NEWS
author wenzelm
Sat Jan 26 19:20:01 2002 +0100 (2002-01-26)
changeset 12853 de505273c971
parent 12850 d3c16021e999
child 12877 b9635eb8a448
permissions -rw-r--r--
Isar cases/induct: no backtracking;
     1 
     2 Isabelle NEWS -- history user-relevant changes
     3 ==============================================
     4 
     5 New in Isabelle2002 (January 2002)
     6 ----------------------------------
     7 
     8 *** Document preparation ***
     9 
    10 * greatly simplified document preparation setup, including more
    11 graceful interpretation of isatool usedir -i/-d/-D options, and more
    12 instructive isatool mkdir; users should basically be able to get
    13 started with "isatool mkdir Test && isatool make";
    14 
    15 * theory dependency graph may now be incorporated into documents;
    16 isatool usedir -g true will produce session_graph.eps/.pdf for use
    17 with \includegraphics of LaTeX;
    18 
    19 * proper spacing of consecutive markup elements, especially text
    20 blocks after section headings;
    21 
    22 * support bold style (for single symbols only), input syntax is like
    23 this: "\<^bold>\<alpha>" or "\<^bold>A";
    24 
    25 * \<bullet> is now output as bold \cdot by default, which looks much
    26 better in printed text;
    27 
    28 * added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
    29 note that these symbols are currently unavailable in Proof General /
    30 X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
    31 
    32 * isatool latex no longer depends on changed TEXINPUTS, instead
    33 isatool document copies the Isabelle style files to the target
    34 location;
    35 
    36 
    37 *** Isar ***
    38 
    39 * Pure/Provers: improved proof by cases and induction;
    40   - 'case' command admits impromptu naming of parameters (such as
    41     "case (Suc n)");
    42   - 'induct' method divinates rule instantiation from the inductive
    43     claim; no longer requires excessive ?P bindings for proper
    44     instantiation of cases;
    45   - 'induct' method properly enumerates all possibilities of set/type
    46     rules; as a consequence facts may be also passed through *type*
    47     rules without further ado;
    48   - 'induct' method now derives symbolic cases from the *rulified*
    49     rule (before it used to rulify cases stemming from the internal
    50     atomized version); this means that the context of a non-atomic
    51     statement becomes is included in the hypothesis, avoiding the
    52     slightly cumbersome show "PROP ?case" form;
    53   - 'induct' may now use elim-style induction rules without chaining
    54     facts, using ``missing'' premises from the goal state; this allows
    55     rules stemming from inductive sets to be applied in unstructured
    56     scripts, while still benefitting from proper handling of non-atomic
    57     statements; NB: major inductive premises need to be put first, all
    58     the rest of the goal is passed through the induction;
    59   - 'induct' proper support for mutual induction involving non-atomic
    60     rule statements (uses the new concept of simultaneous goals, see
    61     below);
    62   - append all possible rule selections, but only use the first
    63     success (no backtracking);
    64   - removed obsolete "(simplified)" and "(stripped)" options of methods;
    65   - undeclared rule case names default to numbers 1, 2, 3, ...;
    66   - added 'print_induct_rules' (covered by help item in recent Proof
    67     General versions);
    68   - moved induct/cases attributes to Pure, methods to Provers;
    69   - generic method setup instantiated for FOL and HOL;
    70 
    71 * Pure: support multiple simultaneous goal statements, for example
    72 "have a: A and b: B" (same for 'theorem' etc.); being a pure
    73 meta-level mechanism, this acts as if several individual goals had
    74 been stated separately; in particular common proof methods need to be
    75 repeated in order to cover all claims; note that a single elimination
    76 step is *not* sufficient to establish the two conjunctions, so this
    77 fails:
    78 
    79   assume "A & B" then have A and B ..   (*".." fails*)
    80 
    81 better use "obtain" in situations as above; alternative refer to
    82 multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
    83 
    84 * Pure: proper integration with ``locales''; unlike the original
    85 version by Florian Kammüller, Isar locales package high-level proof
    86 contexts rather than raw logical ones (e.g. we admit to include
    87 attributes everywhere); operations on locales include merge and
    88 rename; e.g. see HOL/ex/Locales.thy;
    89 
    90 * Pure: the following commands have been ``localized'', supporting a
    91 target locale specification "(in name)": 'lemma', 'theorem',
    92 'corollary', 'lemmas', 'theorems', 'declare'; the results will be
    93 stored both within the locale and at the theory level (exported and
    94 qualified by the locale name);
    95 
    96 * Pure: theory goals now support ad-hoc contexts, which are discharged
    97 in the result, as in ``lemma (assumes A and B) K: A .''; syntax
    98 coincides with that of a locale body;
    99 
   100 * Pure: renamed "antecedent" case to "rule_context";
   101 
   102 * Pure: added 'corollary' command;
   103 
   104 * Pure: fixed 'token_translation' command;
   105 
   106 * Pure: removed obsolete 'exported' attribute;
   107 
   108 * Pure: dummy pattern "_" in is/let is now automatically lifted over
   109 bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
   110 supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
   111 
   112 * Pure: method 'atomize' presents local goal premises as object-level
   113 statements (atomic meta-level propositions); setup controlled via
   114 rewrite rules declarations of 'atomize' attribute; example
   115 application: 'induct' method with proper rule statements in improper
   116 proof *scripts*;
   117 
   118 * Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
   119 now consider the syntactic context of assumptions, giving a better
   120 chance to get type-inference of the arguments right (this is
   121 especially important for locales);
   122 
   123 * Pure: "sorry" no longer requires quick_and_dirty in interactive
   124 mode;
   125 
   126 * Pure/obtain: the formal conclusion "thesis", being marked as
   127 ``internal'', may no longer be reference directly in the text;
   128 potential INCOMPATIBILITY, may need to use "?thesis" in rare
   129 situations;
   130 
   131 * Pure: generic 'sym' attribute which declares a rule both as pure
   132 'elim?' and for the 'symmetric' operation;
   133 
   134 * Pure/Provers/classical: simplified integration with pure rule
   135 attributes and methods; the classical "intro?/elim?/dest?"
   136 declarations coincide with the pure ones; the "rule" method no longer
   137 includes classically swapped intros; "intro" and "elim" methods no
   138 longer pick rules from the context; also got rid of ML declarations
   139 AddXIs/AddXEs/AddXDs; all of this has some potential for
   140 INCOMPATIBILITY;
   141 
   142 * Provers/classical: attribute 'swapped' produces classical inversions
   143 of introduction rules;
   144 
   145 * Provers/simplifier: 'simplified' attribute may refer to explicit
   146 rules instead of full simplifier context; 'iff' attribute handles
   147 conditional rules;
   148 
   149 * HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
   150 
   151 * HOL: 'recdef' now fails on unfinished automated proofs, use
   152 "(permissive)" option to recover old behavior;
   153 
   154 * HOL: 'inductive' no longer features separate (collective) attributes
   155 for 'intros' (was found too confusing);
   156 
   157 * HOL: properly declared induction rules less_induct and
   158 wf_induct_rule;
   159 
   160 
   161 *** HOL ***
   162 
   163 * HOL: moved over to sane numeral syntax; the new policy is as
   164 follows:
   165 
   166   - 0 and 1 are polymorphic constants, which are defined on any
   167   numeric type (nat, int, real etc.);
   168 
   169   - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
   170   binary representation internally;
   171 
   172   - type nat has special constructor Suc, and generally prefers Suc 0
   173   over 1::nat and Suc (Suc 0) over 2::nat;
   174 
   175 This change may cause significant problems of INCOMPATIBILITY; here
   176 are some hints on converting existing sources:
   177 
   178   - due to the new "num" token, "-0" and "-1" etc. are now atomic
   179   entities, so expressions involving "-" (unary or binary minus) need
   180   to be spaced properly;
   181 
   182   - existing occurrences of "1" may need to be constraint "1::nat" or
   183   even replaced by Suc 0; similar for old "2";
   184 
   185   - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
   186 
   187   - remove all special provisions on numerals in proofs;
   188 
   189 * HOL: simp rules nat_number_of expand numerals on nat to Suc/0
   190 representation (depends on bin_arith_simps in the default context);
   191 
   192 * HOL: symbolic syntax for x^2 (numeral 2);
   193 
   194 * HOL: the class of all HOL types is now called "type" rather than
   195 "term"; INCOMPATIBILITY, need to adapt references to this type class
   196 in axclass/classes, instance/arities, and (usually rare) occurrences
   197 in typings (of consts etc.); internally the class is called
   198 "HOL.type", ML programs should refer to HOLogic.typeS;
   199 
   200 * HOL/record package improvements:
   201   - new derived operations "fields" to build a partial record section,
   202     "extend" to promote a fixed record to a record scheme, and
   203     "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
   204     declared as simp by default;
   205   - shared operations ("more", "fields", etc.) now need to be always
   206     qualified) --- potential INCOMPATIBILITY;
   207   - removed "make_scheme" operations (use "make" with "extend") --
   208     INCOMPATIBILITY;
   209   - removed "more" class (simply use "term") -- INCOMPATIBILITY;
   210   - provides cases/induct rules for use with corresponding Isar
   211     methods (for concrete records, record schemes, concrete more
   212     parts, and schematic more parts -- in that order);
   213   - internal definitions directly based on a light-weight abstract
   214     theory of product types over typedef rather than datatype;
   215 
   216 * HOL: canonical cases/induct rules for n-tuples (n = 3..7);
   217 
   218 * HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
   219 (beware of argument permutation!);
   220 
   221 * HOL: linorder_less_split superseded by linorder_cases;
   222 
   223 * HOL: added "The" definite description operator; move Hilbert's "Eps"
   224 to peripheral theory "Hilbert_Choice";
   225 
   226 * HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
   227 in this (rare) case use:
   228 
   229   delSWrapper "split_all_tac"
   230   addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
   231 
   232 * HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
   233 MAY FAIL;
   234 
   235 * HOL: introduced f^n = f o ... o f; warning: due to the limits of
   236 Isabelle's type classes, ^ on functions and relations has too general
   237 a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
   238 necessary to attach explicit type constraints;
   239 
   240 * HOL/Relation: the prefix name of the infix "O" has been changed from "comp"
   241 to "rel_comp"; INCOMPATIBILITY: a few theorems have been renamed accordingly
   242 (eg "compI" -> "rel_compI").
