wenzelm@5363: Isabelle NEWS -- history user-relevant changes wenzelm@5363: ============================================== wenzelm@2553: wenzelm@4981: New in this Isabelle version wenzelm@4981: ---------------------------- oheimb@4649: paulson@5931: *** Overview of INCOMPATIBILITIES (see below for more details) *** paulson@5931: paulson@5931: * HOL: Removed the obsolete syntax "Compl A"; use -A for set complement paulson@5931: paulson@6174: * HOL: the predicate "inj" is now defined by translation to "inj_on" paulson@6174: paulson@6064: * ZF: The con_defs part of an inductive definition may no longer refer to paulson@6064: constants declared in the same theory; nipkow@6057: paulson@6141: * HOL, ZF: the function mk_cases, generated by the inductive definition paulson@6141: package, has lost an argument. To simplify its result, it uses the default paulson@6141: simpset instead of a supplied list of theorems. paulson@6141: nipkow@6057: paulson@6069: *** Proof tools *** paulson@6069: paulson@6069: * Provers/Arith/fast_lin_arith.ML contains a functor for creating a decision paulson@6069: procedure for linear arithmetic. Currently it is used for types `nat' and paulson@6069: `int' in HOL (see below) but can, should and will be instantiated for other paulson@6069: types and logics as well. paulson@6069: paulson@6069: paulson@6014: *** General *** paulson@6014: paulson@6014: * in locales, the "assumes" and "defines" parts may be omitted if empty; wenzelm@5973: oheimb@6028: * new print_mode "xsymbols" for extended symbol support wenzelm@6259: (e.g. genuiene long arrows); wenzelm@6259: wenzelm@6259: * path element specification '~~' refers to '$ISABELLE_HOME'; wenzelm@6259: oheimb@6028: nipkow@6057: *** HOL *** nipkow@6057: nipkow@6063: * There are now decision procedures for linear arithmetic over nat and int: nipkow@6131: nipkow@6131: 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=', `+', nipkow@6157: `-', `Suc', `min', `max' and numerical constants; other subterms are treated nipkow@6157: as atomic; subformulae not involving type `nat' or `int' are ignored; nipkow@6157: quantified subformulae are ignored unless they are positive universal or nipkow@6157: negative existential. The tactic has to be invoked by hand and can be a nipkow@6157: little bit slow. In particular, the running time is exponential in the number nipkow@6157: of occurrences of `min' and `max', and `-' on `nat'. nipkow@6131: nipkow@6131: 2. fast_arith_tac is a cut-down version of arith_tac: it only takes (negated) nipkow@6131: (in)equalities among the premises and the conclusion into account (i.e. no nipkow@6157: compound formulae) and does not know about `min' and `max', and `-' on nipkow@6157: `nat'. It is fast and is used automatically by the simplifier. nipkow@6131: nipkow@6063: NB: At the moment, these decision procedures do not cope with mixed nat/int nipkow@6131: formulae where the two parts interact, such as `m < n ==> int(m) < int(n)'. oheimb@6028: wenzelm@5973: *** Internal programming interfaces *** wenzelm@5781: wenzelm@5781: * tuned current_goals_markers semantics: begin / end goal avoids wenzelm@5781: printing empty lines; wenzelm@5781: wenzelm@5973: * removed prs and prs_fn hook, which was broken because it did not wenzelm@5973: include \n in its semantics, forcing writeln to add one wenzelm@5973: uncoditionally; replaced prs_fn by writeln_fn; consider std_output: wenzelm@5973: string -> unit if you really want to output text without newline; wenzelm@5973: wenzelm@5781: paulson@6064: *** ZF *** paulson@6064: paulson@6064: * new primrec section allows primitive recursive functions to be given paulson@6070: directly (as in HOL) over datatypes and the natural numbers; paulson@6064: paulson@6064: * new tactics induct_tac and exhaust_tac for induction (or case analysis) paulson@6070: over datatypes and the natural numbers; paulson@6064: paulson@6064: * the datatype declaration of type T now defines the recursor T_rec; paulson@6064: paulson@6141: * simplification automatically does freeness reasoning for datatype paulson@6141: constructors; paulson@6141: paulson@6155: * automatic type-inference, with AddTCs command to insert new type-checking paulson@6155: rules; paulson@6155: paulson@6155: * datatype introduction rules are now added as Safe Introduction rules to paulson@6155: the claset; paulson@6155: paulson@6069: * The syntax "if P then x else y" is now available in addition to if(P,x,y). paulson@6069: paulson@6064: wenzelm@5781: New in Isabelle98-1 (October 1998) wenzelm@5781: ---------------------------------- wenzelm@5781: wenzelm@5127: *** Overview of INCOMPATIBILITIES (see below for more details) *** wenzelm@4842: wenzelm@5726: * several changes of automated proof tools; wenzelm@5373: wenzelm@5726: * HOL: major changes to the inductive and datatype packages, including wenzelm@5726: some minor incompatibilities of theory syntax; berghofe@5214: wenzelm@5726: * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now wenzelm@5217: called `inj_on'; paulson@5160: nipkow@5275: * HOL: removed duplicate thms in Arith: nipkow@5275: less_imp_add_less should be replaced by trans_less_add1 nipkow@5275: le_imp_add_le should be replaced by trans_le_add1 paulson@5160: wenzelm@5726: * HOL: unary minus is now overloaded (new type constraints may be wenzelm@5726: required); paulson@5490: wenzelm@5726: * HOL and ZF: unary minus for integers is now #- instead of #~. In wenzelm@5726: ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is wenzelm@5726: now taken as an integer constant. paulson@5541: wenzelm@5726: * Pure: ML function 'theory_of' renamed to 'theory'; wenzelm@5397: wenzelm@5363: wenzelm@5127: *** Proof tools *** nipkow@4880: nipkow@5657: * Simplifier: nipkow@5657: 1. Asm_full_simp_tac is now more aggressive. nipkow@5657: 1. It will sometimes reorient premises if that increases their power to nipkow@5657: simplify. nipkow@5657: 2. It does no longer proceed strictly from left to right but may also nipkow@5657: rotate premises to achieve further simplification. nipkow@5657: For compatibility reasons there is now Asm_lr_simp_tac which is like the nipkow@5657: old Asm_full_simp_tac in that it does not rotate premises. nipkow@5657: 2. The simplifier now knows a little bit about nat-arithmetic. nipkow@4880: wenzelm@5127: * Classical reasoner: wrapper mechanism for the classical reasoner now wenzelm@5127: allows for selected deletion of wrappers, by introduction of names for wenzelm@5127: wrapper functionals. This implies that addbefore, addSbefore, wenzelm@5127: addaltern, and addSaltern now take a pair (name, tactic) as argument, wenzelm@5127: and that adding two tactics with the same name overwrites the first wenzelm@5127: one (emitting a warning). oheimb@4824: type wrapper = (int -> tactic) -> (int -> tactic) oheimb@4649: setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by oheimb@4824: addWrapper, addSWrapper: claset * (string * wrapper) -> claset oheimb@4824: delWrapper, delSWrapper: claset * string -> claset oheimb@4649: getWrapper is renamed to appWrappers, getSWrapper to appSWrappers; oheimb@4649: wenzelm@5705: * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE wenzelm@5726: semantics; addbefore now affects only the unsafe part of step_tac wenzelm@5726: etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY wenzelm@5726: FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac wenzelm@5726: by Force_tac; oheimb@5524: wenzelm@5726: * Classical reasoner: setwrapper to setWrapper and compwrapper to wenzelm@5726: compWrapper; added safe wrapper (and access functions for it); oheimb@5524: wenzelm@5127: * HOL/split_all_tac is now much faster and fails if there is nothing wenzelm@5726: to split. Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order wenzelm@5726: and the names of the automatically generated variables have changed. wenzelm@5726: split_all_tac has moved within claset() from unsafe wrappers to safe wenzelm@5726: wrappers, which means that !!