional
   326 weight argument for the rule (just as the Pure versions).  Weights are
   327 ignored by automated tools, but determine the search order of single
   328 rule steps.
   329 
   330 * Syntax: input syntax now supports dummy variable binding "%_. b",
   331 where the body does not mention the bound variable.  Note that dummy
   332 patterns implicitly depend on their context of bounds, which makes
   333 "{_. _}" match any set comprehension as expected.  Potential
   334 INCOMPATIBILITY -- parse translations need to cope with syntactic
   335 constant "_idtdummy" in the binding position.
   336 
   337 * Syntax: removed obsolete syntactic constant "_K" and its associated
   338 parse translation.  INCOMPATIBILITY -- use dummy abstraction instead,
   339 for example "A -> B" => "Pi A (%_. B)".
   340 
   341 
   342 *** HOL ***
   343 
   344 * Alternative iff syntax "A <-> B" for equality on bool (with priority
   345 25 like -->); output depends on the "iff" print_mode, the default is
   346 "A = B" (with priority 50).
   347 
   348 * Renamed constants in HOL.thy and Orderings.thy:
   349     op +   ~> HOL.plus
   350     op -   ~> HOL.minus
   351     uminus ~> HOL.uminus
   352     op *   ~> HOL.times
   353     op <   ~> Orderings.less
   354     op <=  ~> Orderings.less_eq
   355 
   356 Adaptions may be required in the following cases:
   357 
   358 a) User-defined constants using any of the names "plus", "minus", "times", "less"
   359 or "less_eq". The standard syntax translations for "+", "-" and "*" may go wrong.
   360 INCOMPATIBILITY: use more specific names.
   361 
   362 b) Variables named "plus", "minus", "times", "less", "less_eq"
   363 INCOMPATIBILITY: use more specific names.
   364 
   365 c) Commutative equations theorems (e. g. "a + b = b + a")
   366 Since the changing of names also changes the order of terms, commutative rewrites
   367 perhaps will be applied when not applied before or the other way round.
   368 Experience shows that thiis is rarely the case (only two adaptions in the whole
   369 Isabelle distribution).
   370 INCOMPATIBILITY: rewrite proof; sometimes already an exchange of a distributive
   371 law may suffice.
   372 
   373 d) ML code directly refering to constant names
   374 This in general only affects hand-written proof tactics, simprocs and so on.
   375 INCOMPATIBILITY: grep your sourcecode and replace names.
   376 
   377 * "LEAST x:A. P" expands to "LEAST x. x:A & P" (input only).
   378 
   379 * In the context of the assumption "~(s = t)" the Simplifier rewrites
   380 "t = s" to False (by simproc "neq_simproc").  For backward
   381 compatibility this can be disabled by ML "reset use_neq_simproc".
   382 
   383 * "m dvd n" where m and n are numbers is evaluated to True/False by simp.
   384 
   385 * Theorem Cons_eq_map_conv no longer has attribute `simp'.
   386 
   387 * Theorem setsum_mult renamed to setsum_right_distrib.
   388 
   389 * Prefer ex1I over ex_ex1I in single-step reasoning, e.g. by the
   390 'rule' method.
   391 
   392 * Tactics 'sat' and 'satx' reimplemented, several improvements: goals
   393 no longer need to be stated as "<prems> ==> False", equivalences (i.e.
   394 "=" on type bool) are handled, variable names of the form "lit_<n>"
   395 are no longer reserved, significant speedup.
   396 
   397 * inductive and datatype: provide projections of mutual rules, bundled
   398 as foo_bar.inducts;
   399 
   400 * Library: added theory Coinductive_List of potentially infinite lists
   401 as greatest fixed-point.
   402 
   403 * Library: added theory AssocList which implements (finite) maps as
   404 association lists.
   405 
   406 
   407 *** ML ***
   408 
   409 * Pure/library:
   410 
   411   val burrow: ('a list -> 'b list) -> 'a list list -> 'b list list
   412   val fold_burrow: ('a list -> 'c -> 'b list * 'd) -> 'a list list -> 'c -> 'b list list * 'd
   413 
   414 The semantics of "burrow" is: "take a function with *simulatanously*
   415 transforms a list of value, and apply it *simulatanously* to a list of
   416 list of values of the appropriate type". Confer this with "map" which
   417 would *not* apply its argument function simulatanously but in
   418 sequence. "fold_burrow" has an additional context.
   419 
   420 Both actually avoid the usage of "unflat" since they hide away
   421 "unflat" from the user.
   422 
   423 * Pure/library: functions map2 and fold2 with curried syntax for
   424 simultanous mapping and folding:
   425 
   426     val map2: ('a -> 'b -> 'c) -> 'a list -> 'b list -> 'c list
   427     val fold2: ('a -> 'b -> 'c -> 'c) -> 'a list -> 'b list -> 'c -> 'c
   428 
   429 * Pure/library: indexed lists - some functions in the Isabelle library
   430 treating lists over 'a as finite mappings from [0...n] to 'a have been
   431 given more convenient names and signatures reminiscent of similar
   432 functions for alists, tables, etc:
   433 
   434   val nth: 'a list -> int -> 'a 
   435   val nth_update: int * 'a -> 'a list -> 'a list
   436   val nth_map: int -> ('a -> 'a) -> 'a list -> 'a list
   437   val fold_index: (int * 'a -> 'b -> 'b) -> 'a list -> 'b -> 'b
   438 
   439 Note that fold_index starts counting at index 0, not 1 like foldln
   440 used to.
   441 
   442 * Pure/General/name_mangler.ML provides a functor for generic name
   443 mangling (bijective mapping from any expression values to strings).
   444 
   445 * Pure/General/rat.ML implements rational numbers.
   446 
   447 * Pure/General/table.ML: the join operations now works via exceptions
   448 DUP/SAME instead of type option.  This is simpler in simple cases, and
   449 admits slightly more efficient complex applications.
   450 
   451 * Pure: datatype Context.generic joins theory/Proof.context and
   452 provides some facilities for code that works in either kind of
   453 context, notably GenericDataFun for uniform theory and proof data.
   454 
   455 * Pure: 'advanced' translation functions (parse_translation etc.) now
   456 use Context.generic instead of just theory.
   457 
   458 * Pure: simplified internal attribute type, which is now always
   459 Context.generic * thm -> Context.generic * thm.  Global (theory)
   460 vs. local (Proof.context) attributes have been discontinued, while
   461 minimizing code duplication.  Thm.rule_attribute and
   462 Thm.declaration_attribute build canonical attributes; see also
   463 structure Context for further operations on Context.generic, notably
   464 GenericDataFun.  INCOMPATIBILITY, need to adapt attribute type
   465 declarations and definitions.
   466 
   467 * Pure/Isar: Args/Attrib parsers operate on Context.generic --
   468 global/local versions on theory vs. Proof.context have been
   469 discontinued; Attrib.syntax and Method.syntax have been adapted
   470 accordingly.  INCOMPATIBILITY, need to adapt parser expressions for
   471 attributes, methods, etc.
   472 
   473 * Pure: several functions of signature "... -> theory -> theory * ..."
   474 have been reoriented to "... -> theory -> ... * theory" in order to
   475 allow natural usage in combination with the ||>, ||>>, |-> and
   476 fold_map combinators.
   477 
   478 * Pure: primitive rule lift_rule now takes goal cterm instead of an
   479 actual goal state (thm).  Use Thm.lift_rule (Thm.cprem_of st i) to
   480 achieve the old behaviour.
   481 
   482 * Pure: the "Goal" constant is now called "prop", supporting a
   483 slightly more general idea of ``protecting'' meta-level rule
   484 statements.
   485 
   486 * Pure: structure Goal provides simple interfaces for
   487 init/conclude/finish and tactical prove operations (replacing former
   488 Tactic.prove).  Note that OldGoals.prove_goalw_cterm has long been
   489 obsolete, it is ill-behaved in a local proof context (e.g. with local
   490 fixes/assumes or in a locale context).
   491 
   492 * Isar: simplified treatment of user-level errors, using exception
   493 ERROR of string uniformly.  Function error now merely raises ERROR,
   494 without any side effect on output channels.  The Isar toplevel takes
   495 care of proper display of ERROR exceptions.  ML code may use plain
   496 handle/can/try; cat_error may be used to concatenate errors like this:
   497 
   498   ... handle ERROR msg => cat_error msg "..."
   499 
   500 Toplevel ML code (run directly or through the Isar toplevel) may be
   501 embedded into the Isar toplevel with exception display/debug like
   502 this:
   503 
   504   Isar.toplevel (fn () => ...)
   505 
   506 INCOMPATIBILITY, removed special transform_error facilities, removed
   507 obsolete variants of user-level exceptions (ERROR_MESSAGE,
   508 Context.PROOF, ProofContext.CONTEXT, Proof.STATE, ProofHistory.FAIL)
   509 -- use plain ERROR instead.
   510 
   511 * Isar: theory setup now has type (theory -> theory), instead of a
   512 list.  INCOMPATIBILITY, may use #> to compose setup functions.
   513 
   514 * Isar: installed ML toplevel pretty printer for type Proof.context,
   515 subject to ProofContext.debug/verbose flags.
   516 
   517 * Isar: Toplevel.theory_to_proof admits transactions that modify the
   518 theory before entering a proof state.  Transactions now always see a
   519 quasi-functional intermediate checkpoint, both in interactive and
   520 batch mode.
   521 
   522 * Simplifier: the simpset of a running simplification process now
   523 contains a proof context (cf. Simplifier.the_context), which is the
   524 very context that the initial simpset has been retrieved from (by
   525 simpset_of/local_simpset_of).  Consequently, all plug-in components
   526 (solver, looper etc.) may depend on arbitrary proof data.
   527 
   528 * Simplifier.inherit_context inherits the proof context (plus the
   529 local bounds) of the current simplification process; any simproc
   530 etc. that calls the Simplifier recursively should do this!  Removed
   531 former Simplifier.inherit_bounds, which is already included here --
   532 INCOMPATIBILITY.  Tools based on low-level rewriting may even have to
   533 specify an explicit context using Simplifier.context/theory_context.
   534 
   535 * Simplifier/Classical Reasoner: more abstract interfaces
   536 change_simpset/claset for modifying the simpset/claset reference of a
   537 theory; raw versions simpset/claset_ref etc. have been discontinued --
   538 INCOMPATIBILITY.
   539 
   540 * Provers: more generic wrt. syntax of object-logics, avoid hardwired
   541 "Trueprop" etc.
   542 
   543 
   544 
   545 New in Isabelle2005 (October 2005)
   546 ----------------------------------
   547 
   548 *** General ***
   549 
   550 * Theory headers: the new header syntax for Isar theories is
   551 
   552   theory <name>
   553   imports <theory1> ... <theoryN>
   554   uses <file1> ... <fileM>
   555   begin
   556 
   557 where the 'uses' part is optional.  The previous syntax
   558 
   559   theory <name> = <theory1> + ... + <theoryN>:
   560 
   561 will disappear in the next release.  Use isatool fixheaders to convert
   562 existing theory files.  Note that there is no change in ancient
   563 non-Isar theories now, but these will disappear soon.
   564 
   565 * Theory loader: parent theories can now also be referred to via
   566 relative and absolute paths.
   567 
   568 * Command 'find_theorems' searches for a list of criteria instead of a
   569 list of constants. Known criteria are: intro, elim, dest, name:string,
   570 simp:term, and any term. Criteria can be preceded by '-' to select
   571 theorems that do not match. Intro, elim, dest select theorems that
   572 match the current goal, name:s selects theorems whose fully qualified
   573 name contain s, and simp:term selects all simplification rules whose
   574 lhs match term.  Any other term is interpreted as pattern and selects
   575 all theorems matching the pattern. Available in ProofGeneral under
   576 'ProofGeneral -> Find Theorems' or C-c C-f.  Example:
   577 
   578   C-c C-f (100) "(_::nat) + _ + _" intro -name: "HOL."
   579 
   580 prints the last 100 theorems matching the pattern "(_::nat) + _ + _",
   581 matching the current goal as introduction rule and not having "HOL."
   582 in their name (i.e. not being defined in theory HOL).
   583 
   584 * Command 'thms_containing' has been discontinued in favour of
   585 'find_theorems'; INCOMPATIBILITY.
   586 
   587 * Communication with Proof General is now 8bit clean, which means that
   588 Unicode text in UTF-8 encoding may be used within theory texts (both
   589 formal and informal parts).  Cf. option -U of the Isabelle Proof
   590 General interface.  Here are some simple examples (cf. src/HOL/ex):
   591 
   592   http://isabelle.in.tum.de/library/HOL/ex/Hebrew.html
   593   http://isabelle.in.tum.de/library/HOL/ex/Chinese.html
   594 
   595 * Improved efficiency of the Simplifier and, to a lesser degree, the
   596 Classical Reasoner.  Typical big applications run around 2 times
   597 faster.
   598 
   599 
   600 *** Document preparation ***
   601 
   602 * Commands 'display_drafts' and 'print_drafts' perform simple output
   603 of raw sources.  Only those symbols that do not require additional
   604 LaTeX packages (depending on comments in isabellesym.sty) are
   605 displayed properly, everything else is left verbatim.  isatool display
   606 and isatool print are used as front ends (these are subject to the
   607 DVI/PDF_VIEWER and PRINT_COMMAND settings, respectively).
   608 
   609 * Command tags control specific markup of certain regions of text,
   610 notably folding and hiding.  Predefined tags include "theory" (for
   611 theory begin and end), "proof" for proof commands, and "ML" for
   612 commands involving ML code; the additional tags "visible" and
   613 "invisible" are unused by default.  Users may give explicit tag
   614 specifications in the text, e.g. ''by %invisible (auto)''.  The
   615 interpretation of tags is determined by the LaTeX job during document
   616 preparation: see option -V of isatool usedir, or options -n and -t of
   617 isatool document, or even the LaTeX macros \isakeeptag, \isafoldtag,
   618 \isadroptag.
   619 
   620 Several document versions may be produced at the same time via isatool
   621 usedir (the generated index.html will link all of them).  Typical
   622 specifications include ''-V document=theory,proof,ML'' to present
   623 theory/proof/ML parts faithfully, ''-V outline=/proof,/ML'' to fold
   624 proof and ML commands, and ''-V mutilated=-theory,-proof,-ML'' to omit
   625 these parts without any formal replacement text.  The Isabelle site
   626 default settings produce ''document'' and ''outline'' versions as
   627 specified above.
   628 
   629 * Several new antiquotations:
   630 
   631   @{term_type term} prints a term with its type annotated;
   632 
   633   @{typeof term} prints the type of a term;
   634 
   635   @{const const} is the same as @{term const}, but checks that the
   636   argument is a known logical constant;
   637 
   638   @{term_style style term} and @{thm_style style thm} print a term or
   639   theorem applying a "style" to it
   640 
   641   @{ML text}
   642 
   643 Predefined styles are 'lhs' and 'rhs' printing the lhs/rhs of
   644 definitions, equations, inequations etc., 'concl' printing only the
   645 conclusion of a meta-logical statement theorem, and 'prem1' .. 'prem19'
   646 to print the specified premise.  TermStyle.add_style provides an ML
   647 interface for introducing further styles.  See also the "LaTeX Sugar"
   648 document practical applications.  The ML antiquotation prints
   649 type-checked ML expressions verbatim.
   650 
   651 * Markup commands 'chapter', 'section', 'subsection', 'subsubsection',
   652 and 'text' support optional locale specification '(in loc)', which
   653 specifies the default context for interpreting antiquotations.  For
   654 example: 'text (in lattice) {* @{thm inf_assoc}*}'.
   655 
   656 * Option 'locale=NAME' of antiquotations specifies an alternative
   657 context interpreting the subsequent argument.  For example: @{thm
   658 [locale=lattice] inf_assoc}.
   659 
   660 * Proper output of proof terms (@{prf ...} and @{full_prf ...}) within
   661 a proof context.
   662 
   663 * Proper output of antiquotations for theory commands involving a
   664 proof context (such as 'locale' or 'theorem (in loc) ...').
   665 
   666 * Delimiters of outer tokens (string etc.) now produce separate LaTeX
   667 macros (\isachardoublequoteopen, isachardoublequoteclose etc.).
   668 
   669 * isatool usedir: new option -C (default true) controls whether option
   670 -D should include a copy of the original document directory; -C false
   671 prevents unwanted effects such as copying of administrative CVS data.
   672 
   673 
   674 *** Pure ***
   675 
   676 * Considerably improved version of 'constdefs' command.  Now performs
   677 automatic type-inference of declared constants; additional support for
   678 local structure declarations (cf. locales and HOL records), see also
   679 isar-ref manual.  Potential INCOMPATIBILITY: need to observe strictly
   680 sequential dependencies of definitions within a single 'constdefs'
   681 section; moreover, the declared name needs to be an identifier.  If
   682 all fails, consider to fall back on 'consts' and 'defs' separately.
   683 
   684 * Improved indexed syntax and implicit structures.  First of all,
   685 indexed syntax provides a notational device for subscripted
   686 application, using the new syntax \<^bsub>term\<^esub> for arbitrary
   687 expressions.  Secondly, in a local context with structure
   688 declarations, number indexes \<^sub>n or the empty index (default
   689 number 1) refer to a certain fixed variable implicitly; option
   690 show_structs controls printing of implicit structures.  Typical
   691 applications of these concepts involve record types and locales.
   692 
   693 * New command 'no_syntax' removes grammar declarations (and
   694 translations) resulting from the given syntax specification, which is
   695 interpreted in the same manner as for the 'syntax' command.
   696 
   697 * 'Advanced' translation functions (parse_translation etc.) may depend
   698 on the signature of the theory context being presently used for
   699 parsing/printing, see also isar-ref manual.
   700 
   701 * Improved 'oracle' command provides a type-safe interface to turn an
   702 ML expression of type theory -> T -> term into a primitive rule of
   703 type theory -> T -> thm (i.e. the functionality of Thm.invoke_oracle
   704 is already included here); see also FOL/ex/IffExample.thy;
   705 INCOMPATIBILITY.
   706 
   707 * axclass: name space prefix for class "c" is now "c_class" (was "c"
   708 before); "cI" is no longer bound, use "c.intro" instead.
   709 INCOMPATIBILITY.  This change avoids clashes of fact bindings for
   710 axclasses vs. locales.
   711 
   712 * Improved internal renaming of symbolic identifiers -- attach primes
   713 instead of base 26 numbers.
   714 
   715 * New flag show_question_marks controls printing of leading question
   716 marks in schematic variable names.
   717 
   718 * In schematic variable names, *any* symbol following \<^isub> or
   719 \<^isup> is now treated as part of the base name.  For example, the
   720 following works without printing of awkward ".0" indexes:
   721 
   722   lemma "x\<^isub>1 = x\<^isub>2 ==> x\<^isub>2 = x\<^isub>1"
   723     by simp
   724 
   725 * Inner syntax includes (*(*nested*) comments*).
   726 
   727 * Pretty printer now supports unbreakable blocks, specified in mixfix
   728 annotations as "(00...)".
   729 
   730 * Clear separation of logical types and nonterminals, where the latter
   731 may only occur in 'syntax' specifications or type abbreviations.
   732 Before that distinction was only partially implemented via type class
   733 "logic" vs. "{}".  Potential INCOMPATIBILITY in rare cases of improper
   734 use of 'types'/'consts' instead of 'nonterminals'/'syntax'.  Some very
   735 exotic syntax specifications may require further adaption
   736 (e.g. Cube/Cube.thy).
   737 
   738 * Removed obsolete type class "logic", use the top sort {} instead.
   739 Note that non-logical types should be declared as 'nonterminals'
   740 rather than 'types'.  INCOMPATIBILITY for new object-logic
   741 specifications.
   742 
   743 * Attributes 'induct' and 'cases': type or set names may now be
   744 locally fixed variables as well.
   745 
   746 * Simplifier: can now control the depth to which conditional rewriting
   747 is traced via the PG menu Isabelle -> Settings -> Trace Simp Depth
   748 Limit.
   749 
   750 * Simplifier: simplification procedures may now take the current
   751 simpset into account (cf. Simplifier.simproc(_i) / mk_simproc
   752 interface), which is very useful for calling the Simplifier
   753 recursively.  Minor INCOMPATIBILITY: the 'prems' argument of simprocs
   754 is gone -- use prems_of_ss on the simpset instead.  Moreover, the
   755 low-level mk_simproc no longer applies Logic.varify internally, to
   756 allow for use in a context of fixed variables.
   757 
   758 * thin_tac now works even if the assumption being deleted contains !!
   759 or ==>.  More generally, erule now works even if the major premise of
   760 the elimination rule contains !! or ==>.
   761 
   762 * Method 'rules' has been renamed to 'iprover'. INCOMPATIBILITY.
   763 
   764 * Reorganized bootstrapping of the Pure theories; CPure is now derived
   765 from Pure, which contains all common declarations already.  Both
   766 theories are defined via plain Isabelle/Isar .thy files.
   767 INCOMPATIBILITY: elements of CPure (such as the CPure.intro /
   768 CPure.elim / CPure.dest attributes) now appear in the Pure name space;
   769 use isatool fixcpure to adapt your theory and ML sources.
   770 
   771 * New syntax 'name(i-j, i-, i, ...)' for referring to specific
   772 selections of theorems in named facts via index ranges.
   773 
   774 * 'print_theorems': in theory mode, really print the difference
   775 wrt. the last state (works for interactive theory development only),
   776 in proof mode print all local facts (cf. 'print_facts');
   777 
   778 * 'hide': option '(open)' hides only base names.
   779 
   780 * More efficient treatment of intermediate checkpoints in interactive
   781 theory development.
   782 
   783 * Code generator is now invoked via code_module (incremental code
   784 generation) and code_library (modular code generation, ML structures
   785 for each theory).  INCOMPATIBILITY: new keywords 'file' and 'contains'
   786 must be quoted when used as identifiers.
   787 
   788 * New 'value' command for reading, evaluating and printing terms using
   789 the code generator.  INCOMPATIBILITY: command keyword 'value' must be
   790 quoted when used as identifier.
   791 
   792 
   793 *** Locales ***
   794 
   795 * New commands for the interpretation of locale expressions in
   796 theories (1), locales (2) and proof contexts (3).  These generate
   797 proof obligations from the expression specification.  After the
   798 obligations have been discharged, theorems of the expression are added
   799 to the theory, target locale or proof context.  The synopsis of the
   800 commands is a follows:
   801 
   802   (1) interpretation expr inst
   803   (2) interpretation target < expr
   804   (3) interpret expr inst
   805 
   806 Interpretation in theories and proof contexts require a parameter
   807 instantiation of terms from the current context.  This is applied to
   808 specifications and theorems of the interpreted expression.
   809 Interpretation in locales only permits parameter renaming through the
   810 locale expression.  Interpretation is smart in that interpretations
   811 that are active already do not occur in proof obligations, neither are
   812 instantiated theorems stored in duplicate.  Use 'print_interps' to
   813 inspect active interpretations of a particular locale.  For details,
   814 see the Isar Reference manual.  Examples can be found in
   815 HOL/Finite_Set.thy and HOL/Algebra/UnivPoly.thy.
   816 
   817 INCOMPATIBILITY: former 'instantiate' has been withdrawn, use
   818 'interpret' instead.
   819 
   820 * New context element 'constrains' for adding type constraints to
   821 parameters.
   822 
   823 * Context expressions: renaming of parameters with syntax
   824 redeclaration.
   825 
   826 * Locale declaration: 'includes' disallowed.
   827 
   828 * Proper static binding of attribute syntax -- i.e. types / terms /
   829 facts mentioned as arguments are always those of the locale definition
   830 context, independently of the context of later invocations.  Moreover,
   831 locale operations (renaming and type / term instantiation) are applied
   832 to attribute arguments as expected.
   833 
   834 INCOMPATIBILITY of the ML interface: always pass Attrib.src instead of
   835 actual attributes; rare situations may require Attrib.attribute to
   836 embed those attributes into Attrib.src that lack concrete syntax.
   837 Attribute implementations need to cooperate properly with the static
   838 binding mechanism.  Basic parsers Args.XXX_typ/term/prop and
   839 Attrib.XXX_thm etc. already do the right thing without further
   840 intervention.  Only unusual applications -- such as "where" or "of"
   841 (cf. src/Pure/Isar/attrib.ML), which process arguments depending both
   842 on the context and the facts involved -- may have to assign parsed
   843 values to argument tokens explicitly.
   844 
   845 * Changed parameter management in theorem generation for long goal
   846 statements with 'includes'.  INCOMPATIBILITY: produces a different
   847 theorem statement in rare situations.
   848 
   849 * Locale inspection command 'print_locale' omits notes elements.  Use
   850 'print_locale!' to have them included in the output.
   851 
   852 
   853 *** Provers ***
   854 
   855 * Provers/hypsubst.ML: improved version of the subst method, for
   856 single-step rewriting: it now works in bound variable contexts. New is
   857 'subst (asm)', for rewriting an assumption.  INCOMPATIBILITY: may
   858 rewrite a different subterm than the original subst method, which is
   859 still available as 'simplesubst'.
   860 
   861 * Provers/quasi.ML: new transitivity reasoners for transitivity only
   862 and quasi orders.
   863 
   864 * Provers/trancl.ML: new transitivity reasoner for transitive and
   865 reflexive-transitive closure of relations.
   866 
   867 * Provers/blast.ML: new reference depth_limit to make blast's depth
   868 limit (previously hard-coded with a value of 20) user-definable.
   869 
   870 * Provers/simplifier.ML has been moved to Pure, where Simplifier.setup
   871 is peformed already.  Object-logics merely need to finish their
   872 initial simpset configuration as before.  INCOMPATIBILITY.
   873 
   874 
   875 *** HOL ***
   876 
   877 * Symbolic syntax of Hilbert Choice Operator is now as follows:
   878 
   879   syntax (epsilon)
   880     "_Eps" :: "[pttrn, bool] => 'a"    ("(3\<some>_./ _)" [0, 10] 10)
   881 
   882 The symbol \<some> is displayed as the alternative epsilon of LaTeX
   883 and x-symbol; use option '-m epsilon' to get it actually printed.
   884 Moreover, the mathematically important symbolic identifier \<epsilon>
   885 becomes available as variable, constant etc.  INCOMPATIBILITY,
   886 
   887 * "x > y" abbreviates "y < x" and "x >= y" abbreviates "y <= x".
   888 Similarly for all quantifiers: "ALL x > y" etc.  The x-symbol for >=
   889 is \<ge>. New transitivity rules have been added to HOL/Orderings.thy to
   890 support corresponding Isar calculations.
   891 
   892 * "{x:A. P}" abbreviates "{x. x:A & P}", and similarly for "\<in>"
   893 instead of ":".
   894 
   895 * theory SetInterval: changed the syntax for open intervals:
   896 
   897   Old       New
   898   {..n(}    {..<n}
   899   {)n..}    {n<..}
   900   {m..n(}   {m..<n}
   901   {)m..n}   {m<..n}
   902   {)m..n(}  {m<..<n}
   903 
   904 The old syntax is still supported but will disappear in the next
   905 release.  For conversion use the following Emacs search and replace
   906 patterns (these are not perfect but work quite well):
   907 
   908   {)\([^\.]*\)\.\.  ->  {\1<\.\.}
   909   \.\.\([^(}]*\)(}  ->  \.\.<\1}
   910 
   911 * Theory Commutative_Ring (in Library): method comm_ring for proving
   912 equalities in commutative rings; method 'algebra' provides a generic
   913 interface.
   914 
   915 * Theory Finite_Set: changed the syntax for 'setsum', summation over
   916 finite sets: "setsum (%x. e) A", which used to be "\<Sum>x:A. e", is
   917 now either "SUM x:A. e" or "\<Sum>x \<in> A. e". The bound variable can
   918 be a tuple pattern.
   919 
   920 Some new syntax forms are available:
   921 
   922   "\<Sum>x | P. e"      for     "setsum (%x. e) {x. P}"
   923   "\<Sum>x = a..b. e"   for     "setsum (%x. e) {a..b}"
   924   "\<Sum>x = a..<b. e"  for     "setsum (%x. e) {a..<b}"
   925   "\<Sum>x < k. e"      for     "setsum (%x. e) {..<k}"
   926 
   927 The latter form "\<Sum>x < k. e" used to be based on a separate
   928 function "Summation", which has been discontinued.
   929 
   930 * theory Finite_Set: in structured induction proofs, the insert case
   931 is now 'case (insert x F)' instead of the old counterintuitive 'case
   932 (insert F x)'.
   933 
   934 * The 'refute' command has been extended to support a much larger
   935 fragment of HOL, including axiomatic type classes, constdefs and
   936 typedefs, inductive datatypes and recursion.
   937 
   938 * New tactics 'sat' and 'satx' to prove propositional tautologies.
   939 Requires zChaff with proof generation to be installed.  See
   940 HOL/ex/SAT_Examples.thy for examples.
   941 
   942 * Datatype induction via method 'induct' now preserves the name of the
   943 induction variable. For example, when proving P(xs::'a list) by
   944 induction on xs, the induction step is now P(xs) ==> P(a#xs) rather
   945 than P(list) ==> P(a#list) as previously.  Potential INCOMPATIBILITY
   946 in unstructured proof scripts.
   947 
   948 * Reworked implementation of records.  Improved scalability for
   949 records with many fields, avoiding performance problems for type
   950 inference. Records are no longer composed of nested field types, but
   951 of nested extension types. Therefore the record type only grows linear
   952 in the number of extensions and not in the number of fields.  The
   953 top-level (users) view on records is preserved.  Potential
   954 INCOMPATIBILITY only in strange cases, where the theory depends on the
   955 old record representation. The type generated for a record is called
   956 <record_name>_ext_type.
   957 
   958 Flag record_quick_and_dirty_sensitive can be enabled to skip the
   959 proofs triggered by a record definition or a simproc (if
   960 quick_and_dirty is enabled).  Definitions of large records can take
   961 quite long.
   962 
   963 New simproc record_upd_simproc for simplification of multiple record
   964 updates enabled by default.  Moreover, trivial updates are also
   965 removed: r(|x := x r|) = r.  INCOMPATIBILITY: old proofs break
   966 occasionally, since simplification is more powerful by default.
   967 
   968 * typedef: proper support for polymorphic sets, which contain extra
   969 type-variables in the term.
   970 
   971 * Simplifier: automatically reasons about transitivity chains
   972 involving "trancl" (r^+) and "rtrancl" (r^*) by setting up tactics
   973 provided by Provers/trancl.ML as additional solvers.  INCOMPATIBILITY:
   974 old proofs break occasionally as simplification may now solve more
   975 goals than previously.
   976 
   977 * Simplifier: converts x <= y into x = y if assumption y <= x is
   978 present.  Works for all partial orders (class "order"), in particular
   979 numbers and sets.  For linear orders (e.g. numbers) it treats ~ x < y
   980 just like y <= x.
   981 
   982 * Simplifier: new simproc for "let x = a in f x".  If a is a free or
   983 bound variable or a constant then the let is unfolded.  Otherwise
   984 first a is simplified to b, and then f b is simplified to g. If
   985 possible we abstract b from g arriving at "let x = b in h x",
   986 otherwise we unfold the let and arrive at g.  The simproc can be
   987 enabled/disabled by the reference use_let_simproc.  Potential
   988 INCOMPATIBILITY since simplification is more powerful by default.
   989 
   990 * Classical reasoning: the meson method now accepts theorems as arguments.
   991 
   992 * Prover support: pre-release of the Isabelle-ATP linkup, which runs background
   993 jobs to provide advice on the provability of subgoals.
   994 
   995 * Theory OrderedGroup and Ring_and_Field: various additions and
   996 improvements to faciliate calculations involving equalities and
   997 inequalities.
   998 
   999 The following theorems have been eliminated or modified
  1000 (INCOMPATIBILITY):
  1001 
  1002   abs_eq             now named abs_of_nonneg
  1003   abs_of_ge_0        now named abs_of_nonneg
  1004   abs_minus_eq       now named abs_of_nonpos
  1005   imp_abs_id         now named abs_of_nonneg
  1006   imp_abs_neg_id     now named abs_of_nonpos
  1007   mult_pos           now named mult_pos_pos
  1008   mult_pos_le        now named mult_nonneg_nonneg
  1009   mult_pos_neg_le    now named mult_nonneg_nonpos
  1010   mult_pos_neg2_le   now named mult_nonneg_nonpos2
  1011   mult_neg           now named mult_neg_neg
  1012   mult_neg_le        now named mult_nonpos_nonpos
  1013 
  1014 * Theory Parity: added rules for simplifying exponents.
  1015 
  1016 * Theory List:
  1017 
  1018 The following theorems have been eliminated or modified
  1019 (INCOMPATIBILITY):
  1020 
  1021   list_all_Nil       now named list_all.simps(1)
  1022   list_all_Cons      now named list_all.simps(2)
  1023   list_all_conv      now named list_all_iff
  1024   set_mem_eq         now named mem_iff
  1025 
  1026 * Theories SetsAndFunctions and BigO (see HOL/Library) support
  1027 asymptotic "big O" calculations.  See the notes in BigO.thy.
  1028 
  1029 
  1030 *** HOL-Complex ***
  1031 
  1032 * Theory RealDef: better support for embedding natural numbers and
  1033 integers in the reals.
  1034 
  1035 The following theorems have been eliminated or modified
  1036 (INCOMPATIBILITY):
  1037 
  1038   exp_ge_add_one_self  now requires no hypotheses
  1039   real_of_int_add      reversed direction of equality (use [symmetric])
  1040   real_of_int_minus    reversed direction of equality (use [symmetric])
  1041   real_of_int_diff     reversed direction of equality (use [symmetric])
  1042   real_of_int_mult     reversed direction of equality (use [symmetric])
  1043 
  1044 * Theory RComplete: expanded support for floor and ceiling functions.
  1045 
  1046 * Theory Ln is new, with properties of the natural logarithm
  1047 
  1048 * Hyperreal: There is a new type constructor "star" for making
  1049 nonstandard types.  The old type names are now type synonyms:
  1050 
  1051   hypreal = real star
  1052   hypnat = nat star
  1053   hcomplex = complex star
  1054 
  1055 * Hyperreal: Many groups of similarly-defined constants have been
  1056 replaced by polymorphic versions (INCOMPATIBILITY):
  1057 
  1058   star_of <-- hypreal_of_real, hypnat_of_nat, hcomplex_of_complex
  1059 
  1060   starset      <-- starsetNat, starsetC
  1061   *s*          <-- *sNat*, *sc*
  1062   starset_n    <-- starsetNat_n, starsetC_n
  1063   *sn*         <-- *sNatn*, *scn*
  1064   InternalSets <-- InternalNatSets, InternalCSets
  1065 
  1066   starfun      <-- starfun{Nat,Nat2,C,RC,CR}
  1067   *f*          <-- *fNat*, *fNat2*, *fc*, *fRc*, *fcR*
  1068   starfun_n    <-- starfun{Nat,Nat2,C,RC,CR}_n
  1069   *fn*         <-- *fNatn*, *fNat2n*, *fcn*, *fRcn*, *fcRn*
  1070   InternalFuns <-- InternalNatFuns, InternalNatFuns2, Internal{C,RC,CR}Funs
  1071 
  1072 * Hyperreal: Many type-specific theorems have been removed in favor of
  1073 theorems specific to various axiomatic type classes (INCOMPATIBILITY):
  1074 
  1075   add_commute <-- {hypreal,hypnat,hcomplex}_add_commute
  1076   add_assoc   <-- {hypreal,hypnat,hcomplex}_add_assocs
  1077   OrderedGroup.add_0 <-- {hypreal,hypnat,hcomplex}_add_zero_left
  1078   OrderedGroup.add_0_right <-- {hypreal,hcomplex}_add_zero_right
  1079   right_minus <-- hypreal_add_minus
  1080   left_minus <-- {hypreal,hcomplex}_add_minus_left
  1081   mult_commute <-- {hypreal,hypnat,hcomplex}_mult_commute
  1082   mult_assoc <-- {hypreal,hypnat,hcomplex}_mult_assoc
  1083   mult_1_left <-- {hypreal,hypnat}_mult_1, hcomplex_mult_one_left
  1084   mult_1_right <-- hcomplex_mult_one_right
  1085   mult_zero_left <-- hcomplex_mult_zero_left
  1086   left_distrib <-- {hypreal,hypnat,hcomplex}_add_mult_distrib
  1087   right_distrib <-- hypnat_add_mult_distrib2
  1088   zero_neq_one <-- {hypreal,hypnat,hcomplex}_zero_not_eq_one
  1089   right_inverse <-- hypreal_mult_inverse
  1090   left_inverse <-- hypreal_mult_inverse_left, hcomplex_mult_inv_left
  1091   order_refl <-- {hypreal,hypnat}_le_refl
  1092   order_trans <-- {hypreal,hypnat}_le_trans
  1093   order_antisym <-- {hypreal,hypnat}_le_anti_sym
  1094   order_less_le <-- {hypreal,hypnat}_less_le
  1095   linorder_linear <-- {hypreal,hypnat}_le_linear
  1096   add_left_mono <-- {hypreal,hypnat}_add_left_mono
  1097   mult_strict_left_mono <-- {hypreal,hypnat}_mult_less_mono2
  1098   add_nonneg_nonneg <-- hypreal_le_add_order
  1099 
  1100 * Hyperreal: Separate theorems having to do with type-specific
  1101 versions of constants have been merged into theorems that apply to the
  1102 new polymorphic constants (INCOMPATIBILITY):
  1103 
  1104   STAR_UNIV_set <-- {STAR_real,NatStar_real,STARC_complex}_set
  1105   STAR_empty_set <-- {STAR,NatStar,STARC}_empty_set
  1106   STAR_Un <-- {STAR,NatStar,STARC}_Un
  1107   STAR_Int <-- {STAR,NatStar,STARC}_Int
  1108   STAR_Compl <-- {STAR,NatStar,STARC}_Compl
  1109   STAR_subset <-- {STAR,NatStar,STARC}_subset
  1110   STAR_mem <-- {STAR,NatStar,STARC}_mem
  1111   STAR_mem_Compl <-- {STAR,STARC}_mem_Compl
  1112   STAR_diff <-- {STAR,STARC}_diff
  1113   STAR_star_of_image_subset <-- {STAR_hypreal_of_real, NatStar_hypreal_of_real,
  1114     STARC_hcomplex_of_complex}_image_subset
  1115   starset_n_Un <-- starset{Nat,C}_n_Un
  1116   starset_n_Int <-- starset{Nat,C}_n_Int
  1117   starset_n_Compl <-- starset{Nat,C}_n_Compl
  1118   starset_n_diff <-- starset{Nat,C}_n_diff
  1119   InternalSets_Un <-- Internal{Nat,C}Sets_Un
  1120   InternalSets_Int <-- Internal{Nat,C}Sets_Int
  1121   InternalSets_Compl <-- Internal{Nat,C}Sets_Compl
  1122   InternalSets_diff <-- Internal{Nat,C}Sets_diff
  1123   InternalSets_UNIV_diff <-- Internal{Nat,C}Sets_UNIV_diff
  1124   InternalSets_starset_n <-- Internal{Nat,C}Sets_starset{Nat,C}_n
  1125   starset_starset_n_eq <-- starset{Nat,C}_starset{Nat,C}_n_eq
  1126   starset_n_starset <-- starset{Nat,C}_n_starset{Nat,C}
  1127   starfun_n_starfun <-- starfun{Nat,Nat2,C,RC,CR}_n_starfun{Nat,Nat2,C,RC,CR}
  1128   starfun <-- starfun{Nat,Nat2,C,RC,CR}
  1129   starfun_mult <-- starfun{Nat,Nat2,C,RC,CR}_mult
  1130   starfun_add <-- starfun{Nat,Nat2,C,RC,CR}_add
  1131   starfun_minus <-- starfun{Nat,Nat2,C,RC,CR}_minus
  1132   starfun_diff <-- starfun{C,RC,CR}_diff
  1133   starfun_o <-- starfun{NatNat2,Nat2,_stafunNat,C,C_starfunRC,_starfunCR}_o
  1134   starfun_o2 <-- starfun{NatNat2,_stafunNat,C,C_starfunRC,_starfunCR}_o2
  1135   starfun_const_fun <-- starfun{Nat,Nat2,C,RC,CR}_const_fun
  1136   starfun_inverse <-- starfun{Nat,C,RC,CR}_inverse
  1137   starfun_eq <-- starfun{Nat,Nat2,C,RC,CR}_eq
  1138   starfun_eq_iff <-- starfun{C,RC,CR}_eq_iff
  1139   starfun_Id <-- starfunC_Id
  1140   starfun_approx <-- starfun{Nat,CR}_approx
  1141   starfun_capprox <-- starfun{C,RC}_capprox
  1142   starfun_abs <-- starfunNat_rabs
  1143   starfun_lambda_cancel <-- starfun{C,CR,RC}_lambda_cancel
  1144   starfun_lambda_cancel2 <-- starfun{C,CR,RC}_lambda_cancel2
  1145   starfun_mult_HFinite_approx <-- starfunCR_mult_HFinite_capprox
  1146   starfun_mult_CFinite_capprox <-- starfun{C,RC}_mult_CFinite_capprox
  1147   starfun_add_capprox <-- starfun{C,RC}_add_capprox
  1148   starfun_add_approx <-- starfunCR_add_approx
  1149   starfun_inverse_inverse <-- starfunC_inverse_inverse
  1150   starfun_divide <-- starfun{C,CR,RC}_divide
  1151   starfun_n <-- starfun{Nat,C}_n
  1152   starfun_n_mult <-- starfun{Nat,C}_n_mult
  1153   starfun_n_add <-- starfun{Nat,C}_n_add
  1154   starfun_n_add_minus <-- starfunNat_n_add_minus
  1155   starfun_n_const_fun <-- starfun{Nat,C}_n_const_fun
  1156   starfun_n_minus <-- starfun{Nat,C}_n_minus
  1157   starfun_n_eq <-- starfun{Nat,C}_n_eq
  1158 
  1159   star_n_add <-- {hypreal,hypnat,hcomplex}_add
  1160   star_n_minus <-- {hypreal,hcomplex}_minus
  1161   star_n_diff <-- {hypreal,hcomplex}_diff
  1162   star_n_mult <-- {hypreal,hcomplex}_mult
  1163   star_n_inverse <-- {hypreal,hcomplex}_inverse
  1164   star_n_le <-- {hypreal,hypnat}_le
  1165   star_n_less <-- {hypreal,hypnat}_less
  1166   star_n_zero_num <-- {hypreal,hypnat,hcomplex}_zero_num
  1167   star_n_one_num <-- {hypreal,hypnat,hcomplex}_one_num
  1168   star_n_abs <-- hypreal_hrabs
  1169   star_n_divide <-- hcomplex_divide
  1170 
  1171   star_of_add <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_add
  1172   star_of_minus <-- {hypreal_of_real,hcomplex_of_complex}_minus
  1173   star_of_diff <-- hypreal_of_real_diff
  1174   star_of_mult <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_mult
  1175   star_of_one <-- {hypreal_of_real,hcomplex_of_complex}_one
  1176   star_of_zero <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_zero
  1177   star_of_le <-- {hypreal_of_real,hypnat_of_nat}_le_iff
  1178   star_of_less <-- {hypreal_of_real,hypnat_of_nat}_less_iff
  1179   star_of_eq <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_eq_iff
  1180   star_of_inverse <-- {hypreal_of_real,hcomplex_of_complex}_inverse
  1181   star_of_divide <-- {hypreal_of_real,hcomplex_of_complex}_divide
  1182   star_of_of_nat <-- {hypreal_of_real,hcomplex_of_complex}_of_nat
  1183   star_of_of_int <-- {hypreal_of_real,hcomplex_of_complex}_of_int
  1184   star_of_number_of <-- {hypreal,hcomplex}_number_of
  1185   star_of_number_less <-- number_of_less_hypreal_of_real_iff
  1186   star_of_number_le <-- number_of_le_hypreal_of_real_iff
  1187   star_of_eq_number <-- hypreal_of_real_eq_number_of_iff
  1188   star_of_less_number <-- hypreal_of_real_less_number_of_iff
  1189   star_of_le_number <-- hypreal_of_real_le_number_of_iff
  1190   star_of_power <-- hypreal_of_real_power
  1191   star_of_eq_0 <-- hcomplex_of_complex_zero_iff
  1192 
  1193 * Hyperreal: new method "transfer" that implements the transfer
  1194 principle of nonstandard analysis. With a subgoal that mentions
  1195 nonstandard types like "'a star", the command "apply transfer"
  1196 replaces it with an equivalent one that mentions only standard types.