   243 
   244 * HOL: syntax translations now work properly with numerals and records
   245 expressions;
   246 
   247 * HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
   248 of "lam" -- INCOMPATIBILITY;
   249 
   250 * HOL: got rid of some global declarations (potential INCOMPATIBILITY
   251 for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
   252 renamed "Product_Type.unit";
   253 
   254 * HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
   255 
   256 * HOL/GroupTheory: group theory examples including Sylow's theorem (by
   257 Florian Kammüller);
   258 
   259 * HOL/IMP: updated and converted to new-style theory format; several
   260 parts turned into readable document, with proper Isar proof texts and
   261 some explanations (by Gerwin Klein);
   262 
   263 * HOL-Real: added Complex_Numbers (by Gertrud Bauer);
   264 
   265 * HOL-Hyperreal is now a logic image;
   266 
   267 
   268 *** HOLCF ***
   269 
   270 * Isar: consts/constdefs supports mixfix syntax for continuous
   271 operations;
   272 
   273 * Isar: domain package adapted to new-style theory format, e.g. see
   274 HOLCF/ex/Dnat.thy;
   275 
   276 * theory Lift: proper use of rep_datatype lift instead of ML hacks --
   277 potential INCOMPATIBILITY; now use plain induct_tac instead of former
   278 lift.induct_tac, always use UU instead of Undef;
   279 
   280 * HOLCF/IMP: updated and converted to new-style theory;
   281 
   282 
   283 *** ZF ***
   284 
   285 * Isar: proper integration of logic-specific tools and packages,
   286 including theory commands '(co)inductive', '(co)datatype',
   287 'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
   288 'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
   289 
   290 * theory Main no longer includes AC; for the Axiom of Choice, base
   291 your theory on Main_ZFC;
   292 
   293 * the integer library now covers quotients and remainders, with many
   294 laws relating division to addition, multiplication, etc.;
   295 
   296 * ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
   297 typeless version of the formalism;
   298 
   299 * ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory format;
   300 
   301 * ZF/Induct: new directory for examples of inductive definitions,
   302 including theory Multiset for multiset orderings; converted to
   303 new-style theory format;
   304 
   305 * Many new theorems about lists, ordinals, etc.;
   306 
   307 
   308 *** General ***
   309 
   310 * Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
   311 variable proof controls level of detail: 0 = no proofs (only oracle
   312 dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
   313 also ref manual for further ML interfaces;
   314 
   315 * Pure/axclass: removed obsolete ML interface
   316 goal_subclass/goal_arity;
   317 
   318 * Pure/syntax: new token syntax "num" for plain numerals (without "#"
   319 of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
   320 separate tokens, so expressions involving minus need to be spaced
   321 properly;
   322 
   323 * Pure/syntax: support non-oriented infixes, using keyword "infix"
   324 rather than "infixl" or "infixr";
   325 
   326 * Pure/syntax: concrete syntax for dummy type variables admits genuine
   327 sort constraint specifications in type inference; e.g. "x::_::foo"
   328 ensures that the type of "x" is of sort "foo" (but not necessarily a
   329 type variable);
   330 
   331 * Pure/syntax: print modes "type_brackets" and "no_type_brackets"
   332 control output of nested => (types); the default behavior is
   333 "type_brackets";
   334 
   335 * Pure/syntax: builtin parse translation for "_constify" turns valued
   336 tokens into AST constants;
   337 
   338 * Pure/syntax: prefer later declarations of translations and print
   339 translation functions; potential INCOMPATIBILITY: need to reverse
   340 multiple declarations for same syntax element constant;
   341 
   342 * Pure/show_hyps reset by default (in accordance to existing Isar
   343 practice);
   344 
   345 * Provers/classical: renamed addaltern to addafter, addSaltern to
   346 addSafter;
   347 
   348 * Provers/clasimp: ``iff'' declarations now handle conditional rules
   349 as well;
   350 
   351 * system: tested support for MacOS X; should be able to get Isabelle +
   352 Proof General to work in a plain Terminal after installing Poly/ML
   353 (e.g. from the Isabelle distribution area) and GNU bash alone
   354 (e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
   355 support requires further installations, e.g. from
   356 http://fink.sourceforge.net/);
   357 
   358 * system: support Poly/ML 4.1.1 (able to manage larger heaps);
   359 
   360 * system: reduced base memory usage by Poly/ML (approx. 20 MB instead
   361 of 40 MB), cf. ML_OPTIONS;
   362 
   363 * system: Proof General keywords specification is now part of the
   364 Isabelle distribution (see etc/isar-keywords.el);
   365 
   366 * system: support for persistent Proof General sessions (refrain from
   367 outdating all loaded theories on startup); user may create writable
   368 logic images like this: ``isabelle -q HOL Test'';
   369 
   370 * system: smart selection of Isabelle process versus Isabelle
   371 interface, accommodates case-insensitive file systems (e.g. HFS+); may
   372 run both "isabelle" and "Isabelle" even if file names are badly
   373 damaged (executable inspects the case of the first letter of its own
   374 name); added separate "isabelle-process" and "isabelle-interface";
   375 
   376 * system: refrain from any attempt at filtering input streams; no
   377 longer support ``8bit'' encoding of old isabelle font, instead proper
   378 iso-latin characters may now be used; the related isatools
   379 "symbolinput" and "nonascii" have disappeared as well;
   380 
   381 * system: removed old "xterm" interface (the print modes "xterm" and
   382 "xterm_color" are still available for direct use in a suitable
   383 terminal);
   384 
   385 
   386 
   387 New in Isabelle99-2 (February 2001)
   388 -----------------------------------
   389 
   390 *** Overview of INCOMPATIBILITIES ***
   391 
   392 * HOL: please note that theories in the Library and elsewhere often use the
   393 new-style (Isar) format; to refer to their theorems in an ML script you must
   394 bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
   395 
   396 * HOL: inductive package no longer splits induction rule aggressively,
   397 but only as far as specified by the introductions given; the old
   398 format may be recovered via ML function complete_split_rule or attribute
   399 'split_rule (complete)';
   400 
   401 * HOL: induct renamed to lfp_induct, lfp_Tarski to lfp_unfold,
   402 gfp_Tarski to gfp_unfold;
   403 
   404 * HOL: contrapos, contrapos2 renamed to contrapos_nn, contrapos_pp;
   405 
   406 * HOL: infix "dvd" now has priority 50 rather than 70 (because it is a
   407 relation); infix "^^" has been renamed "``"; infix "``" has been
   408 renamed "`"; "univalent" has been renamed "single_valued";
   409 
   410 * HOL/Real: "rinv" and "hrinv" replaced by overloaded "inverse"
   411 operation;
   412 
   413 * HOLCF: infix "`" has been renamed "$"; the symbol syntax is \<cdot>;
   414 
   415 * Isar: 'obtain' no longer declares "that" fact as simp/intro;
   416 
   417 * Isar/HOL: method 'induct' now handles non-atomic goals; as a
   418 consequence, it is no longer monotonic wrt. the local goal context
   419 (which is now passed through the inductive cases);
   420 
   421 * Document preparation: renamed standard symbols \<ll> to \<lless> and
   422 \<gg> to \<ggreater>;
   423 
   424 
   425 *** Document preparation ***
   426 
   427 * \isabellestyle{NAME} selects version of Isabelle output (currently
   428 available: are "it" for near math-mode best-style output, "sl" for
   429 slanted text style, and "tt" for plain type-writer; if no
   430 \isabellestyle command is given, output is according to slanted
   431 type-writer);
   432 
   433 * support sub/super scripts (for single symbols only), input syntax is
   434 like this: "A\<^sup>*" or "A\<^sup>\<star>";
   435 
   436 * some more standard symbols; see Appendix A of the system manual for
   437 the complete list of symbols defined in isabellesym.sty;
   438 
   439 * improved isabelle style files; more abstract symbol implementation
   440 (should now use \isamath{...} and \isatext{...} in custom symbol
   441 definitions);
   442 
   443 * antiquotation @{goals} and @{subgoals} for output of *dynamic* goals
   444 state; Note that presentation of goal states does not conform to
   445 actual human-readable proof documents.  Please do not include goal
   446 states into document output unless you really know what you are doing!
   447 
   448 * proper indentation of antiquoted output with proportional LaTeX
   449 fonts;
   450 
   451 * no_document ML operator temporarily disables LaTeX document
   452 generation;
   453 
   454 * isatool unsymbolize tunes sources for plain ASCII communication;
   455 
   456 
   457 *** Isar ***
   458 
   459 * Pure: Isar now suffers initial goal statements to contain unbound
   460 schematic variables (this does not conform to actual readable proof
   461 documents, due to unpredictable outcome and non-compositional proof
   462 checking); users who know what they are doing may use schematic goals
   463 for Prolog-style synthesis of proven results;
   464 
   465 * Pure: assumption method (an implicit finishing) now handles actual
   466 rules as well;
   467 
   468 * Pure: improved 'obtain' --- moved to Pure, insert "that" into
   469 initial goal, declare "that" only as Pure intro (only for single
   470 steps); the "that" rule assumption may now be involved in implicit
   471 finishing, thus ".." becomes a feasible for trivial obtains;
   472 
   473 * Pure: default proof step now includes 'intro_classes'; thus trivial
   474 instance proofs may be performed by "..";
   475 
   476 * Pure: ?thesis / ?this / "..." now work for pure meta-level
   477 statements as well;
   478 
   479 * Pure: more robust selection of calculational rules;
   480 
   481 * Pure: the builtin notion of 'finished' goal now includes the ==-refl
   482 rule (as well as the assumption rule);
   483 
   484 * Pure: 'thm_deps' command visualizes dependencies of theorems and
   485 lemmas, using the graph browser tool;
   486 
   487 * Pure: predict failure of "show" in interactive mode;
   488 
   489 * Pure: 'thms_containing' now takes actual terms as arguments;
   490 
   491 * HOL: improved method 'induct' --- now handles non-atomic goals
   492 (potential INCOMPATIBILITY); tuned error handling;
   493 
   494 * HOL: cases and induct rules now provide explicit hints about the
   495 number of facts to be consumed (0 for "type" and 1 for "set" rules);
   496 any remaining facts are inserted into the goal verbatim;
   497 
   498 * HOL: local contexts (aka cases) may now contain term bindings as
   499 well; the 'cases' and 'induct' methods new provide a ?case binding for
   500 the result to be shown in each case;
   501 
   502 * HOL: added 'recdef_tc' command;
   503 
   504 * isatool convert assists in eliminating legacy ML scripts;
   505 
   506 
   507 *** HOL ***
   508 
   509 * HOL/Library: a collection of generic theories to be used together
   510 with main HOL; the theory loader path already includes this directory
   511 by default; the following existing theories have been moved here:
   512 HOL/Induct/Multiset, HOL/Induct/Acc (as Accessible_Part), HOL/While
   513 (as While_Combinator), HOL/Lex/Prefix (as List_Prefix);
