summary |
shortlog |
changelog |
graph |
tags |
branches |
files |
changeset |
file |
revisions |
annotate |
diff |
raw

NEWS

author | wenzelm |

Sat Oct 30 20:41:30 1999 +0200 (1999-10-30) | |

changeset 7986 | 9d319a76dbeb |

parent 7919 | 35c18affc1d8 |

child 8007 | c29e27ee4933 |

permissions | -rw-r--r-- |

Isabelle99;

2 Isabelle NEWS -- history user-relevant changes

3 ==============================================

5 New in Isabelle99 (October 1999)

6 --------------------------------

8 *** Overview of INCOMPATIBILITIES (see below for more details) ***

10 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)

11 are no longer simplified. (This allows the simplifier to unfold recursive

12 functional programs.) To restore the old behaviour, declare

14 Delcongs [if_weak_cong];

16 * HOL: Removed the obsolete syntax "Compl A"; use -A for set

17 complement;

19 * HOL: the predicate "inj" is now defined by translation to "inj_on";

21 * HOL/datatype: mutual_induct_tac no longer exists --

22 use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]

24 * HOL/typedef: fixed type inference for representing set; type

25 arguments now have to occur explicitly on the rhs as type constraints;

27 * ZF: The con_defs part of an inductive definition may no longer refer

28 to constants declared in the same theory;

30 * HOL, ZF: the function mk_cases, generated by the inductive

31 definition package, has lost an argument. To simplify its result, it

32 uses the default simpset instead of a supplied list of theorems.

34 * HOL/List: the constructors of type list are now Nil and Cons;

36 * Simplifier: the type of the infix ML functions

37 setSSolver addSSolver setSolver addSolver

38 is now simpset * solver -> simpset where `solver' is a new abstract type

39 for packaging solvers. A solver is created via

40 mk_solver: string -> (thm list -> int -> tactic) -> solver

41 where the string argument is only a comment.

44 *** Proof tools ***

46 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a

47 decision procedure for linear arithmetic. Currently it is used for

48 types `nat', `int', and `real' in HOL (see below); it can, should and

49 will be instantiated for other types and logics as well.

51 * The simplifier now accepts rewrite rules with flexible heads, eg

52 hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y

53 They are applied like any rule with a non-pattern lhs, i.e. by first-order

54 matching.

57 *** General ***

59 * New Isabelle/Isar subsystem provides an alternative to traditional

60 tactical theorem proving; together with the ProofGeneral/isar user

61 interface it offers an interactive environment for developing human

62 readable proof documents (Isar == Intelligible semi-automated

63 reasoning); for further information see isatool doc isar-ref,

64 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/

66 * improved presentation of theories: better HTML markup (including

67 colors), graph views in several sizes; isatool usedir now provides a

68 proper interface for user theories (via -P option); actual document

69 preparation based on (PDF)LaTeX is available as well (for new-style

70 theories only); see isatool doc system for more information;

72 * native support for Proof General, both for classic Isabelle and

73 Isabelle/Isar;

75 * ML function thm_deps visualizes dependencies of theorems and lemmas,

76 using the graph browser tool;

78 * Isabelle manuals now also available as PDF;

80 * theory loader rewritten from scratch (may not be fully

81 bug-compatible); old loadpath variable has been replaced by show_path,

82 add_path, del_path, reset_path functions; new operations such as

83 update_thy, touch_thy, remove_thy, use/update_thy_only (see also

84 isatool doc ref);

86 * improved isatool install: option -k creates KDE application icon,

87 option -p DIR installs standalone binaries;

89 * added ML_PLATFORM setting (useful for cross-platform installations);

90 more robust handling of platform specific ML images for SML/NJ;

92 * the settings environment is now statically scoped, i.e. it is never

93 created again in sub-processes invoked from isabelle, isatool, or

94 Isabelle;

96 * path element specification '~~' refers to '$ISABELLE_HOME';

98 * in locales, the "assumes" and "defines" parts may be omitted if

99 empty;

101 * new print_mode "xsymbols" for extended symbol support (e.g. genuine

102 long arrows);

104 * new print_mode "HTML";

106 * new flag show_tags controls display of tags of theorems (which are

107 basically just comments that may be attached by some tools);

109 * Isamode 2.6 requires patch to accomodate change of Isabelle font

110 mode and goal output format:

112 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el

113 244c244

114 < (list (isa-getenv "ISABELLE") "-msymbols" logic-name)

115 ---

116 > (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)

117 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el

118 181c181

119 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"

120 ---

121 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"

123 * function bind_thms stores lists of theorems (cf. bind_thm);

125 * new shorthand tactics ftac, eatac, datac, fatac;

127 * qed (and friends) now accept "" as result name; in that case the

128 theorem is not stored, but proper checks and presentation of the

129 result still apply;

131 * theorem database now also indexes constants "Trueprop", "all",

132 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;

135 *** HOL ***

137 ** HOL arithmetic **

139 * There are now decision procedures for linear arithmetic over nat and

140 int:

142 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',

143 `+', `-', `Suc', `min', `max' and numerical constants; other subterms

144 are treated as atomic; subformulae not involving type `nat' or `int'

145 are ignored; quantified subformulae are ignored unless they are

146 positive universal or negative existential. The tactic has to be

147 invoked by hand and can be a little bit slow. In particular, the

148 running time is exponential in the number of occurrences of `min' and

149 `max', and `-' on `nat'.