  1197 To be successful, all free variables must have standard types; non-
  1198 standard variables must have explicit universal quantifiers.
  1199 
  1200 * Hyperreal: A theory of Taylor series.
  1201 
  1202 
  1203 *** HOLCF ***
  1204 
  1205 * Discontinued special version of 'constdefs' (which used to support
  1206 continuous functions) in favor of the general Pure one with full
  1207 type-inference.
  1208 
  1209 * New simplification procedure for solving continuity conditions; it
  1210 is much faster on terms with many nested lambda abstractions (cubic
  1211 instead of exponential time).
  1212 
  1213 * New syntax for domain package: selector names are now optional.
  1214 Parentheses should be omitted unless argument is lazy, for example:
  1215 
  1216   domain 'a stream = cons "'a" (lazy "'a stream")
  1217 
  1218 * New command 'fixrec' for defining recursive functions with pattern
  1219 matching; defining multiple functions with mutual recursion is also
  1220 supported.  Patterns may include the constants cpair, spair, up, sinl,
  1221 sinr, or any data constructor defined by the domain package. The given
  1222 equations are proven as rewrite rules. See HOLCF/ex/Fixrec_ex.thy for
  1223 syntax and examples.
  1224 
  1225 * New commands 'cpodef' and 'pcpodef' for defining predicate subtypes
  1226 of cpo and pcpo types. Syntax is exactly like the 'typedef' command,
  1227 but the proof obligation additionally includes an admissibility
  1228 requirement. The packages generate instances of class cpo or pcpo,
  1229 with continuity and strictness theorems for Rep and Abs.
  1230 
  1231 * HOLCF: Many theorems have been renamed according to a more standard naming
  1232 scheme (INCOMPATIBILITY):
  1233 
  1234   foo_inject:  "foo$x = foo$y ==> x = y"
  1235   foo_eq:      "(foo$x = foo$y) = (x = y)"
  1236   foo_less:    "(foo$x << foo$y) = (x << y)"
  1237   foo_strict:  "foo$UU = UU"
  1238   foo_defined: "... ==> foo$x ~= UU"
  1239   foo_defined_iff: "(foo$x = UU) = (x = UU)"
  1240 
  1241 
  1242 *** ZF ***
  1243 
  1244 * ZF/ex: theories Group and Ring provide examples in abstract algebra,
  1245 including the First Isomorphism Theorem (on quotienting by the kernel
  1246 of a homomorphism).
  1247 
  1248 * ZF/Simplifier: install second copy of type solver that actually
  1249 makes use of TC rules declared to Isar proof contexts (or locales);
  1250 the old version is still required for ML proof scripts.
  1251 
  1252 
  1253 *** Cube ***
  1254 
  1255 * Converted to Isar theory format; use locales instead of axiomatic
  1256 theories.
  1257 
  1258 
  1259 *** ML ***
  1260 
  1261 * Pure/library.ML no longer defines its own option datatype, but uses
  1262 that of the SML basis, which has constructors NONE and SOME instead of
  1263 None and Some, as well as exception Option.Option instead of OPTION.
  1264 The functions the, if_none, is_some, is_none have been adapted
  1265 accordingly, while Option.map replaces apsome.
  1266 
  1267 * Pure/library.ML: the exception LIST has been given up in favour of
  1268 the standard exceptions Empty and Subscript, as well as
  1269 Library.UnequalLengths.  Function like Library.hd and Library.tl are
  1270 superceded by the standard hd and tl functions etc.
  1271 
  1272 A number of basic list functions are no longer exported to the ML
  1273 toplevel, as they are variants of predefined functions.  The following
  1274 suggests how one can translate existing code:
  1275 
  1276     rev_append xs ys = List.revAppend (xs, ys)
  1277     nth_elem (i, xs) = List.nth (xs, i)
  1278     last_elem xs = List.last xs
  1279     flat xss = List.concat xss
  1280     seq fs = List.app fs
  1281     partition P xs = List.partition P xs
  1282     mapfilter f xs = List.mapPartial f xs
  1283 
  1284 * Pure/library.ML: several combinators for linear functional
  1285 transformations, notably reverse application and composition:
  1286 
  1287   x |> f                f #> g
  1288   (x, y) |-> f          f #-> g
  1289 
  1290 * Pure/library.ML: introduced/changed precedence of infix operators:
  1291 
  1292   infix 1 |> |-> ||> ||>> |>> |>>> #> #->;
  1293   infix 2 ?;
  1294   infix 3 o oo ooo oooo;
  1295   infix 4 ~~ upto downto;
  1296 
  1297 Maybe INCOMPATIBILITY when any of those is used in conjunction with other
  1298 infix operators.
  1299 
  1300 * Pure/library.ML: natural list combinators fold, fold_rev, and
  1301 fold_map support linear functional transformations and nesting.  For
  1302 example:
  1303 
  1304   fold f [x1, ..., xN] y =
  1305     y |> f x1 |> ... |> f xN
  1306 
  1307   (fold o fold) f [xs1, ..., xsN] y =
  1308     y |> fold f xs1 |> ... |> fold f xsN
  1309 
  1310   fold f [x1, ..., xN] =
  1311     f x1 #> ... #> f xN
  1312 
  1313   (fold o fold) f [xs1, ..., xsN] =
  1314     fold f xs1 #> ... #> fold f xsN
  1315 
  1316 * Pure/library.ML: the following selectors on type 'a option are
  1317 available:
  1318 
  1319   the:               'a option -> 'a  (*partial*)
  1320   these:             'a option -> 'a  where 'a = 'b list
  1321   the_default: 'a -> 'a option -> 'a
  1322   the_list:          'a option -> 'a list
  1323 
  1324 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  1325 basic operations for association lists, following natural argument
  1326 order; moreover the explicit equality predicate passed here avoids
  1327 potentially expensive polymorphic runtime equality checks.
  1328 The old functions may be expressed as follows:
  1329 
  1330   assoc = uncurry (AList.lookup (op =))
  1331   assocs = these oo AList.lookup (op =)
  1332   overwrite = uncurry (AList.update (op =)) o swap
  1333 
  1334 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  1335 
  1336   val make: ('a -> 'b) -> 'a list -> ('a * 'b) list
  1337   val find: ('a * 'b -> bool) -> ('c * 'b) list -> 'a -> 'c list
  1338 
  1339 replacing make_keylist and keyfilter (occassionally used)
  1340 Naive rewrites:
  1341 
  1342   make_keylist = AList.make
  1343   keyfilter = AList.find (op =)
  1344 
  1345 * eq_fst and eq_snd now take explicit equality parameter, thus
  1346   avoiding eqtypes. Naive rewrites:
  1347 
  1348     eq_fst = eq_fst (op =)
  1349     eq_snd = eq_snd (op =)
  1350 
  1351 * Removed deprecated apl and apr (rarely used).
  1352   Naive rewrites:
  1353 
  1354     apl (n, op) =>>= curry op n
  1355     apr (op, m) =>>= fn n => op (n, m)
  1356 
  1357 * Pure/General: structure OrdList (cf. Pure/General/ord_list.ML)
  1358 provides a reasonably efficient light-weight implementation of sets as
  1359 lists.
  1360 
  1361 * Pure/General: generic tables (cf. Pure/General/table.ML) provide a
  1362 few new operations; existing lookup and update are now curried to
  1363 follow natural argument order (for use with fold etc.);
  1364 INCOMPATIBILITY, use (uncurry Symtab.lookup) etc. as last resort.
  1365 
  1366 * Pure/General: output via the Isabelle channels of
  1367 writeln/warning/error etc. is now passed through Output.output, with a
  1368 hook for arbitrary transformations depending on the print_mode
  1369 (cf. Output.add_mode -- the first active mode that provides a output
  1370 function wins).  Already formatted output may be embedded into further
  1371 text via Output.raw; the result of Pretty.string_of/str_of and derived
  1372 functions (string_of_term/cterm/thm etc.) is already marked raw to
  1373 accommodate easy composition of diagnostic messages etc.  Programmers
  1374 rarely need to care about Output.output or Output.raw at all, with
  1375 some notable exceptions: Output.output is required when bypassing the
  1376 standard channels (writeln etc.), or in token translations to produce
  1377 properly formatted results; Output.raw is required when capturing
  1378 already output material that will eventually be presented to the user
  1379 a second time.  For the default print mode, both Output.output and
  1380 Output.raw have no effect.
  1381 
  1382 * Pure/General: Output.time_accumulator NAME creates an operator ('a
  1383 -> 'b) -> 'a -> 'b to measure runtime and count invocations; the
  1384 cumulative results are displayed at the end of a batch session.
  1385 
  1386 * Pure/General: File.sysify_path and File.quote_sysify path have been
  1387 replaced by File.platform_path and File.shell_path (with appropriate
  1388 hooks).  This provides a clean interface for unusual systems where the
  1389 internal and external process view of file names are different.
  1390 
  1391 * Pure: more efficient orders for basic syntactic entities: added
  1392 fast_string_ord, fast_indexname_ord, fast_term_ord; changed sort_ord
  1393 and typ_ord to use fast_string_ord and fast_indexname_ord (term_ord is
  1394 NOT affected); structures Symtab, Vartab, Typtab, Termtab use the fast
  1395 orders now -- potential INCOMPATIBILITY for code that depends on a
  1396 particular order for Symtab.keys, Symtab.dest, etc. (consider using
  1397 Library.sort_strings on result).
  1398 
  1399 * Pure/term.ML: combinators fold_atyps, fold_aterms, fold_term_types,
  1400 fold_types traverse types/terms from left to right, observing natural
  1401 argument order.  Supercedes previous foldl_XXX versions, add_frees,
  1402 add_vars etc. have been adapted as well: INCOMPATIBILITY.
  1403 
  1404 * Pure: name spaces have been refined, with significant changes of the
  1405 internal interfaces -- INCOMPATIBILITY.  Renamed cond_extern(_table)
  1406 to extern(_table).  The plain name entry path is superceded by a
  1407 general 'naming' context, which also includes the 'policy' to produce
  1408 a fully qualified name and external accesses of a fully qualified
  1409 name; NameSpace.extend is superceded by context dependent
  1410 Sign.declare_name.  Several theory and proof context operations modify
  1411 the naming context.  Especially note Theory.restore_naming and
  1412 ProofContext.restore_naming to get back to a sane state; note that
  1413 Theory.add_path is no longer sufficient to recover from
  1414 Theory.absolute_path in particular.
  1415 
  1416 * Pure: new flags short_names (default false) and unique_names
  1417 (default true) for controlling output of qualified names.  If
  1418 short_names is set, names are printed unqualified.  If unique_names is
  1419 reset, the name prefix is reduced to the minimum required to achieve
  1420 the original result when interning again, even if there is an overlap
  1421 with earlier declarations.
  1422 
  1423 * Pure/TheoryDataFun: change of the argument structure; 'prep_ext' is
  1424 now 'extend', and 'merge' gets an additional Pretty.pp argument
  1425 (useful for printing error messages).  INCOMPATIBILITY.
  1426 
  1427 * Pure: major reorganization of the theory context.  Type Sign.sg and
  1428 Theory.theory are now identified, referring to the universal
  1429 Context.theory (see Pure/context.ML).  Actual signature and theory
  1430 content is managed as theory data.  The old code and interfaces were
  1431 spread over many files and structures; the new arrangement introduces
  1432 considerable INCOMPATIBILITY to gain more clarity:
  1433 
  1434   Context -- theory management operations (name, identity, inclusion,
  1435     parents, ancestors, merge, etc.), plus generic theory data;
  1436 
  1437   Sign -- logical signature and syntax operations (declaring consts,
  1438     types, etc.), plus certify/read for common entities;
  1439 
  1440   Theory -- logical theory operations (stating axioms, definitions,
  1441     oracles), plus a copy of logical signature operations (consts,
  1442     types, etc.); also a few basic management operations (Theory.copy,
  1443     Theory.merge, etc.)
  1444 
  1445 The most basic sign_of operations (Theory.sign_of, Thm.sign_of_thm
  1446 etc.) as well as the sign field in Thm.rep_thm etc. have been retained
  1447 for convenience -- they merely return the theory.
  1448 
  1449 * Pure: type Type.tsig is superceded by theory in most interfaces.
  1450 
  1451 * Pure: the Isar proof context type is already defined early in Pure
  1452 as Context.proof (note that ProofContext.context and Proof.context are
  1453 aliases, where the latter is the preferred name).  This enables other
  1454 Isabelle components to refer to that type even before Isar is present.
  1455 
  1456 * Pure/sign/theory: discontinued named name spaces (i.e. classK,
  1457 typeK, constK, axiomK, oracleK), but provide explicit operations for
  1458 any of these kinds.  For example, Sign.intern typeK is now
  1459 Sign.intern_type, Theory.hide_space Sign.typeK is now
  1460 Theory.hide_types.  Also note that former
  1461 Theory.hide_classes/types/consts are now
  1462 Theory.hide_classes_i/types_i/consts_i, while the non '_i' versions
  1463 internalize their arguments!  INCOMPATIBILITY.
  1464 
  1465 * Pure: get_thm interface (of PureThy and ProofContext) expects
  1466 datatype thmref (with constructors Name and NameSelection) instead of
  1467 plain string -- INCOMPATIBILITY;
  1468 
  1469 * Pure: cases produced by proof methods specify options, where NONE
  1470 means to remove case bindings -- INCOMPATIBILITY in
  1471 (RAW_)METHOD_CASES.
  1472 
  1473 * Pure: the following operations retrieve axioms or theorems from a
  1474 theory node or theory hierarchy, respectively:
  1475 
  1476   Theory.axioms_of: theory -> (string * term) list
  1477   Theory.all_axioms_of: theory -> (string * term) list
  1478   PureThy.thms_of: theory -> (string * thm) list
  1479   PureThy.all_thms_of: theory -> (string * thm) list
  1480 
  1481 * Pure: print_tac now outputs the goal through the trace channel.
  1482 
  1483 * Isar toplevel: improved diagnostics, mostly for Poly/ML only.
  1484 Reference Toplevel.debug (default false) controls detailed printing
  1485 and tracing of low-level exceptions; Toplevel.profiling (default 0)
  1486 controls execution profiling -- set to 1 for time and 2 for space
  1487 (both increase the runtime).
  1488 
  1489 * Isar session: The initial use of ROOT.ML is now always timed,
  1490 i.e. the log will show the actual process times, in contrast to the
  1491 elapsed wall-clock time that the outer shell wrapper produces.
  1492 
  1493 * Simplifier: improved handling of bound variables (nameless
  1494 representation, avoid allocating new strings).  Simprocs that invoke
  1495 the Simplifier recursively should use Simplifier.inherit_bounds to
  1496 avoid local name clashes.  Failure to do so produces warnings
  1497 "Simplifier: renamed bound variable ..."; set Simplifier.debug_bounds
  1498 for further details.
  1499 
  1500 * ML functions legacy_bindings and use_legacy_bindings produce ML fact
  1501 bindings for all theorems stored within a given theory; this may help
  1502 in porting non-Isar theories to Isar ones, while keeping ML proof
  1503 scripts for the time being.
  1504 
  1505 * ML operator HTML.with_charset specifies the charset begin used for
  1506 generated HTML files.  For example:
  1507 
  1508   HTML.with_charset "utf-8" use_thy "Hebrew";
  1509   HTML.with_charset "utf-8" use_thy "Chinese";
  1510 
  1511 
  1512 *** System ***
  1513 
  1514 * Allow symlinks to all proper Isabelle executables (Isabelle,
  1515 isabelle, isatool etc.).
  1516 
  1517 * ISABELLE_DOC_FORMAT setting specifies preferred document format (for
  1518 isatool doc, isatool mkdir, display_drafts etc.).
  1519 
  1520 * isatool usedir: option -f allows specification of the ML file to be
  1521 used by Isabelle; default is ROOT.ML.
  1522 
  1523 * New isatool version outputs the version identifier of the Isabelle
  1524 distribution being used.
  1525 
  1526 * HOL: new isatool dimacs2hol converts files in DIMACS CNF format
  1527 (containing Boolean satisfiability problems) into Isabelle/HOL
  1528 theories.
  1529 
  1530 
  1531 
  1532 New in Isabelle2004 (April 2004)
  1533 --------------------------------
  1534 
  1535 *** General ***
  1536 
  1537 * Provers/order.ML:  new efficient reasoner for partial and linear orders.
  1538   Replaces linorder.ML.
  1539 
  1540 * Pure: Greek letters (except small lambda, \<lambda>), as well as Gothic
  1541   (\<aa>...\<zz>\<AA>...\<ZZ>), calligraphic (\<A>...\<Z>), and Euler
  1542   (\<a>...\<z>), are now considered normal letters, and can therefore
  1543   be used anywhere where an ASCII letter (a...zA...Z) has until
  1544   now. COMPATIBILITY: This obviously changes the parsing of some
  1545   terms, especially where a symbol has been used as a binder, say
  1546   '\<Pi>x. ...', which is now a type error since \<Pi>x will be parsed
  1547   as an identifier.  Fix it by inserting a space around former
  1548   symbols.  Call 'isatool fixgreek' to try to fix parsing errors in
  1549   existing theory and ML files.
  1550 
  1551 * Pure: Macintosh and Windows line-breaks are now allowed in theory files.
  1552 
  1553 * Pure: single letter sub/superscripts (\<^isub> and \<^isup>) are now
  1554   allowed in identifiers. Similar to Greek letters \<^isub> is now considered
  1555   a normal (but invisible) letter. For multiple letter subscripts repeat
  1556   \<^isub> like this: x\<^isub>1\<^isub>2.
  1557 
  1558 * Pure: There are now sub-/superscripts that can span more than one
  1559   character. Text between \<^bsub> and \<^esub> is set in subscript in
  1560   ProofGeneral and LaTeX, text between \<^bsup> and \<^esup> in
  1561   superscript. The new control characters are not identifier parts.
  1562 
  1563 * Pure: Control-symbols of the form \<^raw:...> will literally print the
  1564   content of "..." to the latex file instead of \isacntrl... . The "..."
  1565   may consist of any printable characters excluding the end bracket >.
  1566 
  1567 * Pure: Using new Isar command "finalconsts" (or the ML functions
  1568   Theory.add_finals or Theory.add_finals_i) it is now possible to
  1569   declare constants "final", which prevents their being given a definition
  1570   later.  It is useful for constants whose behaviour is fixed axiomatically
  1571   rather than definitionally, such as the meta-logic connectives.
  1572 
  1573 * Pure: 'instance' now handles general arities with general sorts
  1574   (i.e. intersections of classes),
  1575 
  1576 * Presentation: generated HTML now uses a CSS style sheet to make layout
  1577   (somewhat) independent of content. It is copied from lib/html/isabelle.css.
  1578   It can be changed to alter the colors/layout of generated pages.
  1579 
  1580 
  1581 *** Isar ***
  1582 
  1583 * Tactic emulation methods rule_tac, erule_tac, drule_tac, frule_tac,
  1584   cut_tac, subgoal_tac and thin_tac:
  1585   - Now understand static (Isar) contexts.  As a consequence, users of Isar
  1586     locales are no longer forced to write Isar proof scripts.
  1587     For details see Isar Reference Manual, paragraph 4.3.2: Further tactic
  1588     emulations.
  1589   - INCOMPATIBILITY: names of variables to be instantiated may no
  1590     longer be enclosed in quotes.  Instead, precede variable name with `?'.
  1591     This is consistent with the instantiation attribute "where".
  1592 
  1593 * Attributes "where" and "of":
  1594   - Now take type variables of instantiated theorem into account when reading
  1595     the instantiation string.  This fixes a bug that caused instantiated
  1596     theorems to have too special types in some circumstances.
  1597   - "where" permits explicit instantiations of type variables.
  1598 
  1599 * Calculation commands "moreover" and "also" no longer interfere with
  1600   current facts ("this"), admitting arbitrary combinations with "then"
  1601   and derived forms.