   514 
   515 * HOL/Unix: "Some aspects of Unix file-system security", a typical
   516 modelling and verification task performed in Isabelle/HOL +
   517 Isabelle/Isar + Isabelle document preparation (by Markus Wenzel).
   518 
   519 * HOL/Algebra: special summation operator SUM no longer exists, it has
   520 been replaced by setsum; infix 'assoc' now has priority 50 (like
   521 'dvd'); axiom 'one_not_zero' has been moved from axclass 'ring' to
   522 'domain', this makes the theory consistent with mathematical
   523 literature;
   524 
   525 * HOL basics: added overloaded operations "inverse" and "divide"
   526 (infix "/"), syntax for generic "abs" operation, generic summation
   527 operator \<Sum>;
   528 
   529 * HOL/typedef: simplified package, provide more useful rules (see also
   530 HOL/subset.thy);
   531 
   532 * HOL/datatype: induction rule for arbitrarily branching datatypes is
   533 now expressed as a proper nested rule (old-style tactic scripts may
   534 require atomize_strip_tac to cope with non-atomic premises);
   535 
   536 * HOL: renamed theory "Prod" to "Product_Type", renamed "split" rule
   537 to "split_conv" (old name still available for compatibility);
   538 
   539 * HOL: improved concrete syntax for strings (e.g. allows translation
   540 rules with string literals);
   541 
   542 * HOL-Real-Hyperreal: this extends HOL-Real with the hyperreals
   543  and Fleuriot's mechanization of analysis, including the transcendental
   544  functions for the reals;
   545 
   546 * HOL/Real, HOL/Hyperreal: improved arithmetic simplification;
   547 
   548 
   549 *** CTT ***
   550 
   551 * CTT: x-symbol support for Pi, Sigma, -->, : (membership); note that
   552 "lam" is displayed as TWO lambda-symbols
   553 
   554 * CTT: theory Main now available, containing everything (that is, Bool
   555 and Arith);
   556 
   557 
   558 *** General ***
   559 
   560 * Pure: the Simplifier has been implemented properly as a derived rule
   561 outside of the actual kernel (at last!); the overall performance
   562 penalty in practical applications is about 50%, while reliability of
   563 the Isabelle inference kernel has been greatly improved;
   564 
   565 * print modes "brackets" and "no_brackets" control output of nested =>
   566 (types) and ==> (props); the default behaviour is "brackets";
   567 
   568 * Provers: fast_tac (and friends) now handle actual object-logic rules
   569 as assumptions as well;
   570 
   571 * system: support Poly/ML 4.0;
   572 
   573 * system: isatool install handles KDE version 1 or 2;
   574 
   575 
   576 
   577 New in Isabelle99-1 (October 2000)
   578 ----------------------------------
   579 
   580 *** Overview of INCOMPATIBILITIES ***
   581 
   582 * HOL: simplification of natural numbers is much changed; to partly
   583 recover the old behaviour (e.g. to prevent n+n rewriting to #2*n)
   584 issue the following ML commands:
   585 
   586   Delsimprocs Nat_Numeral_Simprocs.cancel_numerals;
   587   Delsimprocs [Nat_Numeral_Simprocs.combine_numerals];
   588 
   589 * HOL: simplification no longer dives into case-expressions; this is
   590 controlled by "t.weak_case_cong" for each datatype t;
   591 
   592 * HOL: nat_less_induct renamed to less_induct;
   593 
   594 * HOL: systematic renaming of the SOME (Eps) rules, may use isatool
   595 fixsome to patch .thy and .ML sources automatically;
   596 
   597   select_equality  -> some_equality
   598   select_eq_Ex     -> some_eq_ex
   599   selectI2EX       -> someI2_ex
   600   selectI2         -> someI2
   601   selectI          -> someI
   602   select1_equality -> some1_equality
   603   Eps_sym_eq       -> some_sym_eq_trivial
   604   Eps_eq           -> some_eq_trivial
   605 
   606 * HOL: exhaust_tac on datatypes superceded by new generic case_tac;
   607 
   608 * HOL: removed obsolete theorem binding expand_if (refer to split_if
   609 instead);
   610 
   611 * HOL: the recursion equations generated by 'recdef' are now called
   612 f.simps instead of f.rules;
   613 
   614 * HOL: qed_spec_mp now also handles bounded ALL as well;
   615 
   616 * HOL: 0 is now overloaded, so the type constraint ":: nat" may
   617 sometimes be needed;
   618 
   619 * HOL: the constant for "f``x" is now "image" rather than "op ``";
   620 
   621 * HOL: the constant for "f-``x" is now "vimage" rather than "op -``";
   622 
   623 * HOL: the disjoint sum is now "<+>" instead of "Plus"; the cartesian
   624 product is now "<*>" instead of "Times"; the lexicographic product is
   625 now "<*lex*>" instead of "**";
   626 
   627 * HOL: theory Sexp is now in HOL/Induct examples (it used to be part
   628 of main HOL, but was unused); better use HOL's datatype package;
   629 
   630 * HOL: removed "symbols" syntax for constant "override" of theory Map;
   631 the old syntax may be recovered as follows:
   632 
   633   syntax (symbols)
   634     override  :: "('a ~=> 'b) => ('a ~=> 'b) => ('a ~=> 'b)"
   635       (infixl "\\<oplus>" 100)
   636 
   637 * HOL/Real: "rabs" replaced by overloaded "abs" function;
   638 
   639 * HOL/ML: even fewer consts are declared as global (see theories Ord,
   640 Lfp, Gfp, WF); this only affects ML packages that refer to const names
   641 internally;
   642 
   643 * HOL and ZF: syntax for quotienting wrt an equivalence relation
   644 changed from A/r to A//r;
   645 
   646 * ZF: new treatment of arithmetic (nat & int) may break some old
   647 proofs;
   648 
   649 * Isar: renamed some attributes (RS -> THEN, simplify -> simplified,
   650 rulify -> rule_format, elimify -> elim_format, ...);
   651 
   652 * Isar/Provers: intro/elim/dest attributes changed; renamed
   653 intro/intro!/intro!! flags to intro!/intro/intro? (in most cases, one
   654 should have to change intro!! to intro? only); replaced "delrule" by
   655 "rule del";
   656 
   657 * Isar/HOL: renamed "intrs" to "intros" in inductive definitions;
   658 
   659 * Provers: strengthened force_tac by using new first_best_tac;
   660 
   661 * LaTeX document preparation: several changes of isabelle.sty (see
   662 lib/texinputs);
   663 
   664 
   665 *** Document preparation ***
   666 
   667 * formal comments (text blocks etc.) in new-style theories may now
   668 contain antiquotations of thm/prop/term/typ/text to be presented
   669 according to latex print mode; concrete syntax is like this:
   670 @{term[show_types] "f(x) = a + x"};
   671 
   672 * isatool mkdir provides easy setup of Isabelle session directories,
   673 including proper document sources;
   674 
   675 * generated LaTeX sources are now deleted after successful run
   676 (isatool document -c); may retain a copy somewhere else via -D option
   677 of isatool usedir;
   678 
   679 * isatool usedir -D now lets isatool latex -o sty update the Isabelle
   680 style files, achieving self-contained LaTeX sources and simplifying
   681 LaTeX debugging;
   682 
   683 * old-style theories now produce (crude) LaTeX output as well;
   684 
   685 * browser info session directories are now self-contained (may be put
   686 on WWW server seperately); improved graphs of nested sessions; removed
   687 graph for 'all sessions';
   688 
   689 * several improvements in isabelle style files; \isabellestyle{it}
   690 produces fake math mode output; \isamarkupheader is now \section by
   691 default; see lib/texinputs/isabelle.sty etc.;
   692 
   693 
   694 *** Isar ***
   695 
   696 * Isar/Pure: local results and corresponding term bindings are now
   697 subject to Hindley-Milner polymorphism (similar to ML); this
   698 accommodates incremental type-inference very nicely;
   699 
   700 * Isar/Pure: new derived language element 'obtain' supports
   701 generalized existence reasoning;
   702 
   703 * Isar/Pure: new calculational elements 'moreover' and 'ultimately'
   704 support accumulation of results, without applying any rules yet;
   705 useful to collect intermediate results without explicit name
   706 references, and for use with transitivity rules with more than 2
   707 premises;
   708 
   709 * Isar/Pure: scalable support for case-analysis type proofs: new
   710 'case' language element refers to local contexts symbolically, as
   711 produced by certain proof methods; internally, case names are attached
   712 to theorems as "tags";
   713 
   714 * Isar/Pure: theory command 'hide' removes declarations from
   715 class/type/const name spaces;
   716 
   717 * Isar/Pure: theory command 'defs' supports option "(overloaded)" to
   718 indicate potential overloading;
   719 
   720 * Isar/Pure: changed syntax of local blocks from {{ }} to { };
   721 
   722 * Isar/Pure: syntax of sorts made 'inner', i.e. have to write
   723 "{a,b,c}" instead of {a,b,c};
   724 
   725 * Isar/Pure now provides its own version of intro/elim/dest
   726 attributes; useful for building new logics, but beware of confusion
   727 with the version in Provers/classical;
   728 
   729 * Isar/Pure: the local context of (non-atomic) goals is provided via
   730 case name 'antecedent';
   731 
   732 * Isar/Pure: removed obsolete 'transfer' attribute (transfer of thms
   733 to the current context is now done automatically);
   734 
   735 * Isar/Pure: theory command 'method_setup' provides a simple interface
   736 for definining proof methods in ML;
   737 
   738 * Isar/Provers: intro/elim/dest attributes changed; renamed
   739 intro/intro!/intro!! flags to intro!/intro/intro? (INCOMPATIBILITY, in
   740 most cases, one should have to change intro!! to intro? only);
   741 replaced "delrule" by "rule del";
   742 
   743 * Isar/Provers: new 'hypsubst' method, plain 'subst' method and
   744 'symmetric' attribute (the latter supercedes [RS sym]);
   745 
   746 * Isar/Provers: splitter support (via 'split' attribute and 'simp'
   747 method modifier); 'simp' method: 'only:' modifier removes loopers as
   748 well (including splits);
   749 
   750 * Isar/Provers: Simplifier and Classical methods now support all kind
   751 of modifiers used in the past, including 'cong', 'iff', etc.
   752 
   753 * Isar/Provers: added 'fastsimp' and 'clarsimp' methods (combination
   754 of Simplifier and Classical reasoner);
   755 
   756 * Isar/HOL: new proof method 'cases' and improved version of 'induct'
   757 now support named cases; major packages (inductive, datatype, primrec,
   758 recdef) support case names and properly name parameters;
   759 
   760 * Isar/HOL: new transitivity rules for substitution in inequalities --
   761 monotonicity conditions are extracted to be proven at end of
   762 calculations;
   763 
   764 * Isar/HOL: removed 'case_split' thm binding, should use 'cases' proof
   765 method anyway;
   766 
   767 * Isar/HOL: removed old expand_if = split_if; theorems if_splits =
   768 split_if split_if_asm; datatype package provides theorems foo.splits =
   769 foo.split foo.split_asm for each datatype;
   770 
   771 * Isar/HOL: tuned inductive package, rename "intrs" to "intros"
   772 (potential INCOMPATIBILITY), emulation of mk_cases feature for proof
   773 scripts: new 'inductive_cases' command and 'ind_cases' method; (Note:
   774 use "(cases (simplified))" method in proper proof texts);
   775 
   776 * Isar/HOL: added global 'arith_split' attribute for 'arith' method;
   777 
   778 * Isar: names of theorems etc. may be natural numbers as well;
   779 
   780 * Isar: 'pr' command: optional arguments for goals_limit and
   781 ProofContext.prems_limit; no longer prints theory contexts, but only
   782 proof states;
   783 
   784 * Isar: diagnostic commands 'pr', 'thm', 'prop', 'term', 'typ' admit
   785 additional print modes to be specified; e.g. "pr(latex)" will print
   786 proof state according to the Isabelle LaTeX style;
   787 
   788 * Isar: improved support for emulating tactic scripts, including proof
   789 methods 'rule_tac' etc., 'cut_tac', 'thin_tac', 'subgoal_tac',
   790 'rename_tac', 'rotate_tac', 'tactic', and 'case_tac' / 'induct_tac'
   791 (for HOL datatypes);
   792 
   793 * Isar: simplified (more robust) goal selection of proof methods: 1st
   794 goal, all goals, or explicit goal specifier (tactic emulation); thus
   795 'proof method scripts' have to be in depth-first order;
   796 
   797 * Isar: tuned 'let' syntax: replaced 'as' keyword by 'and';
   798 
   799 * Isar: removed 'help' command, which hasn't been too helpful anyway;
   800 should instead use individual commands for printing items
   801 (print_commands, print_methods etc.);
   802 
   803 * Isar: added 'nothing' --- the empty list of theorems;
   804 
   805 
   806 *** HOL ***
   807 
   808 * HOL/MicroJava: formalization of a fragment of Java, together with a
   809 corresponding virtual machine and a specification of its bytecode
   810 verifier and a lightweight bytecode verifier, including proofs of
   811 type-safety; by Gerwin Klein, Tobias Nipkow, David von Oheimb, and
   812 Cornelia Pusch (see also the homepage of project Bali at
   813 http://isabelle.in.tum.de/Bali/);
   814 
   815 * HOL/Algebra: new theory of rings and univariate polynomials, by
   816 Clemens Ballarin;
   817 
   818 * HOL/NumberTheory: fundamental Theorem of Arithmetic, Chinese
   819 Remainder Theorem, Fermat/Euler Theorem, Wilson's Theorem, by Thomas M
   820 Rasmussen;
   821 
   822 * HOL/Lattice: fundamental concepts of lattice theory and order
   823 structures, including duals, properties of bounds versus algebraic
   824 laws, lattice operations versus set-theoretic ones, the Knaster-Tarski
   825 Theorem for complete lattices etc.; may also serve as a demonstration
   826 for abstract algebraic reasoning using axiomatic type classes, and
   827 mathematics-style proof in Isabelle/Isar; by Markus Wenzel;
   828 
   829 * HOL/Prolog: a (bare-bones) implementation of Lambda-Prolog, by David
   830 von Oheimb;
   831 
   832 * HOL/IMPP: extension of IMP with local variables and mutually
   833 recursive procedures, by David von Oheimb;
   834 
   835 * HOL/Lambda: converted into new-style theory and document;
   836 
   837 * HOL/ex/Multiquote: example of multiple nested quotations and
   838 anti-quotations -- basically a generalized version of de-Bruijn
   839 representation; very useful in avoiding lifting of operations;
   840 
   841 * HOL/record: added general record equality rule to simpset; fixed
   842 select-update simplification procedure to handle extended records as
   843 well; admit "r" as field name;
   844 
   845 * HOL: 0 is now overloaded over the new sort "zero", allowing its use with
   846 other numeric types and also as the identity of groups, rings, etc.;
   847 
   848 * HOL: new axclass plus_ac0 for addition with the AC-laws and 0 as identity.
   849 Types nat and int belong to this axclass;
   850 
   851 * HOL: greatly improved simplification involving numerals of type nat, int, real:
   852    (i + #8 + j) = Suc k simplifies to  #7 + (i + j) = k
   853    i*j + k + j*#3*i     simplifies to  #4*(i*j) + k
   854   two terms #m*u and #n*u are replaced by #(m+n)*u
   855     (where #m, #n and u can implicitly be 1; this is simproc combine_numerals)
   856   and the term/formula #m*u+x ~~ #n*u+y simplifies simplifies to #(m-n)+x ~~ y
   857     or x ~~ #(n-m)+y, where ~~ is one of = < <= or - (simproc cancel_numerals);
   858 
   859 * HOL: meson_tac is available (previously in ex/meson.ML); it is a
   860 powerful prover for predicate logic but knows nothing of clasets; see
   861 ex/mesontest.ML and ex/mesontest2.ML for example applications;
   862 
   863 * HOL: new version of "case_tac" subsumes both boolean case split and
   864 "exhaust_tac" on datatypes; INCOMPATIBILITY: exhaust_tac no longer
   865 exists, may define val exhaust_tac = case_tac for ad-hoc portability;
   866 
   867 * HOL: simplification no longer dives into case-expressions: only the
   868 selector expression is simplified, but not the remaining arms; to
   869 enable full simplification of case-expressions for datatype t, you may
   870 remove t.weak_case_cong from the simpset, either globally (Delcongs
   871 [thm"t.weak_case_cong"];) or locally (delcongs [...]).
   872 
   873 * HOL/recdef: the recursion equations generated by 'recdef' for
   874 function 'f' are now called f.simps instead of f.rules; if all
   875 termination conditions are proved automatically, these simplification
   876 rules are added to the simpset, as in primrec; rules may be named
   877 individually as well, resulting in a separate list of theorems for
   878 each equation;
   879 
   880 * HOL/While is a new theory that provides a while-combinator. It
   881 permits the definition of tail-recursive functions without the
   882 provision of a termination measure. The latter is necessary once the
   883 invariant proof rule for while is applied.
   884 
   885 * HOL: new (overloaded) notation for the set of elements below/above
   886 some element: {..u}, {..u(}, {l..}, {)l..}. See theory SetInterval.
   887 
   888 * HOL: theorems impI, allI, ballI bound as "strip";
   889 
   890 * HOL: new tactic induct_thm_tac: thm -> string -> int -> tactic
   891 induct_tac th "x1 ... xn" expects th to have a conclusion of the form
   892 P v1 ... vn and abbreviates res_inst_tac [("v1","x1"),...,("vn","xn")] th;
   893 
   894 * HOL/Real: "rabs" replaced by overloaded "abs" function;
   895 
   896 * HOL: theory Sexp now in HOL/Induct examples (it used to be part of
   897 main HOL, but was unused);
   898 
   899 * HOL: fewer consts declared as global (e.g. have to refer to
   900 "Lfp.lfp" instead of "lfp" internally; affects ML packages only);
   901 
   902 * HOL: tuned AST representation of nested pairs, avoiding bogus output
   903 in case of overlap with user translations (e.g. judgements over
   904 tuples); (note that the underlying logical represenation is still
   905 bogus);
   906 
   907 
   908 *** ZF ***
   909 
   910 * ZF: simplification automatically cancels common terms in arithmetic
   911 expressions over nat and int;
   912 
   913 * ZF: new treatment of nat to minimize type-checking: all operators
   914 coerce their operands to a natural number using the function natify,
   915 making the algebraic laws unconditional;
   916 
   917 * ZF: as above, for int: operators coerce their operands to an integer
   918 using the function intify;
   919 
   920 * ZF: the integer library now contains many of the usual laws for the
   921 orderings, including $<=, and monotonicity laws for $+ and $*;
   922 
   923 * ZF: new example ZF/ex/NatSum to demonstrate integer arithmetic
   924 simplification;
   925 
   926 * FOL and ZF: AddIffs now available, giving theorems of the form P<->Q
   927 to the simplifier and classical reasoner simultaneously;
   928 
   929 
   930 *** General ***
   931 
   932 * Provers: blast_tac now handles actual object-logic rules as
   933 assumptions; note that auto_tac uses blast_tac internally as well;
   934 
   935 * Provers: new functions rulify/rulify_no_asm: thm -> thm for turning
   936 outer -->/All/Ball into ==>/!!; qed_spec_mp now uses rulify_no_asm;
   937 
   938 * Provers: delrules now handles destruct rules as well (no longer need
   939 explicit make_elim);
   940 
   941 * Provers: Blast_tac now warns of and ignores "weak elimination rules" e.g.
   942   [| inj ?f;          ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
   943 use instead the strong form,
   944   [| inj ?f; ~ ?W ==> ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
   945 in HOL, FOL and ZF the function cla_make_elim will create such rules
   946 from destruct-rules;
   947 
   948 * Provers: Simplifier.easy_setup provides a fast path to basic
   949 Simplifier setup for new object-logics;
   950 
   951 * Pure: AST translation rules no longer require constant head on LHS;
   952 
   953 * Pure: improved name spaces: ambiguous output is qualified; support
   954 for hiding of names;
   955 
   956 * system: smart setup of canonical ML_HOME, ISABELLE_INTERFACE, and
   957 XSYMBOL_HOME; no longer need to do manual configuration in most
   958 situations;
   959 
   960 * system: compression of ML heaps images may now be controlled via -c
   961 option of isabelle and isatool usedir (currently only observed by
   962 Poly/ML);
   963 
   964 * system: isatool installfonts may handle X-Symbol fonts as well (very
   965 useful for remote X11);
   966 
   967 * system: provide TAGS file for Isabelle sources;
   968 
   969 * ML: infix 'OF' is a version of 'MRS' with more appropriate argument
   970 order;
   971 
   972 * ML: renamed flags Syntax.trace_norm_ast to Syntax.trace_ast; global
   973 timing flag supersedes proof_timing and Toplevel.trace;
   974 
   975 * ML: new combinators |>> and |>>> for incremental transformations
   976 with secondary results (e.g. certain theory extensions):
   977 
   978 * ML: PureThy.add_defs gets additional argument to indicate potential
   979 overloading (usually false);
   980 
   981 * ML: PureThy.add_thms/add_axioms/add_defs now return theorems as
   982 results;
   983 
   984 
   985 
   986 New in Isabelle99 (October 1999)
   987 --------------------------------
   988 
   989 *** Overview of INCOMPATIBILITIES (see below for more details) ***
   990 
   991 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)
   992 are no longer simplified.  (This allows the simplifier to unfold recursive
   993 functional programs.)  To restore the old behaviour, declare
   994 
   995     Delcongs [if_weak_cong];
   996 
   997 * HOL: Removed the obsolete syntax "Compl A"; use -A for set
   998 complement;
   999 
  1000 * HOL: the predicate "inj" is now defined by translation to "inj_on";
  1001 
  1002 * HOL/datatype: mutual_induct_tac no longer exists --
  1003   use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]
  1004 
  1005 * HOL/typedef: fixed type inference for representing set; type
  1006 arguments now have to occur explicitly on the rhs as type constraints;
  1007 
  1008 * ZF: The con_defs part of an inductive definition may no longer refer
  1009 to constants declared in the same theory;
  1010 
  1011 * HOL, ZF: the function mk_cases, generated by the inductive
  1012 definition package, has lost an argument.  To simplify its result, it
  1013 uses the default simpset instead of a supplied list of theorems.
  1014 
  1015 * HOL/List: the constructors of type list are now Nil and Cons;
  1016 
  1017 * Simplifier: the type of the infix ML functions
  1018         setSSolver addSSolver setSolver addSolver
  1019 is now  simpset * solver -> simpset  where `solver' is a new abstract type
  1020 for packaging solvers. A solver is created via
  1021         mk_solver: string -> (thm list -> int -> tactic) -> solver
  1022 where the string argument is only a comment.
  1023 
  1024 
  1025 *** Proof tools ***
  1026 
  1027 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a
  1028 decision procedure for linear arithmetic. Currently it is used for
  1029 types `nat', `int', and `real' in HOL (see below); it can, should and
  1030 will be instantiated for other types and logics as well.
  1031 
  1032 * The simplifier now accepts rewrite rules with flexible heads, eg
  1033      hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y
  1034   They are applied like any rule with a non-pattern lhs, i.e. by first-order
  1035   matching.
  1036 
  1037 
  1038 *** General ***
  1039 
  1040 * New Isabelle/Isar subsystem provides an alternative to traditional
  1041 tactical theorem proving; together with the ProofGeneral/isar user
  1042 interface it offers an interactive environment for developing human
  1043 readable proof documents (Isar == Intelligible semi-automated
  1044 reasoning); for further information see isatool doc isar-ref,
  1045 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/
  1046 
  1047 * improved and simplified presentation of theories: better HTML markup
  1048 (including colors), graph views in several sizes; isatool usedir now
  1049 provides a proper interface for user theories (via -P option); actual
  1050 document preparation based on (PDF)LaTeX is available as well (for
  1051 new-style theories only); see isatool doc system for more information;
  1052 
  1053 * native support for Proof General, both for classic Isabelle and
  1054 Isabelle/Isar;
  1055 
  1056 * ML function thm_deps visualizes dependencies of theorems and lemmas,
  1057 using the graph browser tool;
  1058 
  1059 * Isabelle manuals now also available as PDF;
  1060 
  1061 * theory loader rewritten from scratch (may not be fully
  1062 bug-compatible); old loadpath variable has been replaced by show_path,
  1063 add_path, del_path, reset_path functions; new operations such as
  1064 update_thy, touch_thy, remove_thy, use/update_thy_only (see also
  1065 isatool doc ref);
  1066 
  1067 * improved isatool install: option -k creates KDE application icon,
  1068 option -p DIR installs standalone binaries;
  1069 
  1070 * added ML_PLATFORM setting (useful for cross-platform installations);
  1071 more robust handling of platform specific ML images for SML/NJ;
  1072 
  1073 * the settings environment is now statically scoped, i.e. it is never
  1074 created again in sub-processes invoked from isabelle, isatool, or
  1075 Isabelle;
  1076 
  1077 * path element specification '~~' refers to '$ISABELLE_HOME';
  1078 
  1079 * in locales, the "assumes" and "defines" parts may be omitted if
  1080 empty;
  1081 
  1082 * new print_mode "xsymbols" for extended symbol support (e.g. genuine
  1083 long arrows);
  1084 
  1085 * new print_mode "HTML";
  1086 
  1087 * new flag show_tags controls display of tags of theorems (which are
  1088 basically just comments that may be attached by some tools);
  1089 
  1090 * Isamode 2.6 requires patch to accomodate change of Isabelle font
  1091 mode and goal output format:
  1092 
  1093 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el
  1094 244c244
  1095 <       (list (isa-getenv "ISABELLE") "-msymbols" logic-name)
  1096 ---
  1097 >       (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)
  1098 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el
  1099 181c181
  1100 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"
  1101 ---
  1102 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"
  1103 
  1104 * function bind_thms stores lists of theorems (cf. bind_thm);
  1105 
  1106 * new shorthand tactics ftac, eatac, datac, fatac;
  1107 
  1108 * qed (and friends) now accept "" as result name; in that case the
  1109 theorem is not stored, but proper checks and presentation of the
  1110 result still apply;
  1111 
  1112 * theorem database now also indexes constants "Trueprop", "all",
  1113 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;
  1114 
  1115 
  1116 *** HOL ***
  1117 
  1118 ** HOL arithmetic **
  1119 
  1120 * There are now decision procedures for linear arithmetic over nat and
  1121 int:
  1122 
  1123 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',
  1124 `+', `-', `Suc', `min', `max' and numerical constants; other subterms
  1125 are treated as atomic; subformulae not involving type `nat' or `int'
  1126 are ignored; quantified subformulae are ignored unless they are
  1127 positive universal or negative existential. The tactic has to be
  1128 invoked by hand and can be a little bit slow. In particular, the
  1129 running time is exponential in the number of occurrences of `min' and
  1130 `max', and `-' on `nat'.
  1131 
  1132 2. fast_arith_tac is a cut-down version of arith_tac: it only takes
  1133 (negated) (in)equalities among the premises and the conclusion into
  1134 account (i.e. no compound formulae) and does not know about `min' and
  1135 `max', and `-' on `nat'. It is fast and is used automatically by the
  1136 simplifier.
  1137 
  1138 NB: At the moment, these decision procedures do not cope with mixed
  1139 nat/int formulae where the two parts interact, such as `m < n ==>
  1140 int(m) < int(n)'.
  1141 
  1142 * HOL/Numeral provides a generic theory of numerals (encoded
  1143 efficiently as bit strings); setup for types nat/int/real is in place;
  1144 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than
  1145 int, existing theories and proof scripts may require a few additional
  1146 type constraints;
  1147 
  1148 * integer division and remainder can now be performed on constant
  1149 arguments;
  1150 
  1151 * many properties of integer multiplication, division and remainder
  1152 are now available;
  1153 
  1154 * An interface to the Stanford Validity Checker (SVC) is available through the
  1155 tactic svc_tac.  Propositional tautologies and theorems of linear arithmetic
  1156 are proved automatically.  SVC must be installed separately, and its results
  1157 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any
  1158 invocation of the underlying oracle).  For SVC see
  1159   http://verify.stanford.edu/SVC
  1160 
  1161 * IsaMakefile: the HOL-Real target now builds an actual image;
  1162 
  1163 
  1164 ** HOL misc **
  1165 
  1166 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces
  1167 (in Isabelle/Isar) -- by Gertrud Bauer;
  1168 
  1169 * HOL/BCV: generic model of bytecode verification, i.e. data-flow
  1170 analysis for assembly languages with subtypes;
  1171 
  1172 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization
  1173 -- avoids syntactic ambiguities and treats state, transition, and
  1174 temporal levels more uniformly; introduces INCOMPATIBILITIES due to
  1175 changed syntax and (many) tactics;
  1176 
  1177 * HOL/inductive: Now also handles more general introduction rules such
  1178   as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity
  1179   theorems are now maintained within the theory (maintained via the
  1180   "mono" attribute);
  1181 
  1182 * HOL/datatype: Now also handles arbitrarily branching datatypes
  1183   (using function types) such as
  1184 
  1185   datatype 'a tree = Atom 'a | Branch "nat => 'a tree"
  1186 
  1187 * HOL/record: record_simproc (part of the default simpset) takes care
  1188 of selectors applied to updated records; record_split_tac is no longer
  1189 part of the default claset; update_defs may now be removed from the
  1190 simpset in many cases; COMPATIBILITY: old behavior achieved by
  1191 
  1192   claset_ref () := claset() addSWrapper record_split_wrapper;
  1193   Delsimprocs [record_simproc]
  1194 
  1195 * HOL/typedef: fixed type inference for representing set; type
  1196 arguments now have to occur explicitly on the rhs as type constraints;
  1197 
  1198 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem
  1199 names rather than an ML expression;
  1200 
  1201 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be
  1202 supplied later.  Program schemes can be defined, such as
  1203     "While B C s = (if B s then While B C (C s) else s)"
  1204 where the well-founded relation can be chosen after B and C have been given.
  1205 
  1206 * HOL/List: the constructors of type list are now Nil and Cons;
  1207 INCOMPATIBILITY: while [] and infix # syntax is still there, of
  1208 course, ML tools referring to List.list.op # etc. have to be adapted;
  1209 
  1210 * HOL_quantifiers flag superseded by "HOL" print mode, which is
  1211 disabled by default; run isabelle with option -m HOL to get back to
  1212 the original Gordon/HOL-style output;
  1213 
  1214 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,
  1215 ALL x<=y. P, EX x<y. P, EX x<=y. P;
  1216 
  1217 * HOL basic syntax simplified (more orthogonal): all variants of
  1218 All/Ex now support plain / symbolic / HOL notation; plain syntax for
  1219 Eps operator is provided as well: "SOME x. P[x]";
  1220 
  1221 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;
  1222 
  1223 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made
  1224 thus available for user theories;
  1225 
  1226 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with
  1227 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the
  1228 time;
  1229 
  1230 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec
  1231 several times and then mp;
  1232 
  1233 
  1234 *** LK ***
  1235 
  1236 * the notation <<...>> is now available as a notation for sequences of
  1237 formulas;
  1238 
  1239 * the simplifier is now installed
  1240 
  1241 * the axiom system has been generalized (thanks to Soren Heilmann)
  1242 
  1243 * the classical reasoner now has a default rule database
  1244 
  1245 
  1246 *** ZF ***
  1247 
  1248 * new primrec section allows primitive recursive functions to be given
  1249 directly (as in HOL) over datatypes and the natural numbers;
  1250 
  1251 * new tactics induct_tac and exhaust_tac for induction (or case
  1252 analysis) over datatypes and the natural numbers;
  1253 
  1254 * the datatype declaration of type T now defines the recursor T_rec;
  1255 
  1256 * simplification automatically does freeness reasoning for datatype
  1257 constructors;
  1258 
  1259 * automatic type-inference, with AddTCs command to insert new
  1260 type-checking rules;
  1261 
  1262 * datatype introduction rules are now added as Safe Introduction rules
  1263 to the claset;
  1264 
  1265 * the syntax "if P then x else y" is now available in addition to
  1266 if(P,x,y);
  1267 
  1268 
  1269 *** Internal programming interfaces ***
  1270 
  1271 * tuned simplifier trace output; new flag debug_simp;
  1272 
  1273 * structures Vartab / Termtab (instances of TableFun) offer efficient
  1274 tables indexed by indexname_ord / term_ord (compatible with aconv);
  1275 
  1276 * AxClass.axclass_tac lost the theory argument;
  1277 
  1278 * tuned current_goals_markers semantics: begin / end goal avoids
  1279 printing empty lines;
  1280 
  1281 * removed prs and prs_fn hook, which was broken because it did not
  1282 include \n in its semantics, forcing writeln to add one
  1283 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:
  1284 string -> unit if you really want to output text without newline;
  1285 
  1286 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to
  1287 plain output, interface builders may have to enable 'isabelle_font'
  1288 mode to get Isabelle font glyphs as before;
  1289 
  1290 * refined token_translation interface; INCOMPATIBILITY: output length
  1291 now of type real instead of int;
  1292 
  1293 * theory loader actions may be traced via new ThyInfo.add_hook
  1294 interface (see src/Pure/Thy/thy_info.ML); example application: keep
  1295 your own database of information attached to *whole* theories -- as
  1296 opposed to intra-theory data slots offered via TheoryDataFun;
  1297 
  1298 * proper handling of dangling sort hypotheses (at last!);
  1299 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing
  1300 extra sort hypotheses that can be witnessed from the type signature;
  1301 the force_strip_shyps flag is gone, any remaining shyps are simply
  1302 left in the theorem (with a warning issued by strip_shyps_warning);
  1303 
  1304 
  1305 
  1306 New in Isabelle98-1 (October 1998)
  1307 ----------------------------------
  1308 
  1309 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  1310 
  1311 * several changes of automated proof tools;
  1312 
  1313 * HOL: major changes to the inductive and datatype packages, including
  1314 some minor incompatibilities of theory syntax;
  1315 
  1316 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now
  1317 called `inj_on';
  1318 
  1319 * HOL: removed duplicate thms in Arith:
  1320   less_imp_add_less  should be replaced by  trans_less_add1
  1321   le_imp_add_le      should be replaced by  trans_le_add1
  1322 
  1323 * HOL: unary minus is now overloaded (new type constraints may be
  1324 required);
  1325 
  1326 * HOL and ZF: unary minus for integers is now #- instead of #~.  In
  1327 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is
  1328 now taken as an integer constant.
  1329 
  1330 * Pure: ML function 'theory_of' renamed to 'theory';
  1331 
  1332 
  1333 *** Proof tools ***
  1334 
  1335 * Simplifier:
  1336   1. Asm_full_simp_tac is now more aggressive.
  1337      1. It will sometimes reorient premises if that increases their power to
  1338         simplify.
  1339      2. It does no longer proceed strictly from left to right but may also
  1340         rotate premises to achieve further simplification.
  1341      For compatibility reasons there is now Asm_lr_simp_tac which is like the
  1342      old Asm_full_simp_tac in that it does not rotate premises.
  1343   2. The simplifier now knows a little bit about nat-arithmetic.
  1344 
  1345 * Classical reasoner: wrapper mechanism for the classical reasoner now
  1346 allows for selected deletion of wrappers, by introduction of names for
  1347 wrapper functionals.  This implies that addbefore, addSbefore,
  1348 addaltern, and addSaltern now take a pair (name, tactic) as argument,
  1349 and that adding two tactics with the same name overwrites the first
  1350 one (emitting a warning).
  1351   type wrapper = (int -> tactic) -> (int -> tactic)
  1352   setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by
  1353   addWrapper, addSWrapper: claset * (string * wrapper) -> claset
  1354   delWrapper, delSWrapper: claset *  string            -> claset
  1355   getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;
  1356 
  1357 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE
  1358 semantics; addbefore now affects only the unsafe part of step_tac
  1359 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY
  1360 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac
  1361 by Force_tac;
  1362 
  1363 * Classical reasoner: setwrapper to setWrapper and compwrapper to
  1364 compWrapper; added safe wrapper (and access functions for it);
  1365 
  1366 * HOL/split_all_tac is now much faster and fails if there is nothing
  1367 to split.  Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order
  1368 and the names of the automatically generated variables have changed.
  1369 split_all_tac has moved within claset() from unsafe wrappers to safe
  1370 wrappers, which means that !!-bound variables are split much more
  1371 aggressively, and safe_tac and clarify_tac now split such variables.
  1372 If this splitting is not appropriate, use delSWrapper "split_all_tac".
  1373 Note: the same holds for record_split_tac, which does the job of
  1374 split_all_tac for record fields.
  1375 
  1376 * HOL/Simplifier: Rewrite rules for case distinctions can now be added
  1377 permanently to the default simpset using Addsplits just like
  1378 Addsimps. They can be removed via Delsplits just like
  1379 Delsimps. Lower-case versions are also available.
  1380 
  1381 * HOL/Simplifier: The rule split_if is now part of the default
  1382 simpset. This means that the simplifier will eliminate all occurrences
  1383 of if-then-else in the conclusion of a goal. To prevent this, you can
  1384 either remove split_if completely from the default simpset by
  1385 `Delsplits [split_if]' or remove it in a specific call of the
  1386 simplifier using `... delsplits [split_if]'.  You can also add/delete
  1387 other case splitting rules to/from the default simpset: every datatype
  1388 generates suitable rules `split_t_case' and `split_t_case_asm' (where
  1389 t is the name of the datatype).
  1390 
  1391 * Classical reasoner / Simplifier combination: new force_tac (and
  1392 derivatives Force_tac, force) combines rewriting and classical
  1393 reasoning (and whatever other tools) similarly to auto_tac, but is
  1394 aimed to solve the given subgoal completely.
  1395 
  1396 
  1397 *** General ***
  1398 
  1399 * new top-level commands `Goal' and `Goalw' that improve upon `goal'
  1400 and `goalw': the theory is no longer needed as an explicit argument -
  1401 the current theory context is used; assumptions are no longer returned
  1402 at the ML-level unless one of them starts with ==> or !!; it is
  1403 recommended to convert to these new commands using isatool fixgoal
  1404 (backup your sources first!);
  1405 
  1406 * new top-level commands 'thm' and 'thms' for retrieving theorems from
  1407 the current theory context, and 'theory' to lookup stored theories;
  1408 
  1409 * new theory section 'locale' for declaring constants, assumptions and
  1410 definitions that have local scope;
  1411 
  1412 * new theory section 'nonterminals' for purely syntactic types;
  1413 
  1414 * new theory section 'setup' for generic ML setup functions
  1415 (e.g. package initialization);
  1416 
  1417 * the distribution now includes Isabelle icons: see
  1418 lib/logo/isabelle-{small,tiny}.xpm;
  1419 
  1420 * isatool install - install binaries with absolute references to
  1421 ISABELLE_HOME/bin;
  1422 
  1423 * isatool logo -- create instances of the Isabelle logo (as EPS);
  1424 
  1425 * print mode 'emacs' reserved for Isamode;
  1426 
  1427 * support multiple print (ast) translations per constant name;
  1428 
  1429 * theorems involving oracles are now printed with a suffixed [!];
  1430 
  1431 
  1432 *** HOL ***
  1433 
  1434 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');
  1435 
  1436 * HOL/inductive package reorganized and improved: now supports mutual
  1437 definitions such as
  1438 
  1439   inductive EVEN ODD
  1440     intrs
  1441       null "0 : EVEN"
  1442       oddI "n : EVEN ==> Suc n : ODD"
  1443       evenI "n : ODD ==> Suc n : EVEN"
  1444 
  1445 new theorem list "elims" contains an elimination rule for each of the
  1446 recursive sets; inductive definitions now handle disjunctive premises
  1447 correctly (also ZF);
  1448 
  1449 INCOMPATIBILITIES: requires Inductive as an ancestor; component
  1450 "mutual_induct" no longer exists - the induction rule is always
  1451 contained in "induct";
  1452 
  1453 
  1454 * HOL/datatype package re-implemented and greatly improved: now
  1455 supports mutually recursive datatypes such as
  1456 
  1457   datatype
  1458     'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)
  1459             | SUM ('a aexp) ('a aexp)
  1460             | DIFF ('a aexp) ('a aexp)
  1461             | NUM 'a
  1462   and
  1463     'a bexp = LESS ('a aexp) ('a aexp)
  1464             | AND ('a bexp) ('a bexp)
  1465             | OR ('a bexp) ('a bexp)
  1466 
  1467 as well as indirectly recursive datatypes such as
  1468 
  1469   datatype
  1470     ('a, 'b) term = Var 'a
  1471                   | App 'b ((('a, 'b) term) list)
  1472 
  1473 The new tactic  mutual_induct_tac [<var_1>, ..., <var_n>] i  performs
  1474 induction on mutually / indirectly recursive datatypes.
  1475 
  1476 Primrec equations are now stored in theory and can be accessed via
  1477 <function_name>.simps.
  1478 
  1479 INCOMPATIBILITIES:
  1480 
  1481   - Theories using datatypes must now have theory Datatype as an
  1482     ancestor.
  1483   - The specific <typename>.induct_tac no longer exists - use the
  1484     generic induct_tac instead.
  1485   - natE has been renamed to nat.exhaust - use exhaust_tac
  1486     instead of res_inst_tac ... natE. Note that the variable
  1487     names in nat.exhaust differ from the names in natE, this
  1488     may cause some "fragile" proofs to fail.
  1489   - The theorems split_<typename>_case and split_<typename>_case_asm
  1490     have been renamed to <typename>.split and <typename>.split_asm.
  1491   - Since default sorts of type variables are now handled correctly,
  1492     some datatype definitions may have to be annotated with explicit
  1493     sort constraints.
  1494   - Primrec definitions no longer require function name and type
  1495     of recursive argument.
  1496 
  1497 Consider using isatool fixdatatype to adapt your theories and proof
  1498 scripts to the new package (backup your sources first!).
  1499 
  1500 
  1501 * HOL/record package: considerably improved implementation; now
  1502 includes concrete syntax for record types, terms, updates; theorems
  1503 for surjective pairing and splitting !!-bound record variables; proof
  1504 support is as follows:
  1505 
  1506   1) standard conversions (selectors or updates applied to record
  1507 constructor terms) are part of the standard simpset;
  1508 
  1509   2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are
  1510 made part of standard simpset and claset via addIffs;
  1511 
  1512   3) a tactic for record field splitting (record_split_tac) is part of
  1513 the standard claset (addSWrapper);
  1514 
  1515 To get a better idea about these rules you may retrieve them via
  1516 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is
  1517 the name of your record type.
  1518 
  1519 The split tactic 3) conceptually simplifies by the following rule:
  1520 
  1521   "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"
  1522 
  1523 Thus any record variable that is bound by meta-all will automatically
  1524 blow up into some record constructor term, consequently the
  1525 simplifications of 1), 2) apply.  Thus force_tac, auto_tac etc. shall
  1526 solve record problems automatically.
  1527 
  1528 
  1529 * reorganized the main HOL image: HOL/Integ and String loaded by
  1530 default; theory Main includes everything;
  1531 
  1532 * automatic simplification of integer sums and comparisons, using cancellation;
  1533 
  1534 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;
  1535 
  1536 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;
  1537 
  1538 * many new identities for unions, intersections, set difference, etc.;
  1539 
  1540 * expand_if, expand_split, expand_sum_case and expand_nat_case are now
  1541 called split_if, split_split, split_sum_case and split_nat_case (to go
  1542 with add/delsplits);
  1543 
  1544 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting
  1545 (?x::unit) = (); this is made part of the default simpset, which COULD
  1546 MAKE EXISTING PROOFS FAIL under rare circumstances (consider
  1547 'Delsimprocs [unit_eq_proc];' as last resort); also note that
  1548 unit_abs_eta_conv is added in order to counter the effect of
  1549 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by
  1550 %u.f();
  1551 
  1552 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which
  1553 makes more sense);
  1554 
  1555 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  1556   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  1557   disjointness reasoning but breaking a few old proofs.
  1558 
  1559 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1
  1560 to 'converse' from 'inverse' (for compatibility with ZF and some
  1561 literature);
  1562 
  1563 * HOL/recdef can now declare non-recursive functions, with {} supplied as
  1564 the well-founded relation;
  1565 
  1566 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of
  1567     Compl A.  The "Compl" syntax remains available as input syntax for this
  1568     release ONLY.
  1569 
  1570 * HOL/Update: new theory of function updates:
  1571     f(a:=b) == %x. if x=a then b else f x
  1572 may also be iterated as in f(a:=b,c:=d,...);
  1573 
  1574 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;
  1575 
  1576 * HOL/List:
  1577   - new function list_update written xs[i:=v] that updates the i-th
  1578     list position. May also be iterated as in xs[i:=a,j:=b,...].
  1579   - new function `upt' written [i..j(] which generates the list
  1580     [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper
  1581     bound write [i..j], which is a shorthand for [i..j+1(].
  1582   - new lexicographic orderings and corresponding wellfoundedness theorems.
  1583 
  1584 * HOL/Arith:
  1585   - removed 'pred' (predecessor) function;
  1586   - generalized some theorems about n-1;
  1587   - many new laws about "div" and "mod";
  1588   - new laws about greatest common divisors (see theory ex/Primes);
  1589 
  1590 * HOL/Relation: renamed the relational operator r^-1 "converse"
  1591 instead of "inverse";
  1592 
  1593 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness
  1594   of the multiset ordering;
  1595 
  1596 * directory HOL/Real: a construction of the reals using Dedekind cuts
  1597   (not included by default);
  1598 
  1599 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;
  1600 
  1601 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted
  1602   programs, i.e. different program variables may have different types.
  1603 
  1604 * calling (stac rew i) now fails if "rew" has no effect on the goal
  1605   [previously, this check worked only if the rewrite rule was unconditional]
  1606   Now rew can involve either definitions or equalities (either == or =).
  1607 
  1608 
  1609 *** ZF ***
  1610 
  1611 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains
  1612   only the theorems proved on ZF.ML;
  1613 
  1614 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  1615   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  1616   disjointness reasoning but breaking a few old proofs.
  1617 
  1618 * ZF/Update: new theory of function updates
  1619     with default rewrite rule  f(x:=y) ` z = if(z=x, y, f`z)
  1620   may also be iterated as in f(a:=b,c:=d,...);
  1621 
  1622 * in  let x=t in u(x), neither t nor u(x) has to be an FOL term.
  1623 
  1624 * calling (stac rew i) now fails if "rew" has no effect on the goal
  1625   [previously, this check worked only if the rewrite rule was unconditional]
  1626   Now rew can involve either definitions or equalities (either == or =).
  1627 
  1628 * case_tac provided for compatibility with HOL
  1629     (like the old excluded_middle_tac, but with subgoals swapped)
  1630 
  1631 
  1632 *** Internal programming interfaces ***
  1633 
  1634 * Pure: several new basic modules made available for general use, see
  1635 also src/Pure/README;
  1636 
  1637 * improved the theory data mechanism to support encapsulation (data
  1638 kind name replaced by private Object.kind, acting as authorization
  1639 key); new type-safe user interface via functor TheoryDataFun; generic
  1640 print_data function becomes basically useless;
  1641 
  1642 * removed global_names compatibility flag -- all theory declarations
  1643 are qualified by default;
  1644 
  1645 * module Pure/Syntax now offers quote / antiquote translation
  1646 functions (useful for Hoare logic etc. with implicit dependencies);
  1647 see HOL/ex/Antiquote for an example use;
  1648 
  1649 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->
  1650 cterm -> thm;
  1651 
  1652 * new tactical CHANGED_GOAL for checking that a tactic modifies a
  1653 subgoal;
  1654 
  1655 * Display.print_goals function moved to Locale.print_goals;
  1656 
  1657 * standard print function for goals supports current_goals_markers
  1658 variable for marking begin of proof, end of proof, start of goal; the
  1659 default is ("", "", ""); setting current_goals_markers := ("<proof>",
  1660 "</proof>", "<goal>") causes SGML like tagged proof state printing,
  1661 for example;
  1662 
  1663 
  1664 
  1665 New in Isabelle98 (January 1998)
  1666 --------------------------------
  1667 
  1668 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  1669 
  1670 * changed lexical syntax of terms / types: dots made part of long
  1671 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";
  1672 
  1673 * simpset (and claset) reference variable replaced by functions
  1674 simpset / simpset_ref;
  1675 
  1676 * no longer supports theory aliases (via merge) and non-trivial
  1677 implicit merge of thms' signatures;
  1678 
  1679 * most internal names of constants changed due to qualified names;
  1680 
  1681 * changed Pure/Sequence interface (see Pure/seq.ML);
  1682 
  1683 
  1684 *** General Changes ***
  1685 
  1686 * hierachically structured name spaces (for consts, types, axms, thms
  1687 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of
  1688 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:
  1689 isatool fixdots ensures space after dots (e.g. "%x. x"); set
  1690 long_names for fully qualified output names; NOTE: ML programs
  1691 (special tactics, packages etc.) referring to internal names may have
  1692 to be adapted to cope with fully qualified names; in case of severe
  1693 backward campatibility problems try setting 'global_names' at compile
  1694 time to have enrything declared within a flat name space; one may also
  1695 fine tune name declarations in theories via the 'global' and 'local'
  1696 section;
  1697 
  1698 * reimplemented the implicit simpset and claset using the new anytype
  1699 data filed in signatures; references simpset:simpset ref etc. are
  1700 replaced by functions simpset:unit->simpset and
  1701 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp
  1702 to patch your ML files accordingly;
  1703 
  1704 * HTML output now includes theory graph data for display with Java
  1705 applet or isatool browser; data generated automatically via isatool
  1706 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);
  1707 
  1708 * defs may now be conditional; improved rewrite_goals_tac to handle
  1709 conditional equations;
  1710 
  1711 * defs now admits additional type arguments, using TYPE('a) syntax;
  1712 
  1713 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always
  1714 creates a new theory node; implicit merge of thms' signatures is
  1715 restricted to 'trivial' ones; COMPATIBILITY: one may have to use
  1716 transfer:theory->thm->thm in (rare) cases;
  1717 
  1718 * improved handling of draft signatures / theories; draft thms (and
  1719 ctyps, cterms) are automatically promoted to real ones;
  1720 
  1721 * slightly changed interfaces for oracles: admit many per theory, named
  1722 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;
  1723 
  1724 * print_goals: optional output of const types (set show_consts and
  1725 show_types);
  1726 
  1727 * improved output of warnings (###) and errors (***);
  1728 
  1729 * subgoal_tac displays a warning if the new subgoal has type variables;
  1730 
  1731 * removed old README and Makefiles;
  1732 
  1733 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;
  1734 
  1735 * removed obsolete init_pps and init_database;
  1736 
  1737 * deleted the obsolete tactical STATE, which was declared by
  1738     fun STATE tacfun st = tacfun st st;
  1739 
  1740 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~
  1741 (which abbreviates $HOME);
  1742 
  1743 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:
  1744 use isatool fixseq to adapt your ML programs (this works for fully
  1745 qualified references to the Sequence structure only!);
  1746 
  1747 * use_thy no longer requires writable current directory; it always
  1748 reloads .ML *and* .thy file, if either one is out of date;
  1749 
  1750 
  1751 *** Classical Reasoner ***
  1752 
  1753 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new
  1754 tactics that use classical reasoning to simplify a subgoal without
  1755 splitting it into several subgoals;
  1756 
  1757 * Safe_tac: like safe_tac but uses the default claset;
  1758 
  1759 
  1760 *** Simplifier ***
  1761 
  1762 * added simplification meta rules:
  1763     (asm_)(full_)simplify: simpset -> thm -> thm;
  1764 
  1765 * simplifier.ML no longer part of Pure -- has to be loaded by object
  1766 logics (again);
  1767 
  1768 * added prems argument to simplification procedures;
  1769 
  1770 * HOL, FOL, ZF: added infix function `addsplits':
  1771   instead of `<simpset> setloop (split_tac <thms>)'
  1772   you can simply write `<simpset> addsplits <thms>'
  1773 
  1774 
  1775 *** Syntax ***
  1776 
  1777 * TYPE('a) syntax for type reflection terms;
  1778 
  1779 * no longer handles consts with name "" -- declare as 'syntax' instead;
  1780 
  1781 * pretty printer: changed order of mixfix annotation preference (again!);
  1782 
  1783 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;
  1784 
  1785 
  1786 *** HOL ***
  1787 
  1788 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.
  1789   with `addloop' of the simplifier to faciliate case splitting in premises.
  1790 
  1791 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;
  1792 
  1793 * HOL/Auth: new protocol proofs including some for the Internet
  1794   protocol TLS;
  1795 
  1796 * HOL/Map: new theory of `maps' a la VDM;
  1797 
  1798 * HOL/simplifier: simplification procedures nat_cancel_sums for
  1799 cancelling out common nat summands from =, <, <= (in)equalities, or
  1800 differences; simplification procedures nat_cancel_factor for
  1801 cancelling common factor from =, <, <= (in)equalities over natural
  1802 sums; nat_cancel contains both kinds of procedures, it is installed by
  1803 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;
  1804 
  1805 * HOL/simplifier: terms of the form
  1806   `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)'  (or t=x)
  1807   are rewritten to
  1808   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',
  1809   and those of the form
  1810   `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)'  (or t=x)
  1811   are rewritten to
  1812   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',
  1813 
  1814 * HOL/datatype
  1815   Each datatype `t' now comes with a theorem `split_t_case' of the form
  1816 
  1817   P(t_case f1 ... fn x) =
  1818      ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &
  1819         ...
  1820        (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))
  1821      )
  1822 
  1823   and a theorem `split_t_case_asm' of the form
  1824 
  1825   P(t_case f1 ... fn x) =
  1826     ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |
  1827         ...
  1828        (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))
  1829      )
  1830   which can be added to a simpset via `addsplits'. The existing theorems
  1831   expand_list_case and expand_option_case have been renamed to
  1832   split_list_case and split_option_case.
  1833 
  1834 * HOL/Arithmetic:
  1835   - `pred n' is automatically converted to `n-1'.
  1836     Users are strongly encouraged not to use `pred' any longer,
  1837     because it will disappear altogether at some point.
  1838   - Users are strongly encouraged to write "0 < n" rather than
  1839     "n ~= 0". Theorems and proof tools have been modified towards this
  1840     `standard'.
  1841 
  1842 * HOL/Lists:
  1843   the function "set_of_list" has been renamed "set" (and its theorems too);
  1844   the function "nth" now takes its arguments in the reverse order and
  1845   has acquired the infix notation "!" as in "xs!n".
  1846 
  1847 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};
  1848 
  1849 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its
  1850   specialist theorems (like UN1_I) are gone.  Similarly for (INT x.B x);
  1851 
  1852 * HOL/record: extensible records with schematic structural subtyping
  1853 (single inheritance); EXPERIMENTAL version demonstrating the encoding,
  1854 still lacks various theorems and concrete record syntax;
  1855 
  1856 
  1857 *** HOLCF ***
  1858 
  1859 * removed "axioms" and "generated by" sections;
  1860 
  1861 * replaced "ops" section by extended "consts" section, which is capable of
  1862   handling the continuous function space "->" directly;
  1863 
  1864 * domain package:
  1865   . proves theorems immediately and stores them in the theory,
  1866   . creates hierachical name space,
  1867   . now uses normal mixfix annotations (instead of cinfix...),
  1868   . minor changes to some names and values (for consistency),
  1869   . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,
  1870   . separator between mutual domain defs: changed "," to "and",
  1871   . improved handling of sort constraints;  now they have to
  1872     appear on the left-hand side of the equations only;
  1873 
  1874 * fixed LAM <x,y,zs>.b syntax;
  1875 
  1876 * added extended adm_tac to simplifier in HOLCF -- can now discharge
  1877 adm (%x. P (t x)), where P is chainfinite and t continuous;
  1878 
  1879 
  1880 *** FOL and ZF ***
  1881 
  1882 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.