151 2. fast_arith_tac is a cut-down version of arith_tac: it only takes

152 (negated) (in)equalities among the premises and the conclusion into

153 account (i.e. no compound formulae) and does not know about `min' and

154 `max', and `-' on `nat'. It is fast and is used automatically by the

155 simplifier.

157 NB: At the moment, these decision procedures do not cope with mixed

158 nat/int formulae where the two parts interact, such as `m < n ==>

159 int(m) < int(n)'.

161 * HOL/Numeral provides a generic theory of numerals (encoded

162 efficiently as bit strings); setup for types nat/int/real is in place;

163 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than

164 int, existing theories and proof scripts may require a few additional

165 type constraints;

167 * integer division and remainder can now be performed on constant

168 arguments;

170 * many properties of integer multiplication, division and remainder

171 are now available;

173 * An interface to the Stanford Validity Checker (SVC) is available through the

174 tactic svc_tac. Propositional tautologies and theorems of linear arithmetic

175 are proved automatically. SVC must be installed separately, and its results

176 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any

177 invocation of the underlying oracle). For SVC see

178 http://verify.stanford.edu/SVC

180 * IsaMakefile: the HOL-Real target now builds an actual image;

183 ** HOL misc **

185 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces

186 (in Isabelle/Isar) -- by Gertrud Bauer;

188 * HOL/BCV: generic model of bytecode verification, i.e. data-flow

189 analysis for assembly languages with subtypes;

191 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization

192 -- avoids syntactic ambiguities and treats state, transition, and

193 temporal levels more uniformly; introduces INCOMPATIBILITIES due to

194 changed syntax and (many) tactics;

196 * HOL/inductive: Now also handles more general introduction rules such

197 as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity

198 theorems are now maintained within the theory (maintained via the

199 "mono" attribute);

201 * HOL/datatype: Now also handles arbitrarily branching datatypes

202 (using function types) such as

204 datatype 'a tree = Atom 'a | Branch "nat => 'a tree"

206 * HOL/record: record_simproc (part of the default simpset) takes care

207 of selectors applied to updated records; record_split_tac is no longer

208 part of the default claset; update_defs may now be removed from the

209 simpset in many cases; COMPATIBILITY: old behavior achieved by

211 claset_ref () := claset() addSWrapper record_split_wrapper;

212 Delsimprocs [record_simproc]

214 * HOL/typedef: fixed type inference for representing set; type

215 arguments now have to occur explicitly on the rhs as type constraints;

217 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem

218 names rather than an ML expression;

220 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be

221 supplied later. Program schemes can be defined, such as

222 "While B C s = (if B s then While B C (C s) else s)"

223 where the well-founded relation can be chosen after B and C have been given.

225 * HOL/List: the constructors of type list are now Nil and Cons;

226 INCOMPATIBILITY: while [] and infix # syntax is still there, of

227 course, ML tools referring to List.list.op # etc. have to be adapted;

229 * HOL_quantifiers flag superseded by "HOL" print mode, which is

230 disabled by default; run isabelle with option -m HOL to get back to

231 the original Gordon/HOL-style output;

233 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,

234 ALL x<=y. P, EX x<y. P, EX x<=y. P;

236 * HOL basic syntax simplified (more orthogonal): all variants of

237 All/Ex now support plain / symbolic / HOL notation; plain syntax for

238 Eps operator is provided as well: "SOME x. P[x]";

240 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;

242 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made

243 thus available for user theories;

245 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with

246 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the

247 time;

249 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec

250 several times and then mp;

253 *** LK ***

255 * the notation <<...>> is now available as a notation for sequences of

256 formulas;

258 * the simplifier is now installed

260 * the axiom system has been generalized (thanks to Soren Heilmann)

262 * the classical reasoner now has a default rule database

265 *** ZF ***

267 * new primrec section allows primitive recursive functions to be given

268 directly (as in HOL) over datatypes and the natural numbers;

270 * new tactics induct_tac and exhaust_tac for induction (or case

271 analysis) over datatypes and the natural numbers;

273 * the datatype declaration of type T now defines the recursor T_rec;

275 * simplification automatically does freeness reasoning for datatype

276 constructors;

278 * automatic type-inference, with AddTCs command to insert new

279 type-checking rules;

281 * datatype introduction rules are now added as Safe Introduction rules

282 to the claset;

284 * the syntax "if P then x else y" is now available in addition to

285 if(P,x,y);

288 *** Internal programming interfaces ***

290 * tuned simplifier trace output; new flag debug_simp;