  1602 
  1603 * Locales:
  1604   - Goal statements involving the context element "includes" no longer
  1605     generate theorems with internal delta predicates (those ending on
  1606     "_axioms") in the premise.
  1607     Resolve particular premise with <locale>.intro to obtain old form.
  1608   - Fixed bug in type inference ("unify_frozen") that prevented mix of target
  1609     specification and "includes" elements in goal statement.
  1610   - Rule sets <locale>.intro and <locale>.axioms no longer declared as
  1611     [intro?] and [elim?] (respectively) by default.
  1612   - Experimental command for instantiation of locales in proof contexts:
  1613         instantiate <label>[<attrs>]: <loc>
  1614     Instantiates locale <loc> and adds all its theorems to the current context
  1615     taking into account their attributes.  Label and attrs are optional
  1616     modifiers, like in theorem declarations.  If present, names of
  1617     instantiated theorems are qualified with <label>, and the attributes
  1618     <attrs> are applied after any attributes these theorems might have already.
  1619       If the locale has assumptions, a chained fact of the form
  1620     "<loc> t1 ... tn" is expected from which instantiations of the parameters
  1621     are derived.  The command does not support old-style locales declared
  1622     with "locale (open)".
  1623       A few (very simple) examples can be found in FOL/ex/LocaleInst.thy.
  1624 
  1625 * HOL: Tactic emulation methods induct_tac and case_tac understand static
  1626   (Isar) contexts.
  1627 
  1628 
  1629 *** HOL ***
  1630 
  1631 * Proof import: new image HOL4 contains the imported library from
  1632   the HOL4 system with about 2500 theorems. It is imported by
  1633   replaying proof terms produced by HOL4 in Isabelle. The HOL4 image
  1634   can be used like any other Isabelle image.  See
  1635   HOL/Import/HOL/README for more information.
  1636 
  1637 * Simplifier:
  1638   - Much improved handling of linear and partial orders.
  1639     Reasoners for linear and partial orders are set up for type classes
  1640     "linorder" and "order" respectively, and are added to the default simpset
  1641     as solvers.  This means that the simplifier can build transitivity chains
  1642     to solve goals from the assumptions.
  1643   - INCOMPATIBILITY: old proofs break occasionally.  Typically, applications
  1644     of blast or auto after simplification become unnecessary because the goal
  1645     is solved by simplification already.
  1646 
  1647 * Numerics: new theory Ring_and_Field contains over 250 basic numerical laws,
  1648     all proved in axiomatic type classes for semirings, rings and fields.
  1649 
  1650 * Numerics:
  1651   - Numeric types (nat, int, and in HOL-Complex rat, real, complex, etc.) are
  1652     now formalized using the Ring_and_Field theory mentioned above.
  1653   - INCOMPATIBILITY: simplification and arithmetic behaves somewhat differently
  1654     than before, because now they are set up once in a generic manner.
  1655   - INCOMPATIBILITY: many type-specific arithmetic laws have gone.
  1656     Look for the general versions in Ring_and_Field (and Power if they concern
  1657     exponentiation).
  1658 
  1659 * Type "rat" of the rational numbers is now available in HOL-Complex.
  1660 
  1661 * Records:
  1662   - Record types are now by default printed with their type abbreviation
  1663     instead of the list of all field types. This can be configured via
  1664     the reference "print_record_type_abbr".
  1665   - Simproc "record_upd_simproc" for simplification of multiple updates added
  1666     (not enabled by default).
  1667   - Simproc "record_ex_sel_eq_simproc" to simplify EX x. sel r = x resp.
  1668     EX x. x = sel r to True (not enabled by default).
  1669   - Tactic "record_split_simp_tac" to split and simplify records added.
  1670 
  1671 * 'specification' command added, allowing for definition by
  1672   specification.  There is also an 'ax_specification' command that
  1673   introduces the new constants axiomatically.
  1674 
  1675 * arith(_tac) is now able to generate counterexamples for reals as well.
  1676 
  1677 * HOL-Algebra: new locale "ring" for non-commutative rings.
  1678 
  1679 * HOL-ex: InductiveInvariant_examples illustrates advanced recursive function
  1680   definitions, thanks to Sava Krsti\'{c} and John Matthews.
  1681 
  1682 * HOL-Matrix: a first theory for matrices in HOL with an application of
  1683   matrix theory to linear programming.
  1684 
  1685 * Unions and Intersections:
  1686   The latex output syntax of UN and INT has been changed
  1687   from "\Union x \in A. B" to "\Union_{x \in A} B"
  1688   i.e. the index formulae has become a subscript.
  1689   Similarly for "\Union x. B", and for \Inter instead of \Union.
  1690 
  1691 * Unions and Intersections over Intervals:
  1692   There is new short syntax "UN i<=n. A" for "UN i:{0..n}. A". There is
  1693   also an x-symbol version with subscripts "\<Union>\<^bsub>i <= n\<^esub>. A"
  1694   like in normal math, and corresponding versions for < and for intersection.
  1695 
  1696 * HOL/List: Ordering "lexico" is renamed "lenlex" and the standard
  1697   lexicographic dictonary ordering has been added as "lexord".
  1698 
  1699 * ML: the legacy theory structures Int and List have been removed. They had
  1700   conflicted with ML Basis Library structures having the same names.
  1701 
  1702 * 'refute' command added to search for (finite) countermodels.  Only works
  1703   for a fragment of HOL.  The installation of an external SAT solver is
  1704   highly recommended.  See "HOL/Refute.thy" for details.
  1705 
  1706 * 'quickcheck' command: Allows to find counterexamples by evaluating
  1707   formulae under an assignment of free variables to random values.
  1708   In contrast to 'refute', it can deal with inductive datatypes,
  1709   but cannot handle quantifiers. See "HOL/ex/Quickcheck_Examples.thy"
  1710   for examples.
  1711 
  1712 
  1713 *** HOLCF ***
  1714 
  1715 * Streams now come with concatenation and are part of the HOLCF image
  1716 
  1717 
  1718 
  1719 New in Isabelle2003 (May 2003)
  1720 ------------------------------
  1721 
  1722 *** General ***
  1723 
  1724 * Provers/simplifier:
  1725 
  1726   - Completely reimplemented method simp (ML: Asm_full_simp_tac):
  1727     Assumptions are now subject to complete mutual simplification,
  1728     not just from left to right. The simplifier now preserves
  1729     the order of assumptions.
  1730 
  1731     Potential INCOMPATIBILITY:
  1732 
  1733     -- simp sometimes diverges where the old version did
  1734        not, e.g. invoking simp on the goal
  1735 
  1736         [| P (f x); y = x; f x = f y |] ==> Q
  1737 
  1738        now gives rise to the infinite reduction sequence
  1739 
  1740         P(f x) --(f x = f y)--> P(f y) --(y = x)--> P(f x) --(f x = f y)--> ...
  1741 
  1742        Using "simp (asm_lr)" (ML: Asm_lr_simp_tac) instead often solves this
  1743        kind of problem.
  1744 
  1745     -- Tactics combining classical reasoner and simplification (such as auto)
  1746        are also affected by this change, because many of them rely on
  1747        simp. They may sometimes diverge as well or yield a different numbers
  1748        of subgoals. Try to use e.g. force, fastsimp, or safe instead of auto
  1749        in case of problems. Sometimes subsequent calls to the classical
  1750        reasoner will fail because a preceeding call to the simplifier too
  1751        eagerly simplified the goal, e.g. deleted redundant premises.
  1752 
  1753   - The simplifier trace now shows the names of the applied rewrite rules
  1754 
  1755   - You can limit the number of recursive invocations of the simplifier
  1756     during conditional rewriting (where the simplifie tries to solve the
  1757     conditions before applying the rewrite rule):
  1758     ML "simp_depth_limit := n"
  1759     where n is an integer. Thus you can force termination where previously
  1760     the simplifier would diverge.
  1761 
  1762   - Accepts free variables as head terms in congruence rules.  Useful in Isar.
  1763 
  1764   - No longer aborts on failed congruence proof.  Instead, the
  1765     congruence is ignored.
  1766 
  1767 * Pure: New generic framework for extracting programs from constructive
  1768   proofs. See HOL/Extraction.thy for an example instantiation, as well
  1769   as HOL/Extraction for some case studies.
  1770 
  1771 * Pure: The main goal of the proof state is no longer shown by default, only
  1772 the subgoals. This behaviour is controlled by a new flag.
  1773    PG menu: Isabelle/Isar -> Settings -> Show Main Goal
  1774 (ML: Proof.show_main_goal).
  1775 
  1776 * Pure: You can find all matching introduction rules for subgoal 1, i.e. all
  1777 rules whose conclusion matches subgoal 1:
  1778       PG menu: Isabelle/Isar -> Show me -> matching rules
  1779 The rules are ordered by how closely they match the subgoal.
  1780 In particular, rules that solve a subgoal outright are displayed first
  1781 (or rather last, the way they are printed).
  1782 (ML: ProofGeneral.print_intros())
  1783 
  1784 * Pure: New flag trace_unify_fail causes unification to print
  1785 diagnostic information (PG: in trace buffer) when it fails. This is
  1786 useful for figuring out why single step proofs like rule, erule or
  1787 assumption failed.
  1788 
  1789 * Pure: Locale specifications now produce predicate definitions
  1790 according to the body of text (covering assumptions modulo local
  1791 definitions); predicate "loc_axioms" covers newly introduced text,
  1792 while "loc" is cumulative wrt. all included locale expressions; the
  1793 latter view is presented only on export into the global theory
  1794 context; potential INCOMPATIBILITY, use "(open)" option to fall back
  1795 on the old view without predicates;
  1796 
  1797 * Pure: predefined locales "var" and "struct" are useful for sharing
  1798 parameters (as in CASL, for example); just specify something like
  1799 ``var x + var y + struct M'' as import;
  1800 
  1801 * Pure: improved thms_containing: proper indexing of facts instead of
  1802 raw theorems; check validity of results wrt. current name space;
  1803 include local facts of proof configuration (also covers active
  1804 locales), cover fixed variables in index; may use "_" in term
  1805 specification; an optional limit for the number of printed facts may
  1806 be given (the default is 40);
  1807 
  1808 * Pure: disallow duplicate fact bindings within new-style theory files
  1809 (batch-mode only);
  1810 
  1811 * Provers: improved induct method: assumptions introduced by case
  1812 "foo" are split into "foo.hyps" (from the rule) and "foo.prems" (from
  1813 the goal statement); "foo" still refers to all facts collectively;
  1814 
  1815 * Provers: the function blast.overloaded has been removed: all constants
  1816 are regarded as potentially overloaded, which improves robustness in exchange
  1817 for slight decrease in efficiency;
  1818 
  1819 * Provers/linorder: New generic prover for transitivity reasoning over
  1820 linear orders.  Note: this prover is not efficient!
  1821 
  1822 * Isar: preview of problems to finish 'show' now produce an error
  1823 rather than just a warning (in interactive mode);
  1824 
  1825 
  1826 *** HOL ***
  1827 
  1828 * arith(_tac)
  1829 
  1830  - Produces a counter example if it cannot prove a goal.
  1831    Note that the counter example may be spurious if the goal is not a formula
  1832    of quantifier-free linear arithmetic.
  1833    In ProofGeneral the counter example appears in the trace buffer.
  1834 
  1835  - Knows about div k and mod k where k is a numeral of type nat or int.
  1836 
  1837  - Calls full Presburger arithmetic (by Amine Chaieb) if quantifier-free
  1838    linear arithmetic fails. This takes account of quantifiers and divisibility.
  1839    Presburger arithmetic can also be called explicitly via presburger(_tac).
  1840 
  1841 * simp's arithmetic capabilities have been enhanced a bit: it now
  1842 takes ~= in premises into account (by performing a case split);
  1843 
  1844 * simp reduces "m*(n div m) + n mod m" to n, even if the two summands
  1845 are distributed over a sum of terms;
  1846 
  1847 * New tactic "trans_tac" and method "trans" instantiate
  1848 Provers/linorder.ML for axclasses "order" and "linorder" (predicates
  1849 "<=", "<" and "=").
  1850 
  1851 * function INCOMPATIBILITIES: Pi-sets have been redefined and moved from main
  1852 HOL to Library/FuncSet; constant "Fun.op o" is now called "Fun.comp";
  1853 
  1854 * 'typedef' command has new option "open" to suppress the set
  1855 definition;
  1856 
  1857 * functions Min and Max on finite sets have been introduced (theory
  1858 Finite_Set);
  1859 
  1860 * attribute [symmetric] now works for relations as well; it turns
  1861 (x,y) : R^-1 into (y,x) : R, and vice versa;
  1862 
  1863 * induct over a !!-quantified statement (say !!x1..xn):
  1864   each "case" automatically performs "fix x1 .. xn" with exactly those names.
  1865 
  1866 * Map: `empty' is no longer a constant but a syntactic abbreviation for
  1867 %x. None. Warning: empty_def now refers to the previously hidden definition
  1868 of the empty set.
  1869 
  1870 * Algebra: formalization of classical algebra.  Intended as base for
  1871 any algebraic development in Isabelle.  Currently covers group theory
  1872 (up to Sylow's theorem) and ring theory (Universal Property of
  1873 Univariate Polynomials).  Contributions welcome;
  1874 
  1875 * GroupTheory: deleted, since its material has been moved to Algebra;
  1876 
  1877 * Complex: new directory of the complex numbers with numeric constants,
  1878 nonstandard complex numbers, and some complex analysis, standard and
  1879 nonstandard (Jacques Fleuriot);
  1880 
  1881 * HOL-Complex: new image for analysis, replacing HOL-Real and HOL-Hyperreal;
  1882 
  1883 * Hyperreal: introduced Gauge integration and hyperreal logarithms (Jacques
  1884 Fleuriot);
  1885 
  1886 * Real/HahnBanach: updated and adapted to locales;
  1887 
  1888 * NumberTheory: added Gauss's law of quadratic reciprocity (by Avigad,
  1889 Gray and Kramer);
  1890 
  1891 * UNITY: added the Meier-Sanders theory of progress sets;
  1892 
  1893 * MicroJava: bytecode verifier and lightweight bytecode verifier
  1894 as abstract algorithms, instantiated to the JVM;
  1895 
  1896 * Bali: Java source language formalization. Type system, operational
  1897 semantics, axiomatic semantics. Supported language features:
  1898 classes, interfaces, objects,virtual methods, static methods,
  1899 static/instance fields, arrays, access modifiers, definite
  1900 assignment, exceptions.
  1901 
  1902 
  1903 *** ZF ***
  1904 
  1905 * ZF/Constructible: consistency proof for AC (Gdel's constructible
  1906 universe, etc.);
  1907 
  1908 * Main ZF: virtually all theories converted to new-style format;
  1909 
  1910 
  1911 *** ML ***
  1912 
  1913 * Pure: Tactic.prove provides sane interface for internal proofs;
  1914 omits the infamous "standard" operation, so this is more appropriate
  1915 than prove_goalw_cterm in many situations (e.g. in simprocs);
  1916 
  1917 * Pure: improved error reporting of simprocs;
  1918 
  1919 * Provers: Simplifier.simproc(_i) provides sane interface for setting
  1920 up simprocs;
  1921 
  1922 
  1923 *** Document preparation ***
  1924 
  1925 * uses \par instead of \\ for line breaks in theory text. This may
  1926 shift some page breaks in large documents. To get the old behaviour
  1927 use \renewcommand{\isanewline}{\mbox{}\\\mbox{}} in root.tex.
  1928 
  1929 * minimized dependencies of isabelle.sty and isabellesym.sty on
  1930 other packages
  1931 
  1932 * \<euro> now needs package babel/greek instead of marvosym (which
  1933 broke \Rightarrow)
  1934 
  1935 * normal size for \<zero>...\<nine> (uses \mathbf instead of
  1936 textcomp package)
  1937 
  1938 
  1939 
  1940 New in Isabelle2002 (March 2002)
  1941 --------------------------------
  1942 
  1943 *** Document preparation ***
  1944 
  1945 * greatly simplified document preparation setup, including more
  1946 graceful interpretation of isatool usedir -i/-d/-D options, and more
  1947 instructive isatool mkdir; users should basically be able to get
  1948 started with "isatool mkdir HOL Test && isatool make"; alternatively,
  1949 users may run a separate document processing stage manually like this:
  1950 "isatool usedir -D output HOL Test && isatool document Test/output";
  1951 
  1952 * theory dependency graph may now be incorporated into documents;
  1953 isatool usedir -g true will produce session_graph.eps/.pdf for use
  1954 with \includegraphics of LaTeX;
  1955 
  1956 * proper spacing of consecutive markup elements, especially text
  1957 blocks after section headings;
  1958 
  1959 * support bold style (for single symbols only), input syntax is like
  1960 this: "\<^bold>\<alpha>" or "\<^bold>A";
  1961 
  1962 * \<bullet> is now output as bold \cdot by default, which looks much
  1963 better in printed text;
  1964 
  1965 * added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
  1966 note that these symbols are currently unavailable in Proof General /
  1967 X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
  1968 
  1969 * isatool latex no longer depends on changed TEXINPUTS, instead
  1970 isatool document copies the Isabelle style files to the target
  1971 location;
  1972 
  1973 
  1974 *** Isar ***
  1975 
  1976 * Pure/Provers: improved proof by cases and induction;
  1977   - 'case' command admits impromptu naming of parameters (such as
  1978     "case (Suc n)");
  1979   - 'induct' method divinates rule instantiation from the inductive
  1980     claim; no longer requires excessive ?P bindings for proper
  1981     instantiation of cases;
  1982   - 'induct' method properly enumerates all possibilities of set/type
  1983     rules; as a consequence facts may be also passed through *type*
  1984     rules without further ado;
  1985   - 'induct' method now derives symbolic cases from the *rulified*
  1986     rule (before it used to rulify cases stemming from the internal
  1987     atomized version); this means that the context of a non-atomic
  1988     statement becomes is included in the hypothesis, avoiding the
  1989     slightly cumbersome show "PROP ?case" form;
  1990   - 'induct' may now use elim-style induction rules without chaining
  1991     facts, using ``missing'' premises from the goal state; this allows
  1992     rules stemming from inductive sets to be applied in unstructured
  1993     scripts, while still benefitting from proper handling of non-atomic
  1994     statements; NB: major inductive premises need to be put first, all
  1995     the rest of the goal is passed through the induction;
  1996   - 'induct' proper support for mutual induction involving non-atomic
  1997     rule statements (uses the new concept of simultaneous goals, see
  1998     below);
  1999   - append all possible rule selections, but only use the first
  2000     success (no backtracking);
  2001   - removed obsolete "(simplified)" and "(stripped)" options of methods;
  2002   - undeclared rule case names default to numbers 1, 2, 3, ...;
  2003   - added 'print_induct_rules' (covered by help item in recent Proof
  2004     General versions);
  2005   - moved induct/cases attributes to Pure, methods to Provers;
  2006   - generic method setup instantiated for FOL and HOL;
  2007 
  2008 * Pure: support multiple simultaneous goal statements, for example
  2009 "have a: A and b: B" (same for 'theorem' etc.); being a pure
  2010 meta-level mechanism, this acts as if several individual goals had
  2011 been stated separately; in particular common proof methods need to be
  2012 repeated in order to cover all claims; note that a single elimination
  2013 step is *not* sufficient to establish the two conjunctions, so this
  2014 fails:
  2015 
  2016   assume "A & B" then have A and B ..   (*".." fails*)
  2017 
  2018 better use "obtain" in situations as above; alternative refer to
  2019 multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
  2020 
  2021 * Pure: proper integration with ``locales''; unlike the original
  2022 version by Florian Kammller, Isar locales package high-level proof
  2023 contexts rather than raw logical ones (e.g. we admit to include
  2024 attributes everywhere); operations on locales include merge and
  2025 rename; support for implicit arguments (``structures''); simultaneous
  2026 type-inference over imports and text; see also HOL/ex/Locales.thy for
  2027 some examples;
  2028 
  2029 * Pure: the following commands have been ``localized'', supporting a
  2030 target locale specification "(in name)": 'lemma', 'theorem',
  2031 'corollary', 'lemmas', 'theorems', 'declare'; the results will be
  2032 stored both within the locale and at the theory level (exported and
  2033 qualified by the locale name);
  2034 
  2035 * Pure: theory goals may now be specified in ``long'' form, with
  2036 ad-hoc contexts consisting of arbitrary locale elements. for example
  2037 ``lemma foo: fixes x assumes "A x" shows "B x"'' (local syntax and
  2038 definitions may be given, too); the result is a meta-level rule with
  2039 the context elements being discharged in the obvious way;
  2040 
  2041 * Pure: new proof command 'using' allows to augment currently used
  2042 facts after a goal statement ('using' is syntactically analogous to
  2043 'apply', but acts on the goal's facts only); this allows chained facts
  2044 to be separated into parts given before and after a claim, as in
  2045 ``from a and b have C using d and e <proof>'';
  2046 
  2047 * Pure: renamed "antecedent" case to "rule_context";
  2048 
  2049 * Pure: new 'judgment' command records explicit information about the
  2050 object-logic embedding (used by several tools internally); no longer
  2051 use hard-wired "Trueprop";
  2052 
  2053 * Pure: added 'corollary' command;
  2054 
  2055 * Pure: fixed 'token_translation' command;
  2056 
  2057 * Pure: removed obsolete 'exported' attribute;
  2058 
  2059 * Pure: dummy pattern "_" in is/let is now automatically lifted over
  2060 bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
  2061 supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
  2062 
  2063 * Pure: method 'atomize' presents local goal premises as object-level
  2064 statements (atomic meta-level propositions); setup controlled via
  2065 rewrite rules declarations of 'atomize' attribute; example
  2066 application: 'induct' method with proper rule statements in improper
  2067 proof *scripts*;
  2068 
  2069 * Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
  2070 now consider the syntactic context of assumptions, giving a better
  2071 chance to get type-inference of the arguments right (this is
  2072 especially important for locales);
  2073 
  2074 * Pure: "sorry" no longer requires quick_and_dirty in interactive
  2075 mode;
  2076 
  2077 * Pure/obtain: the formal conclusion "thesis", being marked as
  2078 ``internal'', may no longer be reference directly in the text;
  2079 potential INCOMPATIBILITY, may need to use "?thesis" in rare
  2080 situations;
  2081 
  2082 * Pure: generic 'sym' attribute which declares a rule both as pure
  2083 'elim?' and for the 'symmetric' operation;
  2084 
  2085 * Pure: marginal comments ``--'' may now occur just anywhere in the
  2086 text; the fixed correlation with particular command syntax has been
  2087 discontinued;
  2088 
  2089 * Pure: new method 'rules' is particularly well-suited for proof
  2090 search in intuitionistic logic; a bit slower than 'blast' or 'fast',
  2091 but often produces more compact proof terms with less detours;
  2092 
  2093 * Pure/Provers/classical: simplified integration with pure rule
  2094 attributes and methods; the classical "intro?/elim?/dest?"
  2095 declarations coincide with the pure ones; the "rule" method no longer
  2096 includes classically swapped intros; "intro" and "elim" methods no
  2097 longer pick rules from the context; also got rid of ML declarations
  2098 AddXIs/AddXEs/AddXDs; all of this has some potential for
  2099 INCOMPATIBILITY;
  2100 
  2101 * Provers/classical: attribute 'swapped' produces classical inversions
  2102 of introduction rules;
  2103 
  2104 * Provers/simplifier: 'simplified' attribute may refer to explicit
  2105 rules instead of full simplifier context; 'iff' attribute handles
  2106 conditional rules;
  2107 
  2108 * HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
  2109 
  2110 * HOL: 'recdef' now fails on unfinished automated proofs, use
  2111 "(permissive)" option to recover old behavior;
  2112 
  2113 * HOL: 'inductive' no longer features separate (collective) attributes
  2114 for 'intros' (was found too confusing);
  2115 
  2116 * HOL: properly declared induction rules less_induct and
  2117 wf_induct_rule;
  2118 
  2119 
  2120 *** HOL ***
  2121 
  2122 * HOL: moved over to sane numeral syntax; the new policy is as
  2123 follows:
  2124 
  2125   - 0 and 1 are polymorphic constants, which are defined on any
  2126   numeric type (nat, int, real etc.);
  2127 
  2128   - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
  2129   binary representation internally;
  2130 
  2131   - type nat has special constructor Suc, and generally prefers Suc 0
  2132   over 1::nat and Suc (Suc 0) over 2::nat;
  2133 
  2134 This change may cause significant problems of INCOMPATIBILITY; here
  2135 are some hints on converting existing sources:
  2136 
  2137   - due to the new "num" token, "-0" and "-1" etc. are now atomic
  2138   entities, so expressions involving "-" (unary or binary minus) need
  2139   to be spaced properly;
  2140 
  2141   - existing occurrences of "1" may need to be constraint "1::nat" or
  2142   even replaced by Suc 0; similar for old "2";
  2143 
  2144   - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
  2145 
  2146   - remove all special provisions on numerals in proofs;
  2147 
  2148 * HOL: simp rules nat_number expand numerals on nat to Suc/0
  2149 representation (depends on bin_arith_simps in the default context);
  2150 
  2151 * HOL: symbolic syntax for x^2 (numeral 2);
  2152 
  2153 * HOL: the class of all HOL types is now called "type" rather than
  2154 "term"; INCOMPATIBILITY, need to adapt references to this type class
  2155 in axclass/classes, instance/arities, and (usually rare) occurrences
  2156 in typings (of consts etc.); internally the class is called
  2157 "HOL.type", ML programs should refer to HOLogic.typeS;
  2158 
  2159 * HOL/record package improvements:
  2160   - new derived operations "fields" to build a partial record section,
  2161     "extend" to promote a fixed record to a record scheme, and
  2162     "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
  2163     declared as simp by default;
  2164   - shared operations ("more", "fields", etc.) now need to be always
  2165     qualified) --- potential INCOMPATIBILITY;
  2166   - removed "make_scheme" operations (use "make" with "extend") --
  2167     INCOMPATIBILITY;
  2168   - removed "more" class (simply use "term") -- INCOMPATIBILITY;
  2169   - provides cases/induct rules for use with corresponding Isar
  2170     methods (for concrete records, record schemes, concrete more
  2171     parts, and schematic more parts -- in that order);
  2172   - internal definitions directly based on a light-weight abstract
  2173     theory of product types over typedef rather than datatype;
  2174 
  2175 * HOL: generic code generator for generating executable ML code from
  2176 specifications; specific support for HOL constructs such as inductive
  2177 datatypes and sets, as well as recursive functions; can be invoked
  2178 via 'generate_code' theory section;
  2179 
  2180 * HOL: canonical cases/induct rules for n-tuples (n = 3..7);
  2181 
  2182 * HOL: consolidated and renamed several theories.  In particular:
  2183         Ord.thy has been absorbed into HOL.thy
  2184         String.thy has been absorbed into List.thy
  2185 
  2186 * HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
  2187 (beware of argument permutation!);
  2188 
  2189 * HOL: linorder_less_split superseded by linorder_cases;
  2190 
  2191 * HOL/List: "nodups" renamed to "distinct";
  2192 
  2193 * HOL: added "The" definite description operator; move Hilbert's "Eps"
  2194 to peripheral theory "Hilbert_Choice"; some INCOMPATIBILITIES:
  2195   - Ex_def has changed, now need to use some_eq_ex
  2196 
  2197 * HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
  2198 in this (rare) case use:
  2199 
  2200   delSWrapper "split_all_tac"
  2201   addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
  2202 
  2203 * HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
  2204 MAY FAIL;
  2205 
  2206 * HOL: introduced f^n = f o ... o f; warning: due to the limits of
  2207 Isabelle's type classes, ^ on functions and relations has too general
  2208 a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
  2209 necessary to attach explicit type constraints;
  2210 
  2211 * HOL/Relation: the prefix name of the infix "O" has been changed from
  2212 "comp" to "rel_comp"; INCOMPATIBILITY: a few theorems have been
  2213 renamed accordingly (eg "compI" -> "rel_compI").
  2214 
  2215 * HOL: syntax translations now work properly with numerals and records
  2216 expressions;
  2217 
  2218 * HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
  2219 of "lam" -- INCOMPATIBILITY;
  2220 
  2221 * HOL: got rid of some global declarations (potential INCOMPATIBILITY
  2222 for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
  2223 renamed "Product_Type.unit";
  2224 
  2225 * HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
  2226 
  2227 * HOL: removed obsolete theorem "optionE" (use "option.exhaust", or
  2228 the "cases" method);
  2229 
  2230 * HOL/GroupTheory: group theory examples including Sylow's theorem (by
  2231 Florian Kammller);
  2232 
  2233 * HOL/IMP: updated and converted to new-style theory format; several
  2234 parts turned into readable document, with proper Isar proof texts and
  2235 some explanations (by Gerwin Klein);
  2236 
  2237 * HOL-Real: added Complex_Numbers (by Gertrud Bauer);
  2238 
  2239 * HOL-Hyperreal is now a logic image;
  2240 
  2241 
  2242 *** HOLCF ***
  2243 
  2244 * Isar: consts/constdefs supports mixfix syntax for continuous
  2245 operations;
  2246 
  2247 * Isar: domain package adapted to new-style theory format, e.g. see
  2248 HOLCF/ex/Dnat.thy;
  2249 
  2250 * theory Lift: proper use of rep_datatype lift instead of ML hacks --
  2251 potential INCOMPATIBILITY; now use plain induct_tac instead of former
  2252 lift.induct_tac, always use UU instead of Undef;
  2253 
  2254 * HOLCF/IMP: updated and converted to new-style theory;
  2255 
  2256 
  2257 *** ZF ***
  2258 
  2259 * Isar: proper integration of logic-specific tools and packages,
  2260 including theory commands '(co)inductive', '(co)datatype',
  2261 'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
  2262 'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
  2263 
  2264 * theory Main no longer includes AC; for the Axiom of Choice, base
  2265 your theory on Main_ZFC;
  2266 
  2267 * the integer library now covers quotients and remainders, with many
  2268 laws relating division to addition, multiplication, etc.;
  2269 
  2270 * ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
  2271 typeless version of the formalism;
  2272 
  2273 * ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory
  2274 format;
  2275 
  2276 * ZF/Induct: new directory for examples of inductive definitions,
  2277 including theory Multiset for multiset orderings; converted to
  2278 new-style theory format;
  2279 
  2280 * ZF: many new theorems about lists, ordinals, etc.;
  2281 
  2282 
  2283 *** General ***
  2284 
  2285 * Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
  2286 variable proof controls level of detail: 0 = no proofs (only oracle
  2287 dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
  2288 also ref manual for further ML interfaces;
  2289 
  2290 * Pure/axclass: removed obsolete ML interface
  2291 goal_subclass/goal_arity;
  2292 
  2293 * Pure/syntax: new token syntax "num" for plain numerals (without "#"
  2294 of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
  2295 separate tokens, so expressions involving minus need to be spaced
  2296 properly;
  2297 
  2298 * Pure/syntax: support non-oriented infixes, using keyword "infix"
  2299 rather than "infixl" or "infixr";
  2300 
  2301 * Pure/syntax: concrete syntax for dummy type variables admits genuine
  2302 sort constraint specifications in type inference; e.g. "x::_::foo"
  2303 ensures that the type of "x" is of sort "foo" (but not necessarily a
  2304 type variable);
  2305 
  2306 * Pure/syntax: print modes "type_brackets" and "no_type_brackets"
  2307 control output of nested => (types); the default behavior is
  2308 "type_brackets";
  2309 
  2310 * Pure/syntax: builtin parse translation for "_constify" turns valued
  2311 tokens into AST constants;
  2312 
  2313 * Pure/syntax: prefer later declarations of translations and print
  2314 translation functions; potential INCOMPATIBILITY: need to reverse
  2315 multiple declarations for same syntax element constant;
  2316 
  2317 * Pure/show_hyps reset by default (in accordance to existing Isar
  2318 practice);
  2319 
  2320 * Provers/classical: renamed addaltern to addafter, addSaltern to
  2321 addSafter;
  2322 
  2323 * Provers/clasimp: ``iff'' declarations now handle conditional rules
  2324 as well;
  2325 
  2326 * system: tested support for MacOS X; should be able to get Isabelle +
  2327 Proof General to work in a plain Terminal after installing Poly/ML
  2328 (e.g. from the Isabelle distribution area) and GNU bash alone
  2329 (e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
  2330 support requires further installations, e.g. from
  2331 http://fink.sourceforge.net/);
  2332 
  2333 * system: support Poly/ML 4.1.1 (able to manage larger heaps);
  2334 
  2335 * system: reduced base memory usage by Poly/ML (approx. 20 MB instead
  2336 of 40 MB), cf. ML_OPTIONS;
  2337 
  2338 * system: Proof General keywords specification is now part of the
  2339 Isabelle distribution (see etc/isar-keywords.el);
  2340 
  2341 * system: support for persistent Proof General sessions (refrain from
  2342 outdating all loaded theories on startup); user may create writable
  2343 logic images like this: ``isabelle -q HOL Test'';
  2344 
  2345 * system: smart selection of Isabelle process versus Isabelle
  2346 interface, accommodates case-insensitive file systems (e.g. HFS+); may
  2347 run both "isabelle" and "Isabelle" even if file names are badly
  2348 damaged (executable inspects the case of the first letter of its own
  2349 name); added separate "isabelle-process" and "isabelle-interface";
  2350 
  2351 * system: refrain from any attempt at filtering input streams; no
  2352 longer support ``8bit'' encoding of old isabelle font, instead proper
  2353 iso-latin characters may now be used; the related isatools
  2354 "symbolinput" and "nonascii" have disappeared as well;
  2355 
  2356 * system: removed old "xterm" interface (the print modes "xterm" and
  2357 "xterm_color" are still available for direct use in a suitable
  2358 terminal);
  2359 
  2360 
  2361 
  2362 New in Isabelle99-2 (February 2001)
  2363 -----------------------------------
  2364 
  2365 *** Overview of INCOMPATIBILITIES ***
  2366 
  2367 * HOL: please note that theories in the Library and elsewhere often use the
  2368 new-style (Isar) format; to refer to their theorems in an ML script you must
  2369 bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
  2370 
  2371 * HOL: inductive package no longer splits induction rule aggressively,
  2372 but only as far as specified by the introductions given; the old
  2373 format may be recovered via ML function complete_split_rule or attribute
  2374 'split_rule (complete)';
  2375 
  2376 * HOL: induct renamed to lfp_induct, lfp_Tarski to lfp_unfold,
  2377 gfp_Tarski to gfp_unfold;
  2378 
  2379 * HOL: contrapos, contrapos2 renamed to contrapos_nn, contrapos_pp;
  2380 
  2381 * HOL: infix "dvd" now has priority 50 rather than 70 (because it is a
  2382 relation); infix "^^" has been renamed "``"; infix "``" has been
  2383 renamed "`"; "univalent" has been renamed "single_valued";
  2384 
  2385 * HOL/Real: "rinv" and "hrinv" replaced by overloaded "inverse"
  2386 operation;
  2387 
  2388 * HOLCF: infix "`" has been renamed "$"; the symbol syntax is \<cdot>;
  2389 
  2390 * Isar: 'obtain' no longer declares "that" fact as simp/intro;
  2391 
  2392 * Isar/HOL: method 'induct' now handles non-atomic goals; as a
  2393 consequence, it is no longer monotonic wrt. the local goal context
  2394 (which is now passed through the inductive cases);
  2395 
  2396 * Document preparation: renamed standard symbols \<ll> to \<lless> and
  2397 \<gg> to \<ggreater>;
  2398 
  2399 
  2400 *** Document preparation ***
  2401 
  2402 * \isabellestyle{NAME} selects version of Isabelle output (currently
  2403 available: are "it" for near math-mode best-style output, "sl" for
  2404 slanted text style, and "tt" for plain type-writer; if no
  2405 \isabellestyle command is given, output is according to slanted
  2406 type-writer);
  2407 
  2408 * support sub/super scripts (for single symbols only), input syntax is
  2409 like this: "A\<^sup>*" or "A\<^sup>\<star>";
  2410 
  2411 * some more standard symbols; see Appendix A of the system manual for
  2412 the complete list of symbols defined in isabellesym.sty;
  2413 
  2414 * improved isabelle style files; more abstract symbol implementation
  2415 (should now use \isamath{...} and \isatext{...} in custom symbol
  2416 definitions);
  2417 
  2418 * antiquotation @{goals} and @{subgoals} for output of *dynamic* goals
  2419 state; Note that presentation of goal states does not conform to
  2420 actual human-readable proof documents.  Please do not include goal
  2421 states into document output unless you really know what you are doing!
  2422 
  2423 * proper indentation of antiquoted output with proportional LaTeX
  2424 fonts;
  2425 
  2426 * no_document ML operator temporarily disables LaTeX document
  2427 generation;
  2428 
  2429 * isatool unsymbolize tunes sources for plain ASCII communication;
  2430 
  2431 
  2432 *** Isar ***
  2433 
  2434 * Pure: Isar now suffers initial goal statements to contain unbound
  2435 schematic variables (this does not conform to actual readable proof
  2436 documents, due to unpredictable outcome and non-compositional proof
  2437 checking); users who know what they are doing may use schematic goals
  2438 for Prolog-style synthesis of proven results;
  2439 
  2440 * Pure: assumption method (an implicit finishing) now handles actual
  2441 rules as well;
  2442 
  2443 * Pure: improved 'obtain' --- moved to Pure, insert "that" into
  2444 initial goal, declare "that" only as Pure intro (only for single
  2445 steps); the "that" rule assumption may now be involved in implicit
  2446 finishing, thus ".." becomes a feasible for trivial obtains;
  2447 
  2448 * Pure: default proof step now includes 'intro_classes'; thus trivial
  2449 instance proofs may be performed by "..";
  2450 
  2451 * Pure: ?thesis / ?this / "..." now work for pure meta-level
  2452 statements as well;
  2453 
  2454 * Pure: more robust selection of calculational rules;
  2455 
  2456 * Pure: the builtin notion of 'finished' goal now includes the ==-refl
  2457 rule (as well as the assumption rule);
  2458 
  2459 * Pure: 'thm_deps' command visualizes dependencies of theorems and
  2460 lemmas, using the graph browser tool;
  2461 
  2462 * Pure: predict failure of "show" in interactive mode;
  2463 
  2464 * Pure: 'thms_containing' now takes actual terms as arguments;
  2465 
  2466 * HOL: improved method 'induct' --- now handles non-atomic goals
  2467 (potential INCOMPATIBILITY); tuned error handling;
  2468 
  2469 * HOL: cases and induct rules now provide explicit hints about the
  2470 number of facts to be consumed (0 for "type" and 1 for "set" rules);
  2471 any remaining facts are inserted into the goal verbatim;
  2472 
  2473 * HOL: local contexts (aka cases) may now contain term bindings as
  2474 well; the 'cases' and 'induct' methods new provide a ?case binding for
  2475 the result to be shown in each case;
  2476 
  2477 * HOL: added 'recdef_tc' command;
  2478 
  2479 * isatool convert assists in eliminating legacy ML scripts;
  2480 
  2481 
  2482 *** HOL ***
  2483 
  2484 * HOL/Library: a collection of generic theories to be used together
  2485 with main HOL; the theory loader path already includes this directory
  2486 by default; the following existing theories have been moved here:
  2487 HOL/Induct/Multiset, HOL/Induct/Acc (as Accessible_Part), HOL/While
  2488 (as While_Combinator), HOL/Lex/Prefix (as List_Prefix);
  2489 
  2490 * HOL/Unix: "Some aspects of Unix file-system security", a typical
  2491 modelling and verification task performed in Isabelle/HOL +
  2492 Isabelle/Isar + Isabelle document preparation (by Markus Wenzel).
  2493 
  2494 * HOL/Algebra: special summation operator SUM no longer exists, it has
  2495 been replaced by setsum; infix 'assoc' now has priority 50 (like
  2496 'dvd'); axiom 'one_not_zero' has been moved from axclass 'ring' to
  2497 'domain', this makes the theory consistent with mathematical
  2498 literature;
  2499 
  2500 * HOL basics: added overloaded operations "inverse" and "divide"
  2501 (infix "/"), syntax for generic "abs" operation, generic summation
  2502 operator \<Sum>;
  2503 
  2504 * HOL/typedef: simplified package, provide more useful rules (see also
  2505 HOL/subset.thy);
  2506 
  2507 * HOL/datatype: induction rule for arbitrarily branching datatypes is
  2508 now expressed as a proper nested rule (old-style tactic scripts may
  2509 require atomize_strip_tac to cope with non-atomic premises);
  2510 
  2511 * HOL: renamed theory "Prod" to "Product_Type", renamed "split" rule
  2512 to "split_conv" (old name still available for compatibility);
  2513 
  2514 * HOL: improved concrete syntax for strings (e.g. allows translation
  2515 rules with string literals);
  2516 
  2517 * HOL-Real-Hyperreal: this extends HOL-Real with the hyperreals
  2518  and Fleuriot's mechanization of analysis, including the transcendental
  2519  functions for the reals;
  2520 
  2521 * HOL/Real, HOL/Hyperreal: improved arithmetic simplification;
  2522 
  2523 
  2524 *** CTT ***
  2525 
  2526 * CTT: x-symbol support for Pi, Sigma, -->, : (membership); note that
  2527 "lam" is displayed as TWO lambda-symbols
  2528 
  2529 * CTT: theory Main now available, containing everything (that is, Bool
  2530 and Arith);
  2531 
  2532 
  2533 *** General ***
  2534 
  2535 * Pure: the Simplifier has been implemented properly as a derived rule
  2536 outside of the actual kernel (at last!); the overall performance
  2537 penalty in practical applications is about 50%, while reliability of
  2538 the Isabelle inference kernel has been greatly improved;
  2539 
  2540 * print modes "brackets" and "no_brackets" control output of nested =>
  2541 (types) and ==> (props); the default behaviour is "brackets";
  2542 
  2543 * Provers: fast_tac (and friends) now handle actual object-logic rules
  2544 as assumptions as well;
  2545 
  2546 * system: support Poly/ML 4.0;
  2547 
  2548 * system: isatool install handles KDE version 1 or 2;
  2549 
  2550 
  2551 
  2552 New in Isabelle99-1 (October 2000)
  2553 ----------------------------------
  2554 
  2555 *** Overview of INCOMPATIBILITIES ***
  2556 
  2557 * HOL: simplification of natural numbers is much changed; to partly
  2558 recover the old behaviour (e.g. to prevent n+n rewriting to #2*n)
  2559 issue the following ML commands:
  2560 
  2561   Delsimprocs Nat_Numeral_Simprocs.cancel_numerals;
  2562   Delsimprocs [Nat_Numeral_Simprocs.combine_numerals];
  2563 
  2564 * HOL: simplification no longer dives into case-expressions; this is
  2565 controlled by "t.weak_case_cong" for each datatype t;
  2566 
  2567 * HOL: nat_less_induct renamed to less_induct;
  2568 
  2569 * HOL: systematic renaming of the SOME (Eps) rules, may use isatool
  2570 fixsome to patch .thy and .ML sources automatically;
  2571 
  2572   select_equality  -> some_equality
  2573   select_eq_Ex     -> some_eq_ex
  2574   selectI2EX       -> someI2_ex
  2575   selectI2         -> someI2
  2576   selectI          -> someI
  2577   select1_equality -> some1_equality
  2578   Eps_sym_eq       -> some_sym_eq_trivial
  2579   Eps_eq           -> some_eq_trivial
  2580 
  2581 * HOL: exhaust_tac on datatypes superceded by new generic case_tac;
  2582 
  2583 * HOL: removed obsolete theorem binding expand_if (refer to split_if
  2584 instead);
  2585 
  2586 * HOL: the recursion equations generated by 'recdef' are now called
  2587 f.simps instead of f.rules;
  2588 
  2589 * HOL: qed_spec_mp now also handles bounded ALL as well;
  2590 
  2591 * HOL: 0 is now overloaded, so the type constraint ":: nat" may
  2592 sometimes be needed;
  2593 
  2594 * HOL: the constant for "f``x" is now "image" rather than "op ``";
  2595 
  2596 * HOL: the constant for "f-``x" is now "vimage" rather than "op -``";
  2597 
  2598 * HOL: the disjoint sum is now "<+>" instead of "Plus"; the cartesian
  2599 product is now "<*>" instead of "Times"; the lexicographic product is
  2600 now "<*lex*>" instead of "**";
  2601 
  2602 * HOL: theory Sexp is now in HOL/Induct examples (it used to be part
  2603 of main HOL, but was unused); better use HOL's datatype package;
  2604 
  2605 * HOL: removed "symbols" syntax for constant "override" of theory Map;
  2606 the old syntax may be recovered as follows:
  2607 
  2608   syntax (symbols)
  2609     override  :: "('a ~=> 'b) => ('a ~=> 'b) => ('a ~=> 'b)"
  2610       (infixl "\\<oplus>" 100)
  2611 
  2612 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  2613 
  2614 * HOL/ML: even fewer consts are declared as global (see theories Ord,
  2615 Lfp, Gfp, WF); this only affects ML packages that refer to const names
  2616 internally;
  2617 
  2618 * HOL and ZF: syntax for quotienting wrt an equivalence relation
  2619 changed from A/r to A//r;
  2620 
  2621 * ZF: new treatment of arithmetic (nat & int) may break some old
  2622 proofs;
  2623 
  2624 * Isar: renamed some attributes (RS -> THEN, simplify -> simplified,
  2625 rulify -> rule_format, elimify -> elim_format, ...);
  2626 
  2627 * Isar/Provers: intro/elim/dest attributes changed; renamed
  2628 intro/intro!/intro!! flags to intro!/intro/intro? (in most cases, one
  2629 should have to change intro!! to intro? only); replaced "delrule" by
  2630 "rule del";
  2631 
  2632 * Isar/HOL: renamed "intrs" to "intros" in inductive definitions;
  2633 
  2634 * Provers: strengthened force_tac by using new first_best_tac;
  2635 
  2636 * LaTeX document preparation: several changes of isabelle.sty (see
  2637 lib/texinputs);
  2638 
  2639 
  2640 *** Document preparation ***
  2641 
  2642 * formal comments (text blocks etc.) in new-style theories may now
  2643 contain antiquotations of thm/prop/term/typ/text to be presented
  2644 according to latex print mode; concrete syntax is like this:
  2645 @{term[show_types] "f(x) = a + x"};
  2646 
  2647 * isatool mkdir provides easy setup of Isabelle session directories,
  2648 including proper document sources;
  2649 
  2650 * generated LaTeX sources are now deleted after successful run
  2651 (isatool document -c); may retain a copy somewhere else via -D option
  2652 of isatool usedir;
  2653 
  2654 * isatool usedir -D now lets isatool latex -o sty update the Isabelle
  2655 style files, achieving self-contained LaTeX sources and simplifying
  2656 LaTeX debugging;
  2657 
  2658 * old-style theories now produce (crude) LaTeX output as well;
  2659 
  2660 * browser info session directories are now self-contained (may be put
  2661 on WWW server seperately); improved graphs of nested sessions; removed
  2662 graph for 'all sessions';
  2663 
  2664 * several improvements in isabelle style files; \isabellestyle{it}
  2665 produces fake math mode output; \isamarkupheader is now \section by
  2666 default; see lib/texinputs/isabelle.sty etc.;
  2667 
  2668 
  2669 *** Isar ***
  2670 
  2671 * Isar/Pure: local results and corresponding term bindings are now
  2672 subject to Hindley-Milner polymorphism (similar to ML); this
  2673 accommodates incremental type-inference very nicely;
  2674 
  2675 * Isar/Pure: new derived language element 'obtain' supports
  2676 generalized existence reasoning;
  2677 
  2678 * Isar/Pure: new calculational elements 'moreover' and 'ultimately'
  2679 support accumulation of results, without applying any rules yet;
  2680 useful to collect intermediate results without explicit name
  2681 references, and for use with transitivity rules with more than 2
  2682 premises;
  2683 
  2684 * Isar/Pure: scalable support for case-analysis type proofs: new
  2685 'case' language element refers to local contexts symbolically, as
  2686 produced by certain proof methods; internally, case names are attached
  2687 to theorems as "tags";
  2688 
  2689 * Isar/Pure: theory command 'hide' removes declarations from
  2690 class/type/const name spaces;
  2691 
  2692 * Isar/Pure: theory command 'defs' supports option "(overloaded)" to
  2693 indicate potential overloading;
  2694 
  2695 * Isar/Pure: changed syntax of local blocks from {{ }} to { };
  2696 
  2697 * Isar/Pure: syntax of sorts made 'inner', i.e. have to write
  2698 "{a,b,c}" instead of {a,b,c};
  2699 
  2700 * Isar/Pure now provides its own version of intro/elim/dest
  2701 attributes; useful for building new logics, but beware of confusion
  2702 with the version in Provers/classical;
  2703 
  2704 * Isar/Pure: the local context of (non-atomic) goals is provided via
  2705 case name 'antecedent';
  2706 
  2707 * Isar/Pure: removed obsolete 'transfer' attribute (transfer of thms
  2708 to the current context is now done automatically);
  2709 
  2710 * Isar/Pure: theory command 'method_setup' provides a simple interface
  2711 for definining proof methods in ML;
  2712 
  2713 * Isar/Provers: intro/elim/dest attributes changed; renamed
  2714 intro/intro!/intro!! flags to intro!/intro/intro? (INCOMPATIBILITY, in
  2715 most cases, one should have to change intro!! to intro? only);
  2716 replaced "delrule" by "rule del";
  2717 
  2718 * Isar/Provers: new 'hypsubst' method, plain 'subst' method and
  2719 'symmetric' attribute (the latter supercedes [RS sym]);
  2720 
  2721 * Isar/Provers: splitter support (via 'split' attribute and 'simp'
  2722 method modifier); 'simp' method: 'only:' modifier removes loopers as
  2723 well (including splits);
  2724 
  2725 * Isar/Provers: Simplifier and Classical methods now support all kind
  2726 of modifiers used in the past, including 'cong', 'iff', etc.
  2727 
  2728 * Isar/Provers: added 'fastsimp' and 'clarsimp' methods (combination
  2729 of Simplifier and Classical reasoner);
  2730 
  2731 * Isar/HOL: new proof method 'cases' and improved version of 'induct'
  2732 now support named cases; major packages (inductive, datatype, primrec,
  2733 recdef) support case names and properly name parameters;
  2734 
  2735 * Isar/HOL: new transitivity rules for substitution in inequalities --
  2736 monotonicity conditions are extracted to be proven at end of
  2737 calculations;
  2738 
  2739 * Isar/HOL: removed 'case_split' thm binding, should use 'cases' proof
  2740 method anyway;
  2741 
  2742 * Isar/HOL: removed old expand_if = split_if; theorems if_splits =
  2743 split_if split_if_asm; datatype package provides theorems foo.splits =
  2744 foo.split foo.split_asm for each datatype;
  2745 
  2746 * Isar/HOL: tuned inductive package, rename "intrs" to "intros"
  2747 (potential INCOMPATIBILITY), emulation of mk_cases feature for proof
  2748 scripts: new 'inductive_cases' command and 'ind_cases' method; (Note:
  2749 use "(cases (simplified))" method in proper proof texts);
  2750 
  2751 * Isar/HOL: added global 'arith_split' attribute for 'arith' method;
  2752 
  2753 * Isar: names of theorems etc. may be natural numbers as well;
  2754 
  2755 * Isar: 'pr' command: optional arguments for goals_limit and
  2756 ProofContext.prems_limit; no longer prints theory contexts, but only
  2757 proof states;
  2758 
  2759 * Isar: diagnostic commands 'pr', 'thm', 'prop', 'term', 'typ' admit
  2760 additional print modes to be specified; e.g. "pr(latex)" will print
  2761 proof state according to the Isabelle LaTeX style;
  2762 
  2763 * Isar: improved support for emulating tactic scripts, including proof
  2764 methods 'rule_tac' etc., 'cut_tac', 'thin_tac', 'subgoal_tac',
  2765 'rename_tac', 'rotate_tac', 'tactic', and 'case_tac' / 'induct_tac'
  2766 (for HOL datatypes);
  2767 
  2768 * Isar: simplified (more robust) goal selection of proof methods: 1st
  2769 goal, all goals, or explicit goal specifier (tactic emulation); thus
  2770 'proof method scripts' have to be in depth-first order;
  2771 
  2772 * Isar: tuned 'let' syntax: replaced 'as' keyword by 'and';
  2773 
  2774 * Isar: removed 'help' command, which hasn't been too helpful anyway;
  2775 should instead use individual commands for printing items
  2776 (print_commands, print_methods etc.);
  2777 
  2778 * Isar: added 'nothing' --- the empty list of theorems;
  2779 
  2780 
  2781 *** HOL ***
  2782 
  2783 * HOL/MicroJava: formalization of a fragment of Java, together with a
  2784 corresponding virtual machine and a specification of its bytecode
  2785 verifier and a lightweight bytecode verifier, including proofs of
  2786 type-safety; by Gerwin Klein, Tobias Nipkow, David von Oheimb, and
  2787 Cornelia Pusch (see also the homepage of project Bali at
  2788 http://isabelle.in.tum.de/Bali/);
  2789 
  2790 * HOL/Algebra: new theory of rings and univariate polynomials, by
  2791 Clemens Ballarin;
  2792 
  2793 * HOL/NumberTheory: fundamental Theorem of Arithmetic, Chinese
  2794 Remainder Theorem, Fermat/Euler Theorem, Wilson's Theorem, by Thomas M
  2795 Rasmussen;
  2796 
  2797 * HOL/Lattice: fundamental concepts of lattice theory and order
  2798 structures, including duals, properties of bounds versus algebraic
  2799 laws, lattice operations versus set-theoretic ones, the Knaster-Tarski
  2800 Theorem for complete lattices etc.; may also serve as a demonstration
  2801 for abstract algebraic reasoning using axiomatic type classes, and
  2802 mathematics-style proof in Isabelle/Isar; by Markus Wenzel;
  2803 
  2804 * HOL/Prolog: a (bare-bones) implementation of Lambda-Prolog, by David
  2805 von Oheimb;
  2806 
  2807 * HOL/IMPP: extension of IMP with local variables and mutually
  2808 recursive procedures, by David von Oheimb;
  2809 
  2810 * HOL/Lambda: converted into new-style theory and document;
  2811 
  2812 * HOL/ex/Multiquote: example of multiple nested quotations and
  2813 anti-quotations -- basically a generalized version of de-Bruijn
  2814 representation; very useful in avoiding lifting of operations;
  2815 
  2816 * HOL/record: added general record equality rule to simpset; fixed
  2817 select-update simplification procedure to handle extended records as
  2818 well; admit "r" as field name;
  2819 
  2820 * HOL: 0 is now overloaded over the new sort "zero", allowing its use with
  2821 other numeric types and also as the identity of groups, rings, etc.;
  2822 
  2823 * HOL: new axclass plus_ac0 for addition with the AC-laws and 0 as identity.
  2824 Types nat and int belong to this axclass;
  2825 
  2826 * HOL: greatly improved simplification involving numerals of type nat, int, real:
  2827    (i + #8 + j) = Suc k simplifies to  #7 + (i + j) = k
  2828    i*j + k + j*#3*i     simplifies to  #4*(i*j) + k
  2829   two terms #m*u and #n*u are replaced by #(m+n)*u
  2830     (where #m, #n and u can implicitly be 1; this is simproc combine_numerals)
  2831   and the term/formula #m*u+x ~~ #n*u+y simplifies simplifies to #(m-n)+x ~~ y
  2832     or x ~~ #(n-m)+y, where ~~ is one of = < <= or - (simproc cancel_numerals);
  2833 
  2834 * HOL: meson_tac is available (previously in ex/meson.ML); it is a
  2835 powerful prover for predicate logic but knows nothing of clasets; see
  2836 ex/mesontest.ML and ex/mesontest2.ML for example applications;
  2837 
  2838 * HOL: new version of "case_tac" subsumes both boolean case split and
  2839 "exhaust_tac" on datatypes; INCOMPATIBILITY: exhaust_tac no longer
  2840 exists, may define val exhaust_tac = case_tac for ad-hoc portability;
  2841 
  2842 * HOL: simplification no longer dives into case-expressions: only the
  2843 selector expression is simplified, but not the remaining arms; to
  2844 enable full simplification of case-expressions for datatype t, you may
  2845 remove t.weak_case_cong from the simpset, either globally (Delcongs
  2846 [thm"t.weak_case_cong"];) or locally (delcongs [...]).
  2847 
  2848 * HOL/recdef: the recursion equations generated by 'recdef' for
  2849 function 'f' are now called f.simps instead of f.rules; if all
  2850 termination conditions are proved automatically, these simplification
  2851 rules are added to the simpset, as in primrec; rules may be named
  2852 individually as well, resulting in a separate list of theorems for
  2853 each equation;
  2854 
  2855 * HOL/While is a new theory that provides a while-combinator. It
  2856 permits the definition of tail-recursive functions without the
  2857 provision of a termination measure. The latter is necessary once the
  2858 invariant proof rule for while is applied.
  2859 
  2860 * HOL: new (overloaded) notation for the set of elements below/above
  2861 some element: {..u}, {..u(}, {l..}, {)l..}. See theory SetInterval.
  2862 
  2863 * HOL: theorems impI, allI, ballI bound as "strip";
  2864 
  2865 * HOL: new tactic induct_thm_tac: thm -> string -> int -> tactic
  2866 induct_tac th "x1 ... xn" expects th to have a conclusion of the form
  2867 P v1 ... vn and abbreviates res_inst_tac [("v1","x1"),...,("vn","xn")] th;
  2868 
  2869 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  2870 
  2871 * HOL: theory Sexp now in HOL/Induct examples (it used to be part of
  2872 main HOL, but was unused);
  2873 
  2874 * HOL: fewer consts declared as global (e.g. have to refer to
  2875 "Lfp.lfp" instead of "lfp" internally; affects ML packages only);
  2876 
  2877 * HOL: tuned AST representation of nested pairs, avoiding bogus output
  2878 in case of overlap with user translations (e.g. judgements over
  2879 tuples); (note that the underlying logical represenation is still
  2880 bogus);
  2881 
  2882 
  2883 *** ZF ***
  2884 
  2885 * ZF: simplification automatically cancels common terms in arithmetic
  2886 expressions over nat and int;
  2887 
  2888 * ZF: new treatment of nat to minimize type-checking: all operators
  2889 coerce their operands to a natural number using the function natify,
  2890 making the algebraic laws unconditional;
  2891 
  2892 * ZF: as above, for int: operators coerce their operands to an integer
  2893 using the function intify;
  2894 
  2895 * ZF: the integer library now contains many of the usual laws for the
  2896 orderings, including $<=, and monotonicity laws for $+ and $*;
  2897 
  2898 * ZF: new example ZF/ex/NatSum to demonstrate integer arithmetic
  2899 simplification;
  2900 
  2901 * FOL and ZF: AddIffs now available, giving theorems of the form P<->Q
  2902 to the simplifier and classical reasoner simultaneously;
  2903 
  2904 
  2905 *** General ***
  2906 
  2907 * Provers: blast_tac now handles actual object-logic rules as
  2908 assumptions; note that auto_tac uses blast_tac internally as well;
  2909 
  2910 * Provers: new functions rulify/rulify_no_asm: thm -> thm for turning
  2911 outer -->/All/Ball into ==>/!!; qed_spec_mp now uses rulify_no_asm;
  2912 
  2913 * Provers: delrules now handles destruct rules as well (no longer need
  2914 explicit make_elim);
  2915 
  2916 * Provers: Blast_tac now warns of and ignores "weak elimination rules" e.g.
  2917   [| inj ?f;          ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  2918 use instead the strong form,
  2919   [| inj ?f; ~ ?W ==> ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  2920 in HOL, FOL and ZF the function cla_make_elim will create such rules
  2921 from destruct-rules;
  2922 
  2923 * Provers: Simplifier.easy_setup provides a fast path to basic
  2924 Simplifier setup for new object-logics;
  2925 
  2926 * Pure: AST translation rules no longer require constant head on LHS;
  2927 
  2928 * Pure: improved name spaces: ambiguous output is qualified; support
  2929 for hiding of names;
  2930 
  2931 * system: smart setup of canonical ML_HOME, ISABELLE_INTERFACE, and
  2932 XSYMBOL_HOME; no longer need to do manual configuration in most
  2933 situations;
  2934 
  2935 * system: compression of ML heaps images may now be controlled via -c
  2936 option of isabelle and isatool usedir (currently only observed by
  2937 Poly/ML);
  2938 
  2939 * system: isatool installfonts may handle X-Symbol fonts as well (very
  2940 useful for remote X11);
  2941 
  2942 * system: provide TAGS file for Isabelle sources;
  2943 
  2944 * ML: infix 'OF' is a version of 'MRS' with more appropriate argument
  2945 order;
  2946 
  2947 * ML: renamed flags Syntax.trace_norm_ast to Syntax.trace_ast; global
  2948 timing flag supersedes proof_timing and Toplevel.trace;
  2949 
  2950 * ML: new combinators |>> and |>>> for incremental transformations
  2951 with secondary results (e.g. certain theory extensions):
  2952 
  2953 * ML: PureThy.add_defs gets additional argument to indicate potential
  2954 overloading (usually false);
  2955 
  2956 * ML: PureThy.add_thms/add_axioms/add_defs now return theorems as
  2957 results;
  2958 
  2959 
  2960 
  2961 New in Isabelle99 (October 1999)
  2962 --------------------------------
  2963 
  2964 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  2965 
  2966 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)
  2967 are no longer simplified.  (This allows the simplifier to unfold recursive
  2968 functional programs.)  To restore the old behaviour, declare
  2969 
  2970     Delcongs [if_weak_cong];
  2971 
  2972 * HOL: Removed the obsolete syntax "Compl A"; use -A for set
  2973 complement;
  2974 
  2975 * HOL: the predicate "inj" is now defined by translation to "inj_on";
  2976 
  2977 * HOL/datatype: mutual_induct_tac no longer exists --
  2978   use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]
  2979 
  2980 * HOL/typedef: fixed type inference for representing set; type
  2981 arguments now have to occur explicitly on the rhs as type constraints;
  2982 
  2983 * ZF: The con_defs part of an inductive definition may no longer refer
  2984 to constants declared in the same theory;
  2985 
  2986 * HOL, ZF: the function mk_cases, generated by the inductive
  2987 definition package, has lost an argument.  To simplify its result, it
  2988 uses the default simpset instead of a supplied list of theorems.
  2989 
  2990 * HOL/List: the constructors of type list are now Nil and Cons;
  2991 
  2992 * Simplifier: the type of the infix ML functions
  2993         setSSolver addSSolver setSolver addSolver
  2994 is now  simpset * solver -> simpset  where `solver' is a new abstract type
  2995 for packaging solvers. A solver is created via
  2996         mk_solver: string -> (thm list -> int -> tactic) -> solver
  2997 where the string argument is only a comment.
  2998 
  2999 
  3000 *** Proof tools ***
  3001 
  3002 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a
  3003 decision procedure for linear arithmetic. Currently it is used for
  3004 types `nat', `int', and `real' in HOL (see below); it can, should and
  3005 will be instantiated for other types and logics as well.
  3006 
  3007 * The simplifier now accepts rewrite rules with flexible heads, eg
  3008      hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y
  3009   They are applied like any rule with a non-pattern lhs, i.e. by first-order
  3010   matching.
  3011 
  3012 
  3013 *** General ***
  3014 
  3015 * New Isabelle/Isar subsystem provides an alternative to traditional
  3016 tactical theorem proving; together with the ProofGeneral/isar user
  3017 interface it offers an interactive environment for developing human
  3018 readable proof documents (Isar == Intelligible semi-automated
  3019 reasoning); for further information see isatool doc isar-ref,
  3020 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/
  3021 
  3022 * improved and simplified presentation of theories: better HTML markup
  3023 (including colors), graph views in several sizes; isatool usedir now
  3024 provides a proper interface for user theories (via -P option); actual
  3025 document preparation based on (PDF)LaTeX is available as well (for
  3026 new-style theories only); see isatool doc system for more information;
  3027 
  3028 * native support for Proof General, both for classic Isabelle and
  3029 Isabelle/Isar;
  3030 
  3031 * ML function thm_deps visualizes dependencies of theorems and lemmas,
  3032 using the graph browser tool;
  3033 
  3034 * Isabelle manuals now also available as PDF;
  3035 
  3036 * theory loader rewritten from scratch (may not be fully
  3037 bug-compatible); old loadpath variable has been replaced by show_path,
  3038 add_path, del_path, reset_path functions; new operations such as
  3039 update_thy, touch_thy, remove_thy, use/update_thy_only (see also
  3040 isatool doc ref);
  3041 
  3042 * improved isatool install: option -k creates KDE application icon,
  3043 option -p DIR installs standalone binaries;
  3044 
  3045 * added ML_PLATFORM setting (useful for cross-platform installations);
  3046 more robust handling of platform specific ML images for SML/NJ;
  3047 
  3048 * the settings environment is now statically scoped, i.e. it is never
  3049 created again in sub-processes invoked from isabelle, isatool, or
  3050 Isabelle;
  3051 
  3052 * path element specification '~~' refers to '$ISABELLE_HOME';
  3053 
  3054 * in locales, the "assumes" and "defines" parts may be omitted if
  3055 empty;
  3056 
  3057 * new print_mode "xsymbols" for extended symbol support (e.g. genuine
  3058 long arrows);
  3059 
  3060 * new print_mode "HTML";
  3061 
  3062 * new flag show_tags controls display of tags of theorems (which are
  3063 basically just comments that may be attached by some tools);
  3064 
  3065 * Isamode 2.6 requires patch to accomodate change of Isabelle font
  3066 mode and goal output format:
  3067 
  3068 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el
  3069 244c244
  3070 <       (list (isa-getenv "ISABELLE") "-msymbols" logic-name)
  3071 ---
  3072 >       (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)
  3073 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el
  3074 181c181
  3075 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"
  3076 ---
  3077 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"
  3078 
  3079 * function bind_thms stores lists of theorems (cf. bind_thm);
  3080 
  3081 * new shorthand tactics ftac, eatac, datac, fatac;
  3082 
  3083 * qed (and friends) now accept "" as result name; in that case the
  3084 theorem is not stored, but proper checks and presentation of the
  3085 result still apply;
  3086 
  3087 * theorem database now also indexes constants "Trueprop", "all",
  3088 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;
  3089 
  3090 
  3091 *** HOL ***
  3092 
  3093 ** HOL arithmetic **
  3094 
  3095 * There are now decision procedures for linear arithmetic over nat and
  3096 int:
  3097 
  3098 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',
  3099 `+', `-', `Suc', `min', `max' and numerical constants; other subterms
  3100 are treated as atomic; subformulae not involving type `nat' or `int'
  3101 are ignored; quantified subformulae are ignored unless they are
  3102 positive universal or negative existential. The tactic has to be
  3103 invoked by hand and can be a little bit slow. In particular, the
  3104 running time is exponential in the number of occurrences of `min' and
  3105 `max', and `-' on `nat'.
  3106 
  3107 2. fast_arith_tac is a cut-down version of arith_tac: it only takes
  3108 (negated) (in)equalities among the premises and the conclusion into
  3109 account (i.e. no compound formulae) and does not know about `min' and
  3110 `max', and `-' on `nat'. It is fast and is used automatically by the
  3111 simplifier.
  3112 
  3113 NB: At the moment, these decision procedures do not cope with mixed
  3114 nat/int formulae where the two parts interact, such as `m < n ==>
  3115 int(m) < int(n)'.
  3116 
  3117 * HOL/Numeral provides a generic theory of numerals (encoded
  3118 efficiently as bit strings); setup for types nat/int/real is in place;
  3119 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than
  3120 int, existing theories and proof scripts may require a few additional
  3121 type constraints;
  3122 
  3123 * integer division and remainder can now be performed on constant
  3124 arguments;
  3125 
  3126 * many properties of integer multiplication, division and remainder
  3127 are now available;
  3128 
  3129 * An interface to the Stanford Validity Checker (SVC) is available through the
  3130 tactic svc_tac.  Propositional tautologies and theorems of linear arithmetic
  3131 are proved automatically.  SVC must be installed separately, and its results
  3132 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any
  3133 invocation of the underlying oracle).  For SVC see
  3134   http://verify.stanford.edu/SVC
  3135 
  3136 * IsaMakefile: the HOL-Real target now builds an actual image;
  3137 
  3138 
  3139 ** HOL misc **
  3140 
  3141 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces
  3142 (in Isabelle/Isar) -- by Gertrud Bauer;
  3143 
  3144 * HOL/BCV: generic model of bytecode verification, i.e. data-flow
  3145 analysis for assembly languages with subtypes;
  3146 
  3147 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization
  3148 -- avoids syntactic ambiguities and treats state, transition, and
  3149 temporal levels more uniformly; introduces INCOMPATIBILITIES due to
  3150 changed syntax and (many) tactics;
  3151 
  3152 * HOL/inductive: Now also handles more general introduction rules such
  3153   as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity
  3154   theorems are now maintained within the theory (maintained via the
  3155   "mono" attribute);
  3156 
  3157 * HOL/datatype: Now also handles arbitrarily branching datatypes
  3158   (using function types) such as
  3159 
  3160   datatype 'a tree = Atom 'a | Branch "nat => 'a tree"
  3161 
  3162 * HOL/record: record_simproc (part of the default simpset) takes care
  3163 of selectors applied to updated records; record_split_tac is no longer
  3164 part of the default claset; update_defs may now be removed from the
  3165 simpset in many cases; COMPATIBILITY: old behavior achieved by
  3166 
  3167   claset_ref () := claset() addSWrapper record_split_wrapper;
  3168   Delsimprocs [record_simproc]
  3169 
  3170 * HOL/typedef: fixed type inference for representing set; type
  3171 arguments now have to occur explicitly on the rhs as type constraints;
  3172 
  3173 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem
  3174 names rather than an ML expression;
  3175 
  3176 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be
  3177 supplied later.  Program schemes can be defined, such as
  3178     "While B C s = (if B s then While B C (C s) else s)"
  3179 where the well-founded relation can be chosen after B and C have been given.
  3180 
  3181 * HOL/List: the constructors of type list are now Nil and Cons;
  3182 INCOMPATIBILITY: while [] and infix # syntax is still there, of
  3183 course, ML tools referring to List.list.op # etc. have to be adapted;
  3184 
  3185 * HOL_quantifiers flag superseded by "HOL" print mode, which is
  3186 disabled by default; run isabelle with option -m HOL to get back to
  3187 the original Gordon/HOL-style output;
  3188 
  3189 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,
  3190 ALL x<=y. P, EX x<y. P, EX x<=y. P;
  3191 
  3192 * HOL basic syntax simplified (more orthogonal): all variants of
  3193 All/Ex now support plain / symbolic / HOL notation; plain syntax for
  3194 Eps operator is provided as well: "SOME x. P[x]";
  3195 
  3196 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;
  3197 
  3198 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made
  3199 thus available for user theories;
  3200 
  3201 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with
  3202 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the
  3203 time;
  3204 
  3205 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec
  3206 several times and then mp;
  3207 
  3208 
  3209 *** LK ***
  3210 
  3211 * the notation <<...>> is now available as a notation for sequences of
  3212 formulas;
  3213 
  3214 * the simplifier is now installed
  3215 
  3216 * the axiom system has been generalized (thanks to Soren Heilmann)
  3217 
  3218 * the classical reasoner now has a default rule database
  3219 
  3220 
  3221 *** ZF ***
  3222 
  3223 * new primrec section allows primitive recursive functions to be given
  3224 directly (as in HOL) over datatypes and the natural numbers;
  3225 
  3226 * new tactics induct_tac and exhaust_tac for induction (or case
  3227 analysis) over datatypes and the natural numbers;
  3228 
  3229 * the datatype declaration of type T now defines the recursor T_rec;
  3230 
  3231 * simplification automatically does freeness reasoning for datatype
  3232 constructors;
  3233 
  3234 * automatic type-inference, with AddTCs command to insert new
  3235 type-checking rules;
  3236 
  3237 * datatype introduction rules are now added as Safe Introduction rules
  3238 to the claset;
  3239 
  3240 * the syntax "if P then x else y" is now available in addition to
  3241 if(P,x,y);
  3242 
  3243 
  3244 *** Internal programming interfaces ***
  3245 
  3246 * tuned simplifier trace output; new flag debug_simp;
  3247 
  3248 * structures Vartab / Termtab (instances of TableFun) offer efficient
  3249 tables indexed by indexname_ord / term_ord (compatible with aconv);
  3250 
  3251 * AxClass.axclass_tac lost the theory argument;
  3252 
  3253 * tuned current_goals_markers semantics: begin / end goal avoids
  3254 printing empty lines;
  3255 
  3256 * removed prs and prs_fn hook, which was broken because it did not
  3257 include \n in its semantics, forcing writeln to add one
  3258 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:
  3259 string -> unit if you really want to output text without newline;
  3260 
  3261 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to
  3262 plain output, interface builders may have to enable 'isabelle_font'
  3263 mode to get Isabelle font glyphs as before;
  3264 
  3265 * refined token_translation interface; INCOMPATIBILITY: output length
  3266 now of type real instead of int;
  3267 
  3268 * theory loader actions may be traced via new ThyInfo.add_hook
  3269 interface (see src/Pure/Thy/thy_info.ML); example application: keep
  3270 your own database of information attached to *whole* theories -- as
  3271 opposed to intra-theory data slots offered via TheoryDataFun;
  3272 
  3273 * proper handling of dangling sort hypotheses (at last!);
  3274 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing
  3275 extra sort hypotheses that can be witnessed from the type signature;
  3276 the force_strip_shyps flag is gone, any remaining shyps are simply
  3277 left in the theorem (with a warning issued by strip_shyps_warning);
  3278 
  3279 
  3280 
  3281 New in Isabelle98-1 (October 1998)
  3282 ----------------------------------
  3283 
  3284 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3285 
  3286 * several changes of automated proof tools;
  3287 
  3288 * HOL: major changes to the inductive and datatype packages, including
  3289 some minor incompatibilities of theory syntax;
  3290 
  3291 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now
  3292 called `inj_on';
  3293 
  3294 * HOL: removed duplicate thms in Arith:
  3295   less_imp_add_less  should be replaced by  trans_less_add1
  3296   le_imp_add_le      should be replaced by  trans_le_add1
  3297 
  3298 * HOL: unary minus is now overloaded (new type constraints may be
  3299 required);
  3300 
  3301 * HOL and ZF: unary minus for integers is now #- instead of #~.  In
  3302 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is
  3303 now taken as an integer constant.
  3304 
  3305 * Pure: ML function 'theory_of' renamed to 'theory';
  3306 
  3307 
  3308 *** Proof tools ***
  3309 
  3310 * Simplifier:
  3311   1. Asm_full_simp_tac is now more aggressive.
  3312      1. It will sometimes reorient premises if that increases their power to
  3313         simplify.
  3314      2. It does no longer proceed strictly from left to right but may also
  3315         rotate premises to achieve further simplification.
  3316      For compatibility reasons there is now Asm_lr_simp_tac which is like the
  3317      old Asm_full_simp_tac in that it does not rotate premises.
  3318   2. The simplifier now knows a little bit about nat-arithmetic.
  3319 
  3320 * Classical reasoner: wrapper mechanism for the classical reasoner now
  3321 allows for selected deletion of wrappers, by introduction of names for
  3322 wrapper functionals.  This implies that addbefore, addSbefore,
  3323 addaltern, and addSaltern now take a pair (name, tactic) as argument,
  3324 and that adding two tactics with the same name overwrites the first
  3325 one (emitting a warning).
  3326   type wrapper = (int -> tactic) -> (int -> tactic)
  3327   setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by
  3328   addWrapper, addSWrapper: claset * (string * wrapper) -> claset
  3329   delWrapper, delSWrapper: claset *  string            -> claset
  3330   getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;
  3331 
  3332 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE
  3333 semantics; addbefore now affects only the unsafe part of step_tac
  3334 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY
  3335 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac
  3336 by Force_tac;
  3337 
  3338 * Classical reasoner: setwrapper to setWrapper and compwrapper to
  3339 compWrapper; added safe wrapper (and access functions for it);
  3340 
  3341 * HOL/split_all_tac is now much faster and fails if there is nothing
  3342 to split.  Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order
  3343 and the names of the automatically generated variables have changed.
  3344 split_all_tac has moved within claset() from unsafe wrappers to safe
  3345 wrappers, which means that !!-bound variables are split much more
  3346 aggressively, and safe_tac and clarify_tac now split such variables.
  3347 If this splitting is not appropriate, use delSWrapper "split_all_tac".
  3348 Note: the same holds for record_split_tac, which does the job of
  3349 split_all_tac for record fields.
  3350 
  3351 * HOL/Simplifier: Rewrite rules for case distinctions can now be added
  3352 permanently to the default simpset using Addsplits just like
  3353 Addsimps. They can be removed via Delsplits just like
  3354 Delsimps. Lower-case versions are also available.
  3355 
  3356 * HOL/Simplifier: The rule split_if is now part of the default
  3357 simpset. This means that the simplifier will eliminate all occurrences
  3358 of if-then-else in the conclusion of a goal. To prevent this, you can
  3359 either remove split_if completely from the default simpset by
  3360 `Delsplits [split_if]' or remove it in a specific call of the
  3361 simplifier using `... delsplits [split_if]'.  You can also add/delete
  3362 other case splitting rules to/from the default simpset: every datatype
  3363 generates suitable rules `split_t_case' and `split_t_case_asm' (where
  3364 t is the name of the datatype).
  3365 
  3366 * Classical reasoner / Simplifier combination: new force_tac (and
  3367 derivatives Force_tac, force) combines rewriting and classical
  3368 reasoning (and whatever other tools) similarly to auto_tac, but is
  3369 aimed to solve the given subgoal completely.
  3370 
  3371 
  3372 *** General ***
  3373 
  3374 * new top-level commands `Goal' and `Goalw' that improve upon `goal'
  3375 and `goalw': the theory is no longer needed as an explicit argument -
  3376 the current theory context is used; assumptions are no longer returned
  3377 at the ML-level unless one of them starts with ==> or !!; it is
  3378 recommended to convert to these new commands using isatool fixgoal
  3379 (backup your sources first!);
  3380 
  3381 * new top-level commands 'thm' and 'thms' for retrieving theorems from
  3382 the current theory context, and 'theory' to lookup stored theories;
  3383 
  3384 * new theory section 'locale' for declaring constants, assumptions and
  3385 definitions that have local scope;
  3386 
  3387 * new theory section 'nonterminals' for purely syntactic types;
  3388 
  3389 * new theory section 'setup' for generic ML setup functions
  3390 (e.g. package initialization);
  3391 
  3392 * the distribution now includes Isabelle icons: see
  3393 lib/logo/isabelle-{small,tiny}.xpm;
  3394 
  3395 * isatool install - install binaries with absolute references to
  3396 ISABELLE_HOME/bin;
  3397 
  3398 * isatool logo -- create instances of the Isabelle logo (as EPS);
  3399 
  3400 * print mode 'emacs' reserved for Isamode;
  3401 
  3402 * support multiple print (ast) translations per constant name;
  3403 
  3404 * theorems involving oracles are now printed with a suffixed [!];
  3405 
  3406 
  3407 *** HOL ***
  3408 
  3409 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');
  3410 
  3411 * HOL/inductive package reorganized and improved: now supports mutual
  3412 definitions such as
  3413 
  3414   inductive EVEN ODD
  3415     intrs
  3416       null "0 : EVEN"
  3417       oddI "n : EVEN ==> Suc n : ODD"
  3418       evenI "n : ODD ==> Suc n : EVEN"
  3419 
  3420 new theorem list "elims" contains an elimination rule for each of the
  3421 recursive sets; inductive definitions now handle disjunctive premises
  3422 correctly (also ZF);
  3423 
  3424 INCOMPATIBILITIES: requires Inductive as an ancestor; component
  3425 "mutual_induct" no longer exists - the induction rule is always
  3426 contained in "induct";
  3427 
  3428 
  3429 * HOL/datatype package re-implemented and greatly improved: now
  3430 supports mutually recursive datatypes such as
  3431 
  3432   datatype
  3433     'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)
  3434             | SUM ('a aexp) ('a aexp)
  3435             | DIFF ('a aexp) ('a aexp)
  3436             | NUM 'a
  3437   and
  3438     'a bexp = LESS ('a aexp) ('a aexp)
  3439             | AND ('a bexp) ('a bexp)
  3440             | OR ('a bexp) ('a bexp)
  3441 
  3442 as well as indirectly recursive datatypes such as
  3443 
  3444   datatype
  3445     ('a, 'b) term = Var 'a
  3446                   | App 'b ((('a, 'b) term) list)
  3447 
  3448 The new tactic  mutual_induct_tac [<var_1>, ..., <var_n>] i  performs
  3449 induction on mutually / indirectly recursive datatypes.
  3450 
  3451 Primrec equations are now stored in theory and can be accessed via
  3452 <function_name>.simps.
  3453 
  3454 INCOMPATIBILITIES:
  3455 
  3456   - Theories using datatypes must now have theory Datatype as an
  3457     ancestor.
  3458   - The specific <typename>.induct_tac no longer exists - use the
  3459     generic induct_tac instead.
  3460   - natE has been renamed to nat.exhaust - use exhaust_tac
  3461     instead of res_inst_tac ... natE. Note that the variable
  3462     names in nat.exhaust differ from the names in natE, this
  3463     may cause some "fragile" proofs to fail.
  3464   - The theorems split_<typename>_case and split_<typename>_case_asm
  3465     have been renamed to <typename>.split and <typename>.split_asm.
  3466   - Since default sorts of type variables are now handled correctly,
  3467     some datatype definitions may have to be annotated with explicit
  3468     sort constraints.
  3469   - Primrec definitions no longer require function name and type
  3470     of recursive argument.
  3471 
  3472 Consider using isatool fixdatatype to adapt your theories and proof
  3473 scripts to the new package (backup your sources first!).
  3474 
  3475 
  3476 * HOL/record package: considerably improved implementation; now
  3477 includes concrete syntax for record types, terms, updates; theorems
  3478 for surjective pairing and splitting !!-bound record variables; proof
  3479 support is as follows:
  3480 
  3481   1) standard conversions (selectors or updates applied to record
  3482 constructor terms) are part of the standard simpset;
  3483 
  3484   2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are
  3485 made part of standard simpset and claset via addIffs;
  3486 
  3487   3) a tactic for record field splitting (record_split_tac) is part of
  3488 the standard claset (addSWrapper);
  3489 
  3490 To get a better idea about these rules you may retrieve them via
  3491 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is
  3492 the name of your record type.
  3493 
  3494 The split tactic 3) conceptually simplifies by the following rule:
  3495 
  3496   "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"
  3497 
  3498 Thus any record variable that is bound by meta-all will automatically
  3499 blow up into some record constructor term, consequently the
  3500 simplifications of 1), 2) apply.  Thus force_tac, auto_tac etc. shall
  3501 solve record problems automatically.
  3502 
  3503 
  3504 * reorganized the main HOL image: HOL/Integ and String loaded by
  3505 default; theory Main includes everything;
  3506 
  3507 * automatic simplification of integer sums and comparisons, using cancellation;
  3508 
  3509 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;
  3510 
  3511 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;
  3512 
  3513 * many new identities for unions, intersections, set difference, etc.;
  3514 
  3515 * expand_if, expand_split, expand_sum_case and expand_nat_case are now
  3516 called split_if, split_split, split_sum_case and split_nat_case (to go
  3517 with add/delsplits);
  3518 
  3519 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting
  3520 (?x::unit) = (); this is made part of the default simpset, which COULD
  3521 MAKE EXISTING PROOFS FAIL under rare circumstances (consider
  3522 'Delsimprocs [unit_eq_proc];' as last resort); also note that
  3523 unit_abs_eta_conv is added in order to counter the effect of
  3524 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by
  3525 %u.f();
  3526 
  3527 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which
  3528 makes more sense);
  3529 
  3530 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  3531   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  3532   disjointness reasoning but breaking a few old proofs.
  3533 
  3534 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1
  3535 to 'converse' from 'inverse' (for compatibility with ZF and some
  3536 literature);
  3537 
  3538 * HOL/recdef can now declare non-recursive functions, with {} supplied as
  3539 the well-founded relation;
  3540 
  3541 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of
  3542     Compl A.  The "Compl" syntax remains available as input syntax for this
  3543     release ONLY.
  3544 
  3545 * HOL/Update: new theory of function updates:
  3546     f(a:=b) == %x. if x=a then b else f x
  3547 may also be iterated as in f(a:=b,c:=d,...);
  3548 
  3549 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;
  3550 
  3551 * HOL/List:
  3552   - new function list_update written xs[i:=v] that updates the i-th
  3553     list position. May also be iterated as in xs[i:=a,j:=b,...].
  3554   - new function `upt' written [i..j(] which generates the list
  3555     [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper
  3556     bound write [i..j], which is a shorthand for [i..j+1(].
  3557   - new lexicographic orderings and corresponding wellfoundedness theorems.
  3558 
  3559 * HOL/Arith:
  3560   - removed 'pred' (predecessor) function;
  3561   - generalized some theorems about n-1;
  3562   - many new laws about "div" and "mod";
  3563   - new laws about greatest common divisors (see theory ex/Primes);
  3564 
  3565 * HOL/Relation: renamed the relational operator r^-1 "converse"
  3566 instead of "inverse";
  3567 
  3568 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness
  3569   of the multiset ordering;
  3570 
  3571 * directory HOL/Real: a construction of the reals using Dedekind cuts
  3572   (not included by default);
  3573 
  3574 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;
  3575 
  3576 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted
  3577   programs, i.e. different program variables may have different types.