  1883   with `addloop' of the simplifier to faciliate case splitting in premises.
  1884 
  1885 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as
  1886 in HOL, they strip ALL and --> from proved theorems;
  1887 
  1888 
  1889 
  1890 New in Isabelle94-8 (May 1997)
  1891 ------------------------------
  1892 
  1893 *** General Changes ***
  1894 
  1895 * new utilities to build / run / maintain Isabelle etc. (in parts
  1896 still somewhat experimental); old Makefiles etc. still functional;
  1897 
  1898 * new 'Isabelle System Manual';
  1899 
  1900 * INSTALL text, together with ./configure and ./build scripts;
  1901 
  1902 * reimplemented type inference for greater efficiency, better error
  1903 messages and clean internal interface;
  1904 
  1905 * prlim command for dealing with lots of subgoals (an easier way of
  1906 setting goals_limit);
  1907 
  1908 
  1909 *** Syntax ***
  1910 
  1911 * supports alternative (named) syntax tables (parser and pretty
  1912 printer); internal interface is provided by add_modesyntax(_i);
  1913 
  1914 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to
  1915 be used in conjunction with the Isabelle symbol font; uses the
  1916 "symbols" syntax table;
  1917 
  1918 * added token_translation interface (may translate name tokens in
  1919 arbitrary ways, dependent on their type (free, bound, tfree, ...) and
  1920 the current print_mode); IMPORTANT: user print translation functions
  1921 are responsible for marking newly introduced bounds
  1922 (Syntax.mark_boundT);
  1923 
  1924 * token translations for modes "xterm" and "xterm_color" that display
  1925 names in bold, underline etc. or colors (which requires a color
  1926 version of xterm);
  1927 
  1928 * infixes may now be declared with names independent of their syntax;
  1929 
  1930 * added typed_print_translation (like print_translation, but may
  1931 access type of constant);
  1932 
  1933 
  1934 *** Classical Reasoner ***
  1935 
  1936 Blast_tac: a new tactic!  It is often more powerful than fast_tac, but has
  1937 some limitations.  Blast_tac...
  1938   + ignores addss, addbefore, addafter; this restriction is intrinsic
  1939   + ignores elimination rules that don't have the correct format
  1940         (the conclusion MUST be a formula variable)
  1941   + ignores types, which can make HOL proofs fail
  1942   + rules must not require higher-order unification, e.g. apply_type in ZF
  1943     [message "Function Var's argument not a bound variable" relates to this]
  1944   + its proof strategy is more general but can actually be slower
  1945 
  1946 * substitution with equality assumptions no longer permutes other
  1947 assumptions;
  1948 
  1949 * minor changes in semantics of addafter (now called addaltern); renamed
  1950 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper
  1951 (and access functions for it);
  1952 
  1953 * improved combination of classical reasoner and simplifier:
  1954   + functions for handling clasimpsets
  1955   + improvement of addss: now the simplifier is called _after_ the
  1956     safe steps.
  1957   + safe variant of addss called addSss: uses safe simplifications
  1958     _during_ the safe steps. It is more complete as it allows multiple
  1959     instantiations of unknowns (e.g. with slow_tac).
  1960 
  1961 *** Simplifier ***
  1962 
  1963 * added interface for simplification procedures (functions that
  1964 produce *proven* rewrite rules on the fly, depending on current
  1965 redex);
  1966 
  1967 * ordering on terms as parameter (used for ordered rewriting);
  1968 
  1969 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;
  1970 
  1971 * the solver is now split into a safe and an unsafe part.
  1972 This should be invisible for the normal user, except that the
  1973 functions setsolver and addsolver have been renamed to setSolver and
  1974 addSolver; added safe_asm_full_simp_tac;
  1975 
  1976 
  1977 *** HOL ***
  1978 
  1979 * a generic induction tactic `induct_tac' which works for all datatypes and
  1980 also for type `nat';
  1981 
  1982 * a generic case distinction tactic `exhaust_tac' which works for all
  1983 datatypes and also for type `nat';
  1984 
  1985 * each datatype comes with a function `size';
  1986 
  1987 * patterns in case expressions allow tuple patterns as arguments to
  1988 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';
  1989 
  1990 * primrec now also works with type nat;
  1991 
  1992 * recdef: a new declaration form, allows general recursive functions to be
  1993 defined in theory files.  See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.
  1994 
  1995 * the constant for negation has been renamed from "not" to "Not" to
  1996 harmonize with FOL, ZF, LK, etc.;
  1997 
  1998 * HOL/ex/LFilter theory of a corecursive "filter" functional for
  1999 infinite lists;
  2000 
  2001 * HOL/Modelcheck demonstrates invocation of model checker oracle;
  2002 
  2003 * HOL/ex/Ring.thy declares cring_simp, which solves equational
  2004 problems in commutative rings, using axiomatic type classes for + and *;
  2005 
  2006 * more examples in HOL/MiniML and HOL/Auth;
  2007 
  2008 * more default rewrite rules for quantifiers, union/intersection;
  2009 
  2010 * a new constant `arbitrary == @x.False';
  2011 
  2012 * HOLCF/IOA replaces old HOL/IOA;
  2013 
  2014 * HOLCF changes: derived all rules and arities
  2015   + axiomatic type classes instead of classes
  2016   + typedef instead of faking type definitions
  2017   + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.
  2018   + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po
  2019   + eliminated the types void, one, tr
  2020   + use unit lift and bool lift (with translations) instead of one and tr
  2021   + eliminated blift from Lift3.thy (use Def instead of blift)
  2022   all eliminated rules are derived as theorems --> no visible changes ;
  2023 
  2024 
  2025 *** ZF ***
  2026 
  2027 * ZF now has Fast_tac, Simp_tac and Auto_tac.  Union_iff is a now a default
  2028 rewrite rule; this may affect some proofs.  eq_cs is gone but can be put back
  2029 as ZF_cs addSIs [equalityI];
  2030 
  2031 
  2032 
  2033 New in Isabelle94-7 (November 96)
  2034 ---------------------------------
  2035 
  2036 * allowing negative levels (as offsets) in prlev and choplev;
  2037 
  2038 * super-linear speedup for large simplifications;
  2039 
  2040 * FOL, ZF and HOL now use miniscoping: rewriting pushes
  2041 quantifications in as far as possible (COULD MAKE EXISTING PROOFS
  2042 FAIL); can suppress it using the command Delsimps (ex_simps @
  2043 all_simps); De Morgan laws are also now included, by default;
  2044 
  2045 * improved printing of ==>  :  ~:
  2046 
  2047 * new object-logic "Sequents" adds linear logic, while replacing LK
  2048 and Modal (thanks to Sara Kalvala);
  2049 
  2050 * HOL/Auth: correctness proofs for authentication protocols;
  2051 
  2052 * HOL: new auto_tac combines rewriting and classical reasoning (many
  2053 examples on HOL/Auth);
  2054 
  2055 * HOL: new command AddIffs for declaring theorems of the form P=Q to
  2056 the rewriter and classical reasoner simultaneously;
  2057 
  2058 * function uresult no longer returns theorems in "standard" format;
  2059 regain previous version by: val uresult = standard o uresult;
  2060 
  2061 
  2062 
  2063 New in Isabelle94-6
  2064 -------------------
  2065 
  2066 * oracles -- these establish an interface between Isabelle and trusted
  2067 external reasoners, which may deliver results as theorems;
  2068 
  2069 * proof objects (in particular record all uses of oracles);
  2070 
  2071 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;
  2072 
  2073 * "constdefs" section in theory files;
  2074 
  2075 * "primrec" section (HOL) no longer requires names;
  2076 
  2077 * internal type "tactic" now simply "thm -> thm Sequence.seq";
  2078 
  2079 
  2080 
  2081 New in Isabelle94-5
  2082 -------------------
  2083 
  2084 * reduced space requirements;
  2085 
  2086 * automatic HTML generation from theories;
  2087 
  2088 * theory files no longer require "..." (quotes) around most types;
  2089 
  2090 * new examples, including two proofs of the Church-Rosser theorem;
  2091 
  2092 * non-curried (1994) version of HOL is no longer distributed;
  2093 
  2094 
  2095 
  2096 New in Isabelle94-4
  2097 -------------------
  2098 
  2099 * greatly reduced space requirements;
  2100 
  2101 * theory files (.thy) no longer require \...\ escapes at line breaks;
  2102 
  2103 * searchable theorem database (see the section "Retrieving theorems" on
  2104 page 8 of the Reference Manual);
  2105 
  2106 * new examples, including Grabczewski's monumental case study of the
  2107 Axiom of Choice;
  2108 
  2109 * The previous version of HOL renamed to Old_HOL;
  2110 
  2111 * The new version of HOL (previously called CHOL) uses a curried syntax
  2112 for functions.  Application looks like f a b instead of f(a,b);
  2113 
  2114 * Mutually recursive inductive definitions finally work in HOL;
  2115 
  2116 * In ZF, pattern-matching on tuples is now available in all abstractions and
  2117 translates to the operator "split";
  2118 
  2119 
  2120 
  2121 New in Isabelle94-3
  2122 -------------------
  2123 
  2124 * new infix operator, addss, allowing the classical reasoner to
  2125 perform simplification at each step of its search.  Example:
  2126         fast_tac (cs addss ss)
  2127 
  2128 * a new logic, CHOL, the same as HOL, but with a curried syntax
  2129 for functions.  Application looks like f a b instead of f(a,b).  Also pairs
  2130 look like (a,b) instead of <a,b>;
  2131 
  2132 * PLEASE NOTE: CHOL will eventually replace HOL!
  2133 
  2134 * In CHOL, pattern-matching on tuples is now available in all abstractions.
  2135 It translates to the operator "split".  A new theory of integers is available;
  2136 
  2137 * In ZF, integer numerals now denote two's-complement binary integers.
  2138 Arithmetic operations can be performed by rewriting.  See ZF/ex/Bin.ML;
  2139 
  2140 * Many new examples: I/O automata, Church-Rosser theorem, equivalents
  2141 of the Axiom of Choice;
  2142 
  2143 
  2144 
  2145 New in Isabelle94-2
  2146 -------------------
  2147 
  2148 * Significantly faster resolution;
  2149 
  2150 * the different sections in a .thy file can now be mixed and repeated
  2151 freely;
  2152 
  2153 * Database of theorems for FOL, HOL and ZF.  New
  2154 commands including qed, qed_goal and bind_thm store theorems in the database.
  2155 
  2156 * Simple database queries: return a named theorem (get_thm) or all theorems of
  2157 a given theory (thms_of), or find out what theory a theorem was proved in
  2158 (theory_of_thm);
  2159 
  2160 * Bugs fixed in the inductive definition and datatype packages;
  2161 
  2162 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs
  2163 and HOL_dup_cs obsolete;
  2164 
  2165 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1
  2166 have been removed;
  2167 
  2168 * Simpler definition of function space in ZF;
  2169 
  2170 * new results about cardinal and ordinal arithmetic in ZF;
  2171 
  2172 * 'subtype' facility in HOL for introducing new types as subsets of existing
  2173 types;
  2174 
  2175 
  2176 $Id$