292 * structures Vartab / Termtab (instances of TableFun) offer efficient

293 tables indexed by indexname_ord / term_ord (compatible with aconv);

295 * AxClass.axclass_tac lost the theory argument;

297 * tuned current_goals_markers semantics: begin / end goal avoids

298 printing empty lines;

300 * removed prs and prs_fn hook, which was broken because it did not

301 include \n in its semantics, forcing writeln to add one

302 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:

303 string -> unit if you really want to output text without newline;

305 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to

306 plain output, interface builders may have to enable 'isabelle_font'

307 mode to get Isabelle font glyphs as before;

309 * refined token_translation interface; INCOMPATIBILITY: output length

310 now of type real instead of int;

312 * theory loader actions may be traced via new ThyInfo.add_hook

313 interface (see src/Pure/Thy/thy_info.ML); example application: keep

314 your own database of information attached to *whole* theories -- as

315 opposed to intra-theory data slots offered via TheoryDataFun;

317 * proper handling of dangling sort hypotheses (at last!);

318 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing

319 extra sort hypotheses that can be witnessed from the type signature;

320 the force_strip_shyps flag is gone, any remaining shyps are simply

321 left in the theorem (with a warning issued by strip_shyps_warning);

325 New in Isabelle98-1 (October 1998)

326 ----------------------------------

328 *** Overview of INCOMPATIBILITIES (see below for more details) ***

330 * several changes of automated proof tools;

332 * HOL: major changes to the inductive and datatype packages, including

333 some minor incompatibilities of theory syntax;

335 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now

336 called `inj_on';

338 * HOL: removed duplicate thms in Arith:

339 less_imp_add_less should be replaced by trans_less_add1

340 le_imp_add_le should be replaced by trans_le_add1

342 * HOL: unary minus is now overloaded (new type constraints may be

343 required);

345 * HOL and ZF: unary minus for integers is now #- instead of #~. In

346 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is

347 now taken as an integer constant.

349 * Pure: ML function 'theory_of' renamed to 'theory';

352 *** Proof tools ***

354 * Simplifier:

355 1. Asm_full_simp_tac is now more aggressive.

356 1. It will sometimes reorient premises if that increases their power to

357 simplify.

358 2. It does no longer proceed strictly from left to right but may also

359 rotate premises to achieve further simplification.

360 For compatibility reasons there is now Asm_lr_simp_tac which is like the

361 old Asm_full_simp_tac in that it does not rotate premises.

362 2. The simplifier now knows a little bit about nat-arithmetic.

364 * Classical reasoner: wrapper mechanism for the classical reasoner now

365 allows for selected deletion of wrappers, by introduction of names for

366 wrapper functionals. This implies that addbefore, addSbefore,

367 addaltern, and addSaltern now take a pair (name, tactic) as argument,

368 and that adding two tactics with the same name overwrites the first

369 one (emitting a warning).

370 type wrapper = (int -> tactic) -> (int -> tactic)

371 setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by

372 addWrapper, addSWrapper: claset * (string * wrapper) -> claset

373 delWrapper, delSWrapper: claset * string -> claset

374 getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;

376 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE

377 semantics; addbefore now affects only the unsafe part of step_tac

378 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY

379 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac

380 by Force_tac;

382 * Classical reasoner: setwrapper to setWrapper and compwrapper to

383 compWrapper; added safe wrapper (and access functions for it);

385 * HOL/split_all_tac is now much faster and fails if there is nothing

386 to split. Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order

387 and the names of the automatically generated variables have changed.

388 split_all_tac has moved within claset() from unsafe wrappers to safe

389 wrappers, which means that !!-bound variables are split much more

390 aggressively, and safe_tac and clarify_tac now split such variables.

391 If this splitting is not appropriate, use delSWrapper "split_all_tac".

392 Note: the same holds for record_split_tac, which does the job of

393 split_all_tac for record fields.

395 * HOL/Simplifier: Rewrite rules for case distinctions can now be added

396 permanently to the default simpset using Addsplits just like

397 Addsimps. They can be removed via Delsplits just like

398 Delsimps. Lower-case versions are also available.

400 * HOL/Simplifier: The rule split_if is now part of the default

401 simpset. This means that the simplifier will eliminate all occurrences

402 of if-then-else in the conclusion of a goal. To prevent this, you can

403 either remove split_if completely from the default simpset by

404 `Delsplits [split_if]' or remove it in a specific call of the

405 simplifier using `... delsplits [split_if]'. You can also add/delete

406 other case splitting rules to/from the default simpset: every datatype

407 generates suitable rules `split_t_case' and `split_t_case_asm' (where

408 t is the name of the datatype).

410 * Classical reasoner / Simplifier combination: new force_tac (and

411 derivatives Force_tac, force) combines rewriting and classical

412 reasoning (and whatever other tools) similarly to auto_tac, but is

413 aimed to solve the given subgoal completely.

416 *** General ***

418 * new top-level commands `Goal' and `Goalw' that improve upon `goal'

419 and `goalw': the theory is no longer needed as an explicit argument -

420 the current theory context is used; assumptions are no longer returned

421 at the ML-level unless one of them starts with ==> or !!; it is

422 recommended to convert to these new commands using isatool fixgoal

423 (backup your sources first!);

425 * new top-level commands 'thm' and 'thms' for retrieving theorems from

426 the current theory context, and 'theory' to lookup stored theories;

428 * new theory section 'locale' for declaring constants, assumptions and

429 definitions that have local scope;

431 * new theory section 'nonterminals' for purely syntactic types;

433 * new theory section 'setup' for generic ML setup functions

434 (e.g. package initialization);

436 * the distribution now includes Isabelle icons: see

437 lib/logo/isabelle-{small,tiny}.xpm;

439 * isatool install - install binaries with absolute references to

440 ISABELLE_HOME/bin;

442 * isatool logo -- create instances of the Isabelle logo (as EPS);

444 * print mode 'emacs' reserved for Isamode;

446 * support multiple print (ast) translations per constant name;

448 * theorems involving oracles are now printed with a suffixed [!];

451 *** HOL ***

453 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');

455 * HOL/inductive package reorganized and improved: now supports mutual

456 definitions such as

458 inductive EVEN ODD

459 intrs

460 null "0 : EVEN"

461 oddI "n : EVEN ==> Suc n : ODD"

462 evenI "n : ODD ==> Suc n : EVEN"

464 new theorem list "elims" contains an elimination rule for each of the

465 recursive sets; inductive definitions now handle disjunctive premises

466 correctly (also ZF);

468 INCOMPATIBILITIES: requires Inductive as an ancestor; component

469 "mutual_induct" no longer exists - the induction rule is always

470 contained in "induct";

473 * HOL/datatype package re-implemented and greatly improved: now

474 supports mutually recursive datatypes such as

476 datatype

477 'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)

478 | SUM ('a aexp) ('a aexp)

479 | DIFF ('a aexp) ('a aexp)

480 | NUM 'a

481 and

482 'a bexp = LESS ('a aexp) ('a aexp)

483 | AND ('a bexp) ('a bexp)

484 | OR ('a bexp) ('a bexp)

486 as well as indirectly recursive datatypes such as

488 datatype

489 ('a, 'b) term = Var 'a

490 | App 'b ((('a, 'b) term) list)

492 The new tactic mutual_induct_tac [<var_1>, ..., <var_n>] i performs

493 induction on mutually / indirectly recursive datatypes.

495 Primrec equations are now stored in theory and can be accessed via

496 <function_name>.simps.

498 INCOMPATIBILITIES:

500 - Theories using datatypes must now have theory Datatype as an

501 ancestor.

502 - The specific <typename>.induct_tac no longer exists - use the

503 generic induct_tac instead.

504 - natE has been renamed to nat.exhaust - use exhaust_tac

505 instead of res_inst_tac ... natE. Note that the variable

506 names in nat.exhaust differ from the names in natE, this

507 may cause some "fragile" proofs to fail.

508 - The theorems split_<typename>_case and split_<typename>_case_asm

509 have been renamed to <typename>.split and <typename>.split_asm.

510 - Since default sorts of type variables are now handled correctly,

511 some datatype definitions may have to be annotated with explicit

512 sort constraints.

513 - Primrec definitions no longer require function name and type

514 of recursive argument.

516 Consider using isatool fixdatatype to adapt your theories and proof

517 scripts to the new package (backup your sources first!).

520 * HOL/record package: considerably improved implementation; now

521 includes concrete syntax for record types, terms, updates; theorems

522 for surjective pairing and splitting !!-bound record variables; proof

523 support is as follows:

525 1) standard conversions (selectors or updates applied to record

526 constructor terms) are part of the standard simpset;

528 2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are

529 made part of standard simpset and claset via addIffs;

531 3) a tactic for record field splitting (record_split_tac) is part of

532 the standard claset (addSWrapper);

534 To get a better idea about these rules you may retrieve them via

535 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is

536 the name of your record type.

538 The split tactic 3) conceptually simplifies by the following rule:

540 "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"

542 Thus any record variable that is bound by meta-all will automatically

543 blow up into some record constructor term, consequently the

544 simplifications of 1), 2) apply. Thus force_tac, auto_tac etc. shall

545 solve record problems automatically.

548 * reorganized the main HOL image: HOL/Integ and String loaded by

549 default; theory Main includes everything;

551 * automatic simplification of integer sums and comparisons, using cancellation;

553 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;

555 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;

557 * many new identities for unions, intersections, set difference, etc.;

559 * expand_if, expand_split, expand_sum_case and expand_nat_case are now

560 called split_if, split_split, split_sum_case and split_nat_case (to go

561 with add/delsplits);

563 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting

564 (?x::unit) = (); this is made part of the default simpset, which COULD

565 MAKE EXISTING PROOFS FAIL under rare circumstances (consider

566 'Delsimprocs [unit_eq_proc];' as last resort); also note that

567 unit_abs_eta_conv is added in order to counter the effect of

568 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by

569 %u.f();

571 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which

572 makes more sense);

574 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;

575 It and 'sym RS equals0D' are now in the default claset, giving automatic

576 disjointness reasoning but breaking a few old proofs.

578 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1

579 to 'converse' from 'inverse' (for compatibility with ZF and some

580 literature);

582 * HOL/recdef can now declare non-recursive functions, with {} supplied as

583 the well-founded relation;

585 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of

586 Compl A. The "Compl" syntax remains available as input syntax for this

587 release ONLY.

589 * HOL/Update: new theory of function updates:

590 f(a:=b) == %x. if x=a then b else f x

591 may also be iterated as in f(a:=b,c:=d,...);

593 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;

595 * HOL/List:

596 - new function list_update written xs[i:=v] that updates the i-th

597 list position. May also be iterated as in xs[i:=a,j:=b,...].

598 - new function `upt' written [i..j(] which generates the list

599 [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper

600 bound write [i..j], which is a shorthand for [i..j+1(].

601 - new lexicographic orderings and corresponding wellfoundedness theorems.

603 * HOL/Arith:

604 - removed 'pred' (predecessor) function;

605 - generalized some theorems about n-1;

606 - many new laws about "div" and "mod";

607 - new laws about greatest common divisors (see theory ex/Primes);

609 * HOL/Relation: renamed the relational operator r^-1 "converse"

610 instead of "inverse";

612 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness

613 of the multiset ordering;

615 * directory HOL/Real: a construction of the reals using Dedekind cuts

616 (not included by default);

618 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;

620 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted

621 programs, i.e. different program variables may have different types.

623 * calling (stac rew i) now fails if "rew" has no effect on the goal

624 [previously, this check worked only if the rewrite rule was unconditional]

625 Now rew can involve either definitions or equalities (either == or =).

628 *** ZF ***

630 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains

631 only the theorems proved on ZF.ML;

633 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;

634 It and 'sym RS equals0D' are now in the default claset, giving automatic

635 disjointness reasoning but breaking a few old proofs.

637 * ZF/Update: new theory of function updates

638 with default rewrite rule f(x:=y) ` z = if(z=x, y, f`z)

639 may also be iterated as in f(a:=b,c:=d,...);

641 * in let x=t in u(x), neither t nor u(x) has to be an FOL term.

643 * calling (stac rew i) now fails if "rew" has no effect on the goal

644 [previously, this check worked only if the rewrite rule was unconditional]

645 Now rew can involve either definitions or equalities (either == or =).

647 * case_tac provided for compatibility with HOL

648 (like the old excluded_middle_tac, but with subgoals swapped)

651 *** Internal programming interfaces ***

653 * Pure: several new basic modules made available for general use, see

654 also src/Pure/README;

656 * improved the theory data mechanism to support encapsulation (data

657 kind name replaced by private Object.kind, acting as authorization

658 key); new type-safe user interface via functor TheoryDataFun; generic

659 print_data function becomes basically useless;

661 * removed global_names compatibility flag -- all theory declarations

662 are qualified by default;

664 * module Pure/Syntax now offers quote / antiquote translation

665 functions (useful for Hoare logic etc. with implicit dependencies);

666 see HOL/ex/Antiquote for an example use;

668 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->

669 cterm -> thm;

671 * new tactical CHANGED_GOAL for checking that a tactic modifies a

672 subgoal;

674 * Display.print_goals function moved to Locale.print_goals;

676 * standard print function for goals supports current_goals_markers

677 variable for marking begin of proof, end of proof, start of goal; the

678 default is ("", "", ""); setting current_goals_markers := ("<proof>",

679 "</proof>", "<goal>") causes SGML like tagged proof state printing,

680 for example;

684 New in Isabelle98 (January 1998)

685 --------------------------------

687 *** Overview of INCOMPATIBILITIES (see below for more details) ***

689 * changed lexical syntax of terms / types: dots made part of long

690 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";

692 * simpset (and claset) reference variable replaced by functions

693 simpset / simpset_ref;

695 * no longer supports theory aliases (via merge) and non-trivial

696 implicit merge of thms' signatures;

698 * most internal names of constants changed due to qualified names;

700 * changed Pure/Sequence interface (see Pure/seq.ML);

703 *** General Changes ***

705 * hierachically structured name spaces (for consts, types, axms, thms

706 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of

707 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:

708 isatool fixdots ensures space after dots (e.g. "%x. x"); set

709 long_names for fully qualified output names; NOTE: ML programs

710 (special tactics, packages etc.) referring to internal names may have

711 to be adapted to cope with fully qualified names; in case of severe

712 backward campatibility problems try setting 'global_names' at compile

713 time to have enrything declared within a flat name space; one may also

714 fine tune name declarations in theories via the 'global' and 'local'

715 section;

717 * reimplemented the implicit simpset and claset using the new anytype

718 data filed in signatures; references simpset:simpset ref etc. are

719 replaced by functions simpset:unit->simpset and

720 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp

721 to patch your ML files accordingly;

723 * HTML output now includes theory graph data for display with Java

724 applet or isatool browser; data generated automatically via isatool

725 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);

727 * defs may now be conditional; improved rewrite_goals_tac to handle

728 conditional equations;

730 * defs now admits additional type arguments, using TYPE('a) syntax;

732 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always

733 creates a new theory node; implicit merge of thms' signatures is

734 restricted to 'trivial' ones; COMPATIBILITY: one may have to use

735 transfer:theory->thm->thm in (rare) cases;

737 * improved handling of draft signatures / theories; draft thms (and

738 ctyps, cterms) are automatically promoted to real ones;

740 * slightly changed interfaces for oracles: admit many per theory, named

741 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;

743 * print_goals: optional output of const types (set show_consts and

744 show_types);

746 * improved output of warnings (###) and errors (***);

748 * subgoal_tac displays a warning if the new subgoal has type variables;

750 * removed old README and Makefiles;

752 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;

754 * removed obsolete init_pps and init_database;

756 * deleted the obsolete tactical STATE, which was declared by

757 fun STATE tacfun st = tacfun st st;

759 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~

760 (which abbreviates $HOME);

762 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:

763 use isatool fixseq to adapt your ML programs (this works for fully

764 qualified references to the Sequence structure only!);

766 * use_thy no longer requires writable current directory; it always

767 reloads .ML *and* .thy file, if either one is out of date;

770 *** Classical Reasoner ***

772 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new

773 tactics that use classical reasoning to simplify a subgoal without

774 splitting it into several subgoals;

776 * Safe_tac: like safe_tac but uses the default claset;

779 *** Simplifier ***

781 * added simplification meta rules:

782 (asm_)(full_)simplify: simpset -> thm -> thm;

784 * simplifier.ML no longer part of Pure -- has to be loaded by object

785 logics (again);

787 * added prems argument to simplification procedures;

789 * HOL, FOL, ZF: added infix function `addsplits':

790 instead of `<simpset> setloop (split_tac <thms>)'

791 you can simply write `<simpset> addsplits <thms>'

794 *** Syntax ***

796 * TYPE('a) syntax for type reflection terms;

798 * no longer handles consts with name "" -- declare as 'syntax' instead;

800 * pretty printer: changed order of mixfix annotation preference (again!);

802 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;

805 *** HOL ***

807 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.

808 with `addloop' of the simplifier to faciliate case splitting in premises.

810 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;

812 * HOL/Auth: new protocol proofs including some for the Internet

813 protocol TLS;

815 * HOL/Map: new theory of `maps' a la VDM;

817 * HOL/simplifier: simplification procedures nat_cancel_sums for

818 cancelling out common nat summands from =, <, <= (in)equalities, or

819 differences; simplification procedures nat_cancel_factor for

820 cancelling common factor from =, <, <= (in)equalities over natural

821 sums; nat_cancel contains both kinds of procedures, it is installed by

822 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;

824 * HOL/simplifier: terms of the form

825 `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)' (or t=x)

826 are rewritten to

827 `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',

828 and those of the form

829 `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)' (or t=x)

830 are rewritten to

831 `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',

833 * HOL/datatype

834 Each datatype `t' now comes with a theorem `split_t_case' of the form

836 P(t_case f1 ... fn x) =

837 ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &

838 ...

839 (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))

840 )

842 and a theorem `split_t_case_asm' of the form

844 P(t_case f1 ... fn x) =

845 ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |

846 ...

847 (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))

848 )

849 which can be added to a simpset via `addsplits'. The existing theorems

850 expand_list_case and expand_option_case have been renamed to

851 split_list_case and split_option_case.

853 * HOL/Arithmetic:

854 - `pred n' is automatically converted to `n-1'.

855 Users are strongly encouraged not to use `pred' any longer,

856 because it will disappear altogether at some point.

857 - Users are strongly encouraged to write "0 < n" rather than

858 "n ~= 0". Theorems and proof tools have been modified towards this

859 `standard'.

861 * HOL/Lists:

862 the function "set_of_list" has been renamed "set" (and its theorems too);

863 the function "nth" now takes its arguments in the reverse order and

864 has acquired the infix notation "!" as in "xs!n".

866 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};

868 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its

869 specialist theorems (like UN1_I) are gone. Similarly for (INT x.B x);

871 * HOL/record: extensible records with schematic structural subtyping

872 (single inheritance); EXPERIMENTAL version demonstrating the encoding,

873 still lacks various theorems and concrete record syntax;

876 *** HOLCF ***

878 * removed "axioms" and "generated by" sections;

880 * replaced "ops" section by extended "consts" section, which is capable of

881 handling the continuous function space "->" directly;

883 * domain package:

884 . proves theorems immediately and stores them in the theory,

885 . creates hierachical name space,

886 . now uses normal mixfix annotations (instead of cinfix...),

887 . minor changes to some names and values (for consistency),

888 . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,

889 . separator between mutual domain defs: changed "," to "and",

890 . improved handling of sort constraints; now they have to

891 appear on the left-hand side of the equations only;

893 * fixed LAM <x,y,zs>.b syntax;

895 * added extended adm_tac to simplifier in HOLCF -- can now discharge

896 adm (%x. P (t x)), where P is chainfinite and t continuous;

899 *** FOL and ZF ***

901 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.

902 with `addloop' of the simplifier to faciliate case splitting in premises.

904 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as

905 in HOL, they strip ALL and --> from proved theorems;

909 New in Isabelle94-8 (May 1997)

910 ------------------------------

912 *** General Changes ***

914 * new utilities to build / run / maintain Isabelle etc. (in parts

915 still somewhat experimental); old Makefiles etc. still functional;

917 * new 'Isabelle System Manual';

919 * INSTALL text, together with ./configure and ./build scripts;

921 * reimplemented type inference for greater efficiency, better error

922 messages and clean internal interface;

924 * prlim command for dealing with lots of subgoals (an easier way of

925 setting goals_limit);

928 *** Syntax ***

930 * supports alternative (named) syntax tables (parser and pretty

931 printer); internal interface is provided by add_modesyntax(_i);

933 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to

934 be used in conjunction with the Isabelle symbol font; uses the

935 "symbols" syntax table;

937 * added token_translation interface (may translate name tokens in

938 arbitrary ways, dependent on their type (free, bound, tfree, ...) and

939 the current print_mode); IMPORTANT: user print translation functions

940 are responsible for marking newly introduced bounds

941 (Syntax.mark_boundT);

943 * token translations for modes "xterm" and "xterm_color" that display

944 names in bold, underline etc. or colors (which requires a color

945 version of xterm);

947 * infixes may now be declared with names independent of their syntax;

949 * added typed_print_translation (like print_translation, but may

950 access type of constant);

953 *** Classical Reasoner ***

955 Blast_tac: a new tactic! It is often more powerful than fast_tac, but has

956 some limitations. Blast_tac...

957 + ignores addss, addbefore, addafter; this restriction is intrinsic

958 + ignores elimination rules that don't have the correct format

959 (the conclusion MUST be a formula variable)

960 + ignores types, which can make HOL proofs fail

961 + rules must not require higher-order unification, e.g. apply_type in ZF

962 [message "Function Var's argument not a bound variable" relates to this]

963 + its proof strategy is more general but can actually be slower

965 * substitution with equality assumptions no longer permutes other

966 assumptions;

968 * minor changes in semantics of addafter (now called addaltern); renamed

969 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper

970 (and access functions for it);

972 * improved combination of classical reasoner and simplifier:

973 + functions for handling clasimpsets

974 + improvement of addss: now the simplifier is called _after_ the

975 safe steps.

976 + safe variant of addss called addSss: uses safe simplifications

977 _during_ the safe steps. It is more complete as it allows multiple

978 instantiations of unknowns (e.g. with slow_tac).

980 *** Simplifier ***

982 * added interface for simplification procedures (functions that

983 produce *proven* rewrite rules on the fly, depending on current

984 redex);

986 * ordering on terms as parameter (used for ordered rewriting);

988 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;

990 * the solver is now split into a safe and an unsafe part.

991 This should be invisible for the normal user, except that the

992 functions setsolver and addsolver have been renamed to setSolver and

993 addSolver; added safe_asm_full_simp_tac;

996 *** HOL ***

998 * a generic induction tactic `induct_tac' which works for all datatypes and

999 also for type `nat';

1001 * a generic case distinction tactic `exhaust_tac' which works for all

1002 datatypes and also for type `nat';

1004 * each datatype comes with a function `size';

1006 * patterns in case expressions allow tuple patterns as arguments to

1007 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';

1009 * primrec now also works with type nat;

1011 * recdef: a new declaration form, allows general recursive functions to be

1012 defined in theory files. See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.

1014 * the constant for negation has been renamed from "not" to "Not" to

1015 harmonize with FOL, ZF, LK, etc.;

1017 * HOL/ex/LFilter theory of a corecursive "filter" functional for

1018 infinite lists;

1020 * HOL/Modelcheck demonstrates invocation of model checker oracle;

1022 * HOL/ex/Ring.thy declares cring_simp, which solves equational

1023 problems in commutative rings, using axiomatic type classes for + and *;

1025 * more examples in HOL/MiniML and HOL/Auth;

1027 * more default rewrite rules for quantifiers, union/intersection;

1029 * a new constant `arbitrary == @x.False';

1031 * HOLCF/IOA replaces old HOL/IOA;

1033 * HOLCF changes: derived all rules and arities

1034 + axiomatic type classes instead of classes

1035 + typedef instead of faking type definitions

1036 + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.

1037 + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po

1038 + eliminated the types void, one, tr

1039 + use unit lift and bool lift (with translations) instead of one and tr

1040 + eliminated blift from Lift3.thy (use Def instead of blift)

1041 all eliminated rules are derived as theorems --> no visible changes ;

1044 *** ZF ***

1046 * ZF now has Fast_tac, Simp_tac and Auto_tac. Union_iff is a now a default

1047 rewrite rule; this may affect some proofs. eq_cs is gone but can be put back

1048 as ZF_cs addSIs [equalityI];

1052 New in Isabelle94-7 (November 96)

1053 ---------------------------------

1055 * allowing negative levels (as offsets) in prlev and choplev;

1057 * super-linear speedup for large simplifications;

1059 * FOL, ZF and HOL now use miniscoping: rewriting pushes

1060 quantifications in as far as possible (COULD MAKE EXISTING PROOFS

1061 FAIL); can suppress it using the command Delsimps (ex_simps @

1062 all_simps); De Morgan laws are also now included, by default;

1064 * improved printing of ==> : ~:

1066 * new object-logic "Sequents" adds linear logic, while replacing LK

1067 and Modal (thanks to Sara Kalvala);

1069 * HOL/Auth: correctness proofs for authentication protocols;

1071 * HOL: new auto_tac combines rewriting and classical reasoning (many

1072 examples on HOL/Auth);

1074 * HOL: new command AddIffs for declaring theorems of the form P=Q to

1075 the rewriter and classical reasoner simultaneously;

1077 * function uresult no longer returns theorems in "standard" format;

1078 regain previous version by: val uresult = standard o uresult;

1082 New in Isabelle94-6

1083 -------------------

1085 * oracles -- these establish an interface between Isabelle and trusted

1086 external reasoners, which may deliver results as theorems;

1088 * proof objects (in particular record all uses of oracles);

1090 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;

1092 * "constdefs" section in theory files;

1094 * "primrec" section (HOL) no longer requires names;

1096 * internal type "tactic" now simply "thm -> thm Sequence.seq";

1100 New in Isabelle94-5

1101 -------------------

1103 * reduced space requirements;

1105 * automatic HTML generation from theories;

1107 * theory files no longer require "..." (quotes) around most types;

1109 * new examples, including two proofs of the Church-Rosser theorem;

1111 * non-curried (1994) version of HOL is no longer distributed;

1115 New in Isabelle94-4

1116 -------------------

1118 * greatly reduced space requirements;

1120 * theory files (.thy) no longer require \...\ escapes at line breaks;

1122 * searchable theorem database (see the section "Retrieving theorems" on

1123 page 8 of the Reference Manual);

1125 * new examples, including Grabczewski's monumental case study of the

1126 Axiom of Choice;

1128 * The previous version of HOL renamed to Old_HOL;

1130 * The new version of HOL (previously called CHOL) uses a curried syntax

1131 for functions. Application looks like f a b instead of f(a,b);

1133 * Mutually recursive inductive definitions finally work in HOL;

1135 * In ZF, pattern-matching on tuples is now available in all abstractions and

1136 translates to the operator "split";

1140 New in Isabelle94-3

1141 -------------------

1143 * new infix operator, addss, allowing the classical reasoner to

1144 perform simplification at each step of its search. Example:

1145 fast_tac (cs addss ss)

1147 * a new logic, CHOL, the same as HOL, but with a curried syntax

1148 for functions. Application looks like f a b instead of f(a,b). Also pairs

1149 look like (a,b) instead of <a,b>;

1151 * PLEASE NOTE: CHOL will eventually replace HOL!

1153 * In CHOL, pattern-matching on tuples is now available in all abstractions.

1154 It translates to the operator "split". A new theory of integers is available;

1156 * In ZF, integer numerals now denote two's-complement binary integers.

1157 Arithmetic operations can be performed by rewriting. See ZF/ex/Bin.ML;

1159 * Many new examples: I/O automata, Church-Rosser theorem, equivalents

1160 of the Axiom of Choice;

1164 New in Isabelle94-2

1165 -------------------

1167 * Significantly faster resolution;

1169 * the different sections in a .thy file can now be mixed and repeated

1170 freely;

1172 * Database of theorems for FOL, HOL and ZF. New

1173 commands including qed, qed_goal and bind_thm store theorems in the database.

1175 * Simple database queries: return a named theorem (get_thm) or all theorems of

1176 a given theory (thms_of), or find out what theory a theorem was proved in

1177 (theory_of_thm);

1179 * Bugs fixed in the inductive definition and datatype packages;

1181 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs

1182 and HOL_dup_cs obsolete;

1184 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1

1185 have been removed;

1187 * Simpler definition of function space in ZF;

1189 * new results about cardinal and ordinal arithmetic in ZF;

1191 * 'subtype' facility in HOL for introducing new types as subsets of existing

1192 types;

1195 $Id$