  3578 
  3579 * calling (stac rew i) now fails if "rew" has no effect on the goal
  3580   [previously, this check worked only if the rewrite rule was unconditional]
  3581   Now rew can involve either definitions or equalities (either == or =).
  3582 
  3583 
  3584 *** ZF ***
  3585 
  3586 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains
  3587   only the theorems proved on ZF.ML;
  3588 
  3589 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  3590   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  3591   disjointness reasoning but breaking a few old proofs.
  3592 
  3593 * ZF/Update: new theory of function updates
  3594     with default rewrite rule  f(x:=y) ` z = if(z=x, y, f`z)
  3595   may also be iterated as in f(a:=b,c:=d,...);
  3596 
  3597 * in  let x=t in u(x), neither t nor u(x) has to be an FOL term.
  3598 
  3599 * calling (stac rew i) now fails if "rew" has no effect on the goal
  3600   [previously, this check worked only if the rewrite rule was unconditional]
  3601   Now rew can involve either definitions or equalities (either == or =).
  3602 
  3603 * case_tac provided for compatibility with HOL
  3604     (like the old excluded_middle_tac, but with subgoals swapped)
  3605 
  3606 
  3607 *** Internal programming interfaces ***
  3608 
  3609 * Pure: several new basic modules made available for general use, see
  3610 also src/Pure/README;
  3611 
  3612 * improved the theory data mechanism to support encapsulation (data
  3613 kind name replaced by private Object.kind, acting as authorization
  3614 key); new type-safe user interface via functor TheoryDataFun; generic
  3615 print_data function becomes basically useless;
  3616 
  3617 * removed global_names compatibility flag -- all theory declarations
  3618 are qualified by default;
  3619 
  3620 * module Pure/Syntax now offers quote / antiquote translation
  3621 functions (useful for Hoare logic etc. with implicit dependencies);
  3622 see HOL/ex/Antiquote for an example use;
  3623 
  3624 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->
  3625 cterm -> thm;
  3626 
  3627 * new tactical CHANGED_GOAL for checking that a tactic modifies a
  3628 subgoal;
  3629 
  3630 * Display.print_goals function moved to Locale.print_goals;
  3631 
  3632 * standard print function for goals supports current_goals_markers
  3633 variable for marking begin of proof, end of proof, start of goal; the
  3634 default is ("", "", ""); setting current_goals_markers := ("<proof>",
  3635 "</proof>", "<goal>") causes SGML like tagged proof state printing,
  3636 for example;
  3637 
  3638 
  3639 
  3640 New in Isabelle98 (January 1998)
  3641 --------------------------------
  3642 
  3643 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3644 
  3645 * changed lexical syntax of terms / types: dots made part of long
  3646 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";
  3647 
  3648 * simpset (and claset) reference variable replaced by functions
  3649 simpset / simpset_ref;
  3650 
  3651 * no longer supports theory aliases (via merge) and non-trivial
  3652 implicit merge of thms' signatures;
  3653 
  3654 * most internal names of constants changed due to qualified names;
  3655 
  3656 * changed Pure/Sequence interface (see Pure/seq.ML);
  3657 
  3658 
  3659 *** General Changes ***
  3660 
  3661 * hierachically structured name spaces (for consts, types, axms, thms
  3662 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of
  3663 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:
  3664 isatool fixdots ensures space after dots (e.g. "%x. x"); set
  3665 long_names for fully qualified output names; NOTE: ML programs
  3666 (special tactics, packages etc.) referring to internal names may have
  3667 to be adapted to cope with fully qualified names; in case of severe
  3668 backward campatibility problems try setting 'global_names' at compile
  3669 time to have enrything declared within a flat name space; one may also
  3670 fine tune name declarations in theories via the 'global' and 'local'
  3671 section;
  3672 
  3673 * reimplemented the implicit simpset and claset using the new anytype
  3674 data filed in signatures; references simpset:simpset ref etc. are
  3675 replaced by functions simpset:unit->simpset and
  3676 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp
  3677 to patch your ML files accordingly;
  3678 
  3679 * HTML output now includes theory graph data for display with Java
  3680 applet or isatool browser; data generated automatically via isatool
  3681 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);
  3682 
  3683 * defs may now be conditional; improved rewrite_goals_tac to handle
  3684 conditional equations;
  3685 
  3686 * defs now admits additional type arguments, using TYPE('a) syntax;
  3687 
  3688 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always
  3689 creates a new theory node; implicit merge of thms' signatures is
  3690 restricted to 'trivial' ones; COMPATIBILITY: one may have to use
  3691 transfer:theory->thm->thm in (rare) cases;
  3692 
  3693 * improved handling of draft signatures / theories; draft thms (and
  3694 ctyps, cterms) are automatically promoted to real ones;
  3695 
  3696 * slightly changed interfaces for oracles: admit many per theory, named
  3697 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;
  3698 
  3699 * print_goals: optional output of const types (set show_consts and
  3700 show_types);
  3701 
  3702 * improved output of warnings (###) and errors (***);
  3703 
  3704 * subgoal_tac displays a warning if the new subgoal has type variables;
  3705 
  3706 * removed old README and Makefiles;
  3707 
  3708 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;
  3709 
  3710 * removed obsolete init_pps and init_database;
  3711 
  3712 * deleted the obsolete tactical STATE, which was declared by
  3713     fun STATE tacfun st = tacfun st st;
  3714 
  3715 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~
  3716 (which abbreviates $HOME);
  3717 
  3718 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:
  3719 use isatool fixseq to adapt your ML programs (this works for fully
  3720 qualified references to the Sequence structure only!);
  3721 
  3722 * use_thy no longer requires writable current directory; it always
  3723 reloads .ML *and* .thy file, if either one is out of date;
  3724 
  3725 
  3726 *** Classical Reasoner ***
  3727 
  3728 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new
  3729 tactics that use classical reasoning to simplify a subgoal without
  3730 splitting it into several subgoals;
  3731 
  3732 * Safe_tac: like safe_tac but uses the default claset;
  3733 
  3734 
  3735 *** Simplifier ***
  3736 
  3737 * added simplification meta rules:
  3738     (asm_)(full_)simplify: simpset -> thm -> thm;
  3739 
  3740 * simplifier.ML no longer part of Pure -- has to be loaded by object
  3741 logics (again);
  3742 
  3743 * added prems argument to simplification procedures;
  3744 
  3745 * HOL, FOL, ZF: added infix function `addsplits':
  3746   instead of `<simpset> setloop (split_tac <thms>)'
  3747   you can simply write `<simpset> addsplits <thms>'
  3748 
  3749 
  3750 *** Syntax ***
  3751 
  3752 * TYPE('a) syntax for type reflection terms;
  3753 
  3754 * no longer handles consts with name "" -- declare as 'syntax' instead;
  3755 
  3756 * pretty printer: changed order of mixfix annotation preference (again!);
  3757 
  3758 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;
  3759 
  3760 
  3761 *** HOL ***
  3762 
  3763 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.
  3764   with `addloop' of the simplifier to faciliate case splitting in premises.
  3765 
  3766 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;
  3767 
  3768 * HOL/Auth: new protocol proofs including some for the Internet
  3769   protocol TLS;
  3770 
  3771 * HOL/Map: new theory of `maps' a la VDM;
  3772 
  3773 * HOL/simplifier: simplification procedures nat_cancel_sums for
  3774 cancelling out common nat summands from =, <, <= (in)equalities, or
  3775 differences; simplification procedures nat_cancel_factor for
  3776 cancelling common factor from =, <, <= (in)equalities over natural
  3777 sums; nat_cancel contains both kinds of procedures, it is installed by
  3778 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;
  3779 
  3780 * HOL/simplifier: terms of the form
  3781   `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)'  (or t=x)
  3782   are rewritten to
  3783   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',
  3784   and those of the form
  3785   `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)'  (or t=x)
  3786   are rewritten to
  3787   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',
  3788 
  3789 * HOL/datatype
  3790   Each datatype `t' now comes with a theorem `split_t_case' of the form
  3791 
  3792   P(t_case f1 ... fn x) =
  3793      ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &
  3794         ...
  3795        (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))
  3796      )
  3797 
  3798   and a theorem `split_t_case_asm' of the form
  3799 
  3800   P(t_case f1 ... fn x) =
  3801     ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |
  3802         ...
  3803        (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))
  3804      )
  3805   which can be added to a simpset via `addsplits'. The existing theorems
  3806   expand_list_case and expand_option_case have been renamed to
  3807   split_list_case and split_option_case.
  3808 
  3809 * HOL/Arithmetic:
  3810   - `pred n' is automatically converted to `n-1'.
  3811     Users are strongly encouraged not to use `pred' any longer,
  3812     because it will disappear altogether at some point.
  3813   - Users are strongly encouraged to write "0 < n" rather than
  3814     "n ~= 0". Theorems and proof tools have been modified towards this
  3815     `standard'.
  3816 
  3817 * HOL/Lists:
  3818   the function "set_of_list" has been renamed "set" (and its theorems too);
  3819   the function "nth" now takes its arguments in the reverse order and
  3820   has acquired the infix notation "!" as in "xs!n".
  3821 
  3822 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};
  3823 
  3824 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its
  3825   specialist theorems (like UN1_I) are gone.  Similarly for (INT x.B x);
  3826 
  3827 * HOL/record: extensible records with schematic structural subtyping
  3828 (single inheritance); EXPERIMENTAL version demonstrating the encoding,
  3829 still lacks various theorems and concrete record syntax;
  3830 
  3831 
  3832 *** HOLCF ***
  3833 
  3834 * removed "axioms" and "generated by" sections;
  3835 
  3836 * replaced "ops" section by extended "consts" section, which is capable of
  3837   handling the continuous function space "->" directly;
  3838 
  3839 * domain package:
  3840   . proves theorems immediately and stores them in the theory,
  3841   . creates hierachical name space,
  3842   . now uses normal mixfix annotations (instead of cinfix...),
  3843   . minor changes to some names and values (for consistency),
  3844   . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,
  3845   . separator between mutual domain defs: changed "," to "and",
  3846   . improved handling of sort constraints;  now they have to
  3847     appear on the left-hand side of the equations only;
  3848 
  3849 * fixed LAM <x,y,zs>.b syntax;
  3850 
  3851 * added extended adm_tac to simplifier in HOLCF -- can now discharge
  3852 adm (%x. P (t x)), where P is chainfinite and t continuous;
  3853 
  3854 
  3855 *** FOL and ZF ***
  3856 
  3857 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.
  3858   with `addloop' of the simplifier to faciliate case splitting in premises.
  3859 
  3860 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as
  3861 in HOL, they strip ALL and --> from proved theorems;
  3862 
  3863 
  3864 
  3865 New in Isabelle94-8 (May 1997)
  3866 ------------------------------
  3867 
  3868 *** General Changes ***
  3869 
  3870 * new utilities to build / run / maintain Isabelle etc. (in parts
  3871 still somewhat experimental); old Makefiles etc. still functional;
  3872 
  3873 * new 'Isabelle System Manual';
  3874 
  3875 * INSTALL text, together with ./configure and ./build scripts;
  3876 
  3877 * reimplemented type inference for greater efficiency, better error
  3878 messages and clean internal interface;
  3879 
  3880 * prlim command for dealing with lots of subgoals (an easier way of
  3881 setting goals_limit);
  3882 
  3883 
  3884 *** Syntax ***
  3885 
  3886 * supports alternative (named) syntax tables (parser and pretty
  3887 printer); internal interface is provided by add_modesyntax(_i);
  3888 
  3889 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to
  3890 be used in conjunction with the Isabelle symbol font; uses the
  3891 "symbols" syntax table;
  3892 
  3893 * added token_translation interface (may translate name tokens in
  3894 arbitrary ways, dependent on their type (free, bound, tfree, ...) and
  3895 the current print_mode); IMPORTANT: user print translation functions
  3896 are responsible for marking newly introduced bounds
  3897 (Syntax.mark_boundT);
  3898 
  3899 * token translations for modes "xterm" and "xterm_color" that display
  3900 names in bold, underline etc. or colors (which requires a color
  3901 version of xterm);
  3902 
  3903 * infixes may now be declared with names independent of their syntax;
  3904 
  3905 * added typed_print_translation (like print_translation, but may
  3906 access type of constant);
  3907 
  3908 
  3909 *** Classical Reasoner ***
  3910 
  3911 Blast_tac: a new tactic!  It is often more powerful than fast_tac, but has
  3912 some limitations.  Blast_tac...
  3913   + ignores addss, addbefore, addafter; this restriction is intrinsic
  3914   + ignores elimination rules that don't have the correct format
  3915         (the conclusion MUST be a formula variable)
  3916   + ignores types, which can make HOL proofs fail
  3917   + rules must not require higher-order unification, e.g. apply_type in ZF
  3918     [message "Function Var's argument not a bound variable" relates to this]
  3919   + its proof strategy is more general but can actually be slower
  3920 
  3921 * substitution with equality assumptions no longer permutes other
  3922 assumptions;
  3923 
  3924 * minor changes in semantics of addafter (now called addaltern); renamed
  3925 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper
  3926 (and access functions for it);
  3927 
  3928 * improved combination of classical reasoner and simplifier:
  3929   + functions for handling clasimpsets
  3930   + improvement of addss: now the simplifier is called _after_ the
  3931     safe steps.
  3932   + safe variant of addss called addSss: uses safe simplifications
  3933     _during_ the safe steps. It is more complete as it allows multiple
  3934     instantiations of unknowns (e.g. with slow_tac).
  3935 
  3936 *** Simplifier ***
  3937 
  3938 * added interface for simplification procedures (functions that
  3939 produce *proven* rewrite rules on the fly, depending on current
  3940 redex);
  3941 
  3942 * ordering on terms as parameter (used for ordered rewriting);
  3943 
  3944 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;
  3945 
  3946 * the solver is now split into a safe and an unsafe part.
  3947 This should be invisible for the normal user, except that the
  3948 functions setsolver and addsolver have been renamed to setSolver and
  3949 addSolver; added safe_asm_full_simp_tac;
  3950 
  3951 
  3952 *** HOL ***
  3953 
  3954 * a generic induction tactic `induct_tac' which works for all datatypes and
  3955 also for type `nat';
  3956 
  3957 * a generic case distinction tactic `exhaust_tac' which works for all
  3958 datatypes and also for type `nat';
  3959 
  3960 * each datatype comes with a function `size';
  3961 
  3962 * patterns in case expressions allow tuple patterns as arguments to
  3963 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';
  3964 
  3965 * primrec now also works with type nat;
  3966 
  3967 * recdef: a new declaration form, allows general recursive functions to be
  3968 defined in theory files.  See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.
  3969 
  3970 * the constant for negation has been renamed from "not" to "Not" to
  3971 harmonize with FOL, ZF, LK, etc.;
  3972 
  3973 * HOL/ex/LFilter theory of a corecursive "filter" functional for
  3974 infinite lists;
  3975 
  3976 * HOL/Modelcheck demonstrates invocation of model checker oracle;
  3977 
  3978 * HOL/ex/Ring.thy declares cring_simp, which solves equational
  3979 problems in commutative rings, using axiomatic type classes for + and *;
  3980 
  3981 * more examples in HOL/MiniML and HOL/Auth;
  3982 
  3983 * more default rewrite rules for quantifiers, union/intersection;
  3984 
  3985 * a new constant `arbitrary == @x.False';
  3986 
  3987 * HOLCF/IOA replaces old HOL/IOA;
  3988 
  3989 * HOLCF changes: derived all rules and arities
  3990   + axiomatic type classes instead of classes
  3991   + typedef instead of faking type definitions
  3992   + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.
  3993   + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po
  3994   + eliminated the types void, one, tr
  3995   + use unit lift and bool lift (with translations) instead of one and tr
  3996   + eliminated blift from Lift3.thy (use Def instead of blift)
  3997   all eliminated rules are derived as theorems --> no visible changes ;
  3998 
  3999 
  4000 *** ZF ***
  4001 
  4002 * ZF now has Fast_tac, Simp_tac and Auto_tac.  Union_iff is a now a default
  4003 rewrite rule; this may affect some proofs.  eq_cs is gone but can be put back
  4004 as ZF_cs addSIs [equalityI];
  4005 
  4006 
  4007 
  4008 New in Isabelle94-7 (November 96)
  4009 ---------------------------------
  4010 
  4011 * allowing negative levels (as offsets) in prlev and choplev;
  4012 
  4013 * super-linear speedup for large simplifications;
  4014 
  4015 * FOL, ZF and HOL now use miniscoping: rewriting pushes
  4016 quantifications in as far as possible (COULD MAKE EXISTING PROOFS
  4017 FAIL); can suppress it using the command Delsimps (ex_simps @
  4018 all_simps); De Morgan laws are also now included, by default;
  4019 
  4020 * improved printing of ==>  :  ~:
  4021 
  4022 * new object-logic "Sequents" adds linear logic, while replacing LK
  4023 and Modal (thanks to Sara Kalvala);
  4024 
  4025 * HOL/Auth: correctness proofs for authentication protocols;
  4026 
  4027 * HOL: new auto_tac combines rewriting and classical reasoning (many
  4028 examples on HOL/Auth);
  4029 
  4030 * HOL: new command AddIffs for declaring theorems of the form P=Q to
  4031 the rewriter and classical reasoner simultaneously;
  4032 
  4033 * function uresult no longer returns theorems in "standard" format;
  4034 regain previous version by: val uresult = standard o uresult;
  4035 
  4036 
  4037 
  4038 New in Isabelle94-6
  4039 -------------------
  4040 
  4041 * oracles -- these establish an interface between Isabelle and trusted
  4042 external reasoners, which may deliver results as theorems;
  4043 
  4044 * proof objects (in particular record all uses of oracles);
  4045 
  4046 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;
  4047 
  4048 * "constdefs" section in theory files;
  4049 
  4050 * "primrec" section (HOL) no longer requires names;
  4051 
  4052 * internal type "tactic" now simply "thm -> thm Sequence.seq";
  4053 
  4054 
  4055 
  4056 New in Isabelle94-5
  4057 -------------------
  4058 
  4059 * reduced space requirements;
  4060 
  4061 * automatic HTML generation from theories;
  4062 
  4063 * theory files no longer require "..." (quotes) around most types;
  4064 
  4065 * new examples, including two proofs of the Church-Rosser theorem;
  4066 
  4067 * non-curried (1994) version of HOL is no longer distributed;
  4068 
  4069 
  4070 
  4071 New in Isabelle94-4
  4072 -------------------
  4073 
  4074 * greatly reduced space requirements;
  4075 
  4076 * theory files (.thy) no longer require \...\ escapes at line breaks;
  4077 
  4078 * searchable theorem database (see the section "Retrieving theorems" on
  4079 page 8 of the Reference Manual);
  4080 
  4081 * new examples, including Grabczewski's monumental case study of the
  4082 Axiom of Choice;
  4083 
  4084 * The previous version of HOL renamed to Old_HOL;
  4085 
  4086 * The new version of HOL (previously called CHOL) uses a curried syntax
  4087 for functions.  Application looks like f a b instead of f(a,b);
  4088 
  4089 * Mutually recursive inductive definitions finally work in HOL;
  4090 
  4091 * In ZF, pattern-matching on tuples is now available in all abstractions and
  4092 translates to the operator "split";
  4093 
  4094 
  4095 
  4096 New in Isabelle94-3
  4097 -------------------
  4098 
  4099 * new infix operator, addss, allowing the classical reasoner to
  4100 perform simplification at each step of its search.  Example:
  4101         fast_tac (cs addss ss)
  4102 
  4103 * a new logic, CHOL, the same as HOL, but with a curried syntax
  4104 for functions.  Application looks like f a b instead of f(a,b).  Also pairs
  4105 look like (a,b) instead of <a,b>;
  4106 
  4107 * PLEASE NOTE: CHOL will eventually replace HOL!
  4108 
  4109 * In CHOL, pattern-matching on tuples is now available in all abstractions.
  4110 It translates to the operator "split".  A new theory of integers is available;
  4111 
  4112 * In ZF, integer numerals now denote two's-complement binary integers.
  4113 Arithmetic operations can be performed by rewriting.  See ZF/ex/Bin.ML;
  4114 
  4115 * Many new examples: I/O automata, Church-Rosser theorem, equivalents
  4116 of the Axiom of Choice;
  4117 
  4118 
  4119 
  4120 New in Isabelle94-2
  4121 -------------------
  4122 
  4123 * Significantly faster resolution;
  4124 
  4125 * the different sections in a .thy file can now be mixed and repeated
  4126 freely;
  4127 
  4128 * Database of theorems for FOL, HOL and ZF.  New
  4129 commands including qed, qed_goal and bind_thm store theorems in the database.
  4130 
  4131 * Simple database queries: return a named theorem (get_thm) or all theorems of
  4132 a given theory (thms_of), or find out what theory a theorem was proved in
  4133 (theory_of_thm);
  4134 
  4135 * Bugs fixed in the inductive definition and datatype packages;
  4136 
  4137 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs
  4138 and HOL_dup_cs obsolete;
  4139 
  4140 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1
  4141 have been removed;
  4142 
  4143 * Simpler definition of function space in ZF;
  4144 
  4145 * new results about cardinal and ordinal arithmetic in ZF;
  4146 
  4147 * 'subtype' facility in HOL for introducing new types as subsets of existing
  4148 types;
  4149 
  4150 
  4151 $Id$