-bound variables are split much more wenzelm@5726: aggressively, and safe_tac and clarify_tac now split such variables. wenzelm@5726: If this splitting is not appropriate, use delSWrapper "split_all_tac". wenzelm@5726: Note: the same holds for record_split_tac, which does the job of wenzelm@5726: split_all_tac for record fields. wenzelm@5127: wenzelm@5726: * HOL/Simplifier: Rewrite rules for case distinctions can now be added wenzelm@5726: permanently to the default simpset using Addsplits just like wenzelm@5726: Addsimps. They can be removed via Delsplits just like wenzelm@5726: Delsimps. Lower-case versions are also available. wenzelm@5127: wenzelm@5726: * HOL/Simplifier: The rule split_if is now part of the default wenzelm@5726: simpset. This means that the simplifier will eliminate all occurrences wenzelm@5726: of if-then-else in the conclusion of a goal. To prevent this, you can wenzelm@5726: either remove split_if completely from the default simpset by wenzelm@5726: `Delsplits [split_if]' or remove it in a specific call of the wenzelm@5726: simplifier using `... delsplits [split_if]'. You can also add/delete wenzelm@5726: other case splitting rules to/from the default simpset: every datatype wenzelm@5726: generates suitable rules `split_t_case' and `split_t_case_asm' (where wenzelm@5726: t is the name of the datatype). wenzelm@5127: wenzelm@5726: * Classical reasoner / Simplifier combination: new force_tac (and wenzelm@5127: derivatives Force_tac, force) combines rewriting and classical wenzelm@5127: reasoning (and whatever other tools) similarly to auto_tac, but is wenzelm@5726: aimed to solve the given subgoal completely. wenzelm@5127: wenzelm@5127: wenzelm@5127: *** General *** wenzelm@5127: wenzelm@5217: * new top-level commands `Goal' and `Goalw' that improve upon `goal' wenzelm@5127: and `goalw': the theory is no longer needed as an explicit argument - wenzelm@5127: the current theory context is used; assumptions are no longer returned wenzelm@5127: at the ML-level unless one of them starts with ==> or !!; it is wenzelm@5217: recommended to convert to these new commands using isatool fixgoal wenzelm@5217: (backup your sources first!); wenzelm@4842: wenzelm@5217: * new top-level commands 'thm' and 'thms' for retrieving theorems from wenzelm@5207: the current theory context, and 'theory' to lookup stored theories; paulson@4806: paulson@5722: * new theory section 'locale' for declaring constants, assumptions and paulson@5722: definitions that have local scope; paulson@5722: wenzelm@5127: * new theory section 'nonterminals' for purely syntactic types; wenzelm@4858: wenzelm@5127: * new theory section 'setup' for generic ML setup functions wenzelm@5127: (e.g. package initialization); wenzelm@4869: wenzelm@5131: * the distribution now includes Isabelle icons: see wenzelm@5131: lib/logo/isabelle-{small,tiny}.xpm; wenzelm@5131: wenzelm@5363: * isatool install - install binaries with absolute references to wenzelm@5363: ISABELLE_HOME/bin; wenzelm@5363: wenzelm@5572: * isatool logo -- create instances of the Isabelle logo (as EPS); wenzelm@5572: wenzelm@5407: * print mode 'emacs' reserved for Isamode; wenzelm@5407: wenzelm@5726: * support multiple print (ast) translations per constant name; wenzelm@5726: wenzelm@4711: paulson@4661: *** HOL *** paulson@4661: wenzelm@5710: * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial'); nipkow@5709: wenzelm@5217: * HOL/inductive package reorganized and improved: now supports mutual paulson@5267: definitions such as wenzelm@5217: wenzelm@5217: inductive EVEN ODD wenzelm@5217: intrs wenzelm@5217: null "0 : EVEN" wenzelm@5217: oddI "n : EVEN ==> Suc n : ODD" wenzelm@5217: evenI "n : ODD ==> Suc n : EVEN" wenzelm@5217: wenzelm@5217: new theorem list "elims" contains an elimination rule for each of the wenzelm@5217: recursive sets; inductive definitions now handle disjunctive premises wenzelm@5217: correctly (also ZF); berghofe@5214: wenzelm@5217: INCOMPATIBILITIES: requires Inductive as an ancestor; component wenzelm@5217: "mutual_induct" no longer exists - the induction rule is always wenzelm@5217: contained in "induct"; wenzelm@5217: wenzelm@5217: wenzelm@5217: * HOL/datatype package re-implemented and greatly improved: now paulson@5267: supports mutually recursive datatypes such as wenzelm@5217: wenzelm@5217: datatype wenzelm@5217: 'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp) wenzelm@5217: | SUM ('a aexp) ('a aexp) wenzelm@5217: | DIFF ('a aexp) ('a aexp) wenzelm@5217: | NUM 'a wenzelm@5217: and wenzelm@5217: 'a bexp = LESS ('a aexp) ('a aexp) wenzelm@5217: | AND ('a bexp) ('a bexp) wenzelm@5217: | OR ('a bexp) ('a bexp) wenzelm@5217: paulson@5267: as well as indirectly recursive datatypes such as berghofe@5214: wenzelm@5217: datatype wenzelm@5217: ('a, 'b) term = Var 'a wenzelm@5217: | App 'b ((('a, 'b) term) list) berghofe@5214: wenzelm@5217: The new tactic mutual_induct_tac [, ..., ] i performs wenzelm@5217: induction on mutually / indirectly recursive datatypes. wenzelm@5217: wenzelm@5217: Primrec equations are now stored in theory and can be accessed via wenzelm@5217: .simps. wenzelm@5217: wenzelm@5217: INCOMPATIBILITIES: berghofe@5214: wenzelm@5217: - Theories using datatypes must now have theory Datatype as an wenzelm@5217: ancestor. wenzelm@5217: - The specific .induct_tac no longer exists - use the wenzelm@5217: generic induct_tac instead. berghofe@5226: - natE has been renamed to nat.exhaust - use exhaust_tac wenzelm@5217: instead of res_inst_tac ... natE. Note that the variable berghofe@5226: names in nat.exhaust differ from the names in natE, this wenzelm@5217: may cause some "fragile" proofs to fail. wenzelm@5217: - The theorems split__case and split__case_asm wenzelm@5217: have been renamed to .split and .split_asm. wenzelm@5217: - Since default sorts of type variables are now handled correctly, wenzelm@5217: some datatype definitions may have to be annotated with explicit wenzelm@5217: sort constraints. wenzelm@5217: - Primrec definitions no longer require function name and type wenzelm@5217: of recursive argument. berghofe@5214: wenzelm@5217: Consider using isatool fixdatatype to adapt your theories and proof wenzelm@5217: scripts to the new package (backup your sources first!). wenzelm@5217: wenzelm@5217: wenzelm@5726: * HOL/record package: considerably improved implementation; now wenzelm@5726: includes concrete syntax for record types, terms, updates; theorems wenzelm@5726: for surjective pairing and splitting !!-bound record variables; proof wenzelm@5726: support is as follows: wenzelm@5726: wenzelm@5726: 1) standard conversions (selectors or updates applied to record wenzelm@5726: constructor terms) are part of the standard simpset; wenzelm@5726: wenzelm@5726: 2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are wenzelm@5726: made part of standard simpset and claset via addIffs; wenzelm@5726: wenzelm@5726: 3) a tactic for record field splitting (record_split_tac) is part of wenzelm@5726: the standard claset (addSWrapper); wenzelm@5726: wenzelm@5726: To get a better idea about these rules you may retrieve them via wenzelm@5726: something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is wenzelm@5726: the name of your record type. wenzelm@5726: wenzelm@5726: The split tactic 3) conceptually simplifies by the following rule: wenzelm@5726: wenzelm@5726: "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))" wenzelm@5726: wenzelm@5726: Thus any record variable that is bound by meta-all will automatically wenzelm@5726: blow up into some record constructor term, consequently the wenzelm@5726: simplifications of 1), 2) apply. Thus force_tac, auto_tac etc. shall wenzelm@5726: solve record problems automatically. wenzelm@5726: berghofe@5214: wenzelm@5125: * reorganized the main HOL image: HOL/Integ and String loaded by wenzelm@5125: default; theory Main includes everything; wenzelm@5125: paulson@5650: * automatic simplification of integer sums and comparisons, using cancellation; paulson@5650: oheimb@5526: * added option_map_eq_Some and not_Some_eq to the default simpset and claset; wenzelm@5127: wenzelm@5127: * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset; wenzelm@5127: wenzelm@5127: * many new identities for unions, intersections, set difference, etc.; wenzelm@5127: wenzelm@5127: * expand_if, expand_split, expand_sum_case and expand_nat_case are now wenzelm@5127: called split_if, split_split, split_sum_case and split_nat_case (to go wenzelm@5127: with add/delsplits); wenzelm@5125: wenzelm@5127: * HOL/Prod introduces simplification procedure unit_eq_proc rewriting wenzelm@5127: (?x::unit) = (); this is made part of the default simpset, which COULD wenzelm@5127: MAKE EXISTING PROOFS FAIL under rare circumstances (consider wenzelm@5207: 'Delsimprocs [unit_eq_proc];' as last resort); also note that wenzelm@5207: unit_abs_eta_conv is added in order to counter the effect of wenzelm@5207: unit_eq_proc on (%u::unit. f u), replacing it by f rather than by wenzelm@5207: %u.f(); wenzelm@5125: wenzelm@5217: * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which wenzelm@5217: makes more sense); berghofe@5109: paulson@5475: * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule; paulson@5475: It and 'sym RS equals0D' are now in the default claset, giving automatic paulson@5475: disjointness reasoning but breaking a few old proofs. paulson@5267: wenzelm@5217: * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1 wenzelm@5217: to 'converse' from 'inverse' (for compatibility with ZF and some wenzelm@5217: literature); wenzelm@5085: wenzelm@5127: * HOL/recdef can now declare non-recursive functions, with {} supplied as wenzelm@5127: the well-founded relation; oheimb@4838: paulson@5490: * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of paulson@5490: Compl A. The "Compl" syntax remains available as input syntax for this paulson@5490: release ONLY. paulson@5490: wenzelm@5127: * HOL/Update: new theory of function updates: wenzelm@5127: f(a:=b) == %x. if x=a then b else f x wenzelm@5127: may also be iterated as in f(a:=b,c:=d,...); nipkow@5077: wenzelm@5127: * HOL/Vimage: new theory for inverse image of a function, syntax f-``B; paulson@4899: nipkow@5282: * HOL/List: nipkow@5282: - new function list_update written xs[i:=v] that updates the i-th nipkow@5282: list position. May also be iterated as in xs[i:=a,j:=b,...]. nipkow@5428: - new function `upt' written [i..j(] which generates the list nipkow@5428: [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper nipkow@5428: bound write [i..j], which is a shorthand for [i..j+1(]. nipkow@5282: - new lexicographic orderings and corresponding wellfoundedness theorems. paulson@4779: wenzelm@5127: * HOL/Arith: wenzelm@5127: - removed 'pred' (predecessor) function; wenzelm@5127: - generalized some theorems about n-1; wenzelm@5127: - many new laws about "div" and "mod"; wenzelm@5127: - new laws about greatest common divisors (see theory ex/Primes); oheimb@4766: wenzelm@5127: * HOL/Relation: renamed the relational operator r^-1 "converse" wenzelm@4842: instead of "inverse"; wenzelm@4711: nipkow@5651: * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness nipkow@5651: of the multiset ordering; nipkow@5651: wenzelm@5127: * directory HOL/Real: a construction of the reals using Dedekind cuts nipkow@5651: (not included by default); nipkow@4835: wenzelm@5127: * directory HOL/UNITY: Chandy and Misra's UNITY formalism; wenzelm@4711: nipkow@5651: * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted nipkow@5651: programs, i.e. different program variables may have different types. nipkow@5651: paulson@5142: * calling (stac rew i) now fails if "rew" has no effect on the goal paulson@5142: [previously, this check worked only if the rewrite rule was unconditional] paulson@5308: Now rew can involve either definitions or equalities (either == or =). wenzelm@5002: wenzelm@5363: paulson@4879: *** ZF *** paulson@4879: paulson@5332: * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains paulson@5332: only the theorems proved on ZF.ML; paulson@5160: paulson@5475: * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule; paulson@5475: It and 'sym RS equals0D' are now in the default claset, giving automatic paulson@5475: disjointness reasoning but breaking a few old proofs. paulson@5267: paulson@5160: * ZF/Update: new theory of function updates paulson@5160: with default rewrite rule f(x:=y) ` z = if(z=x, y, f`z) paulson@5160: may also be iterated as in f(a:=b,c:=d,...); paulson@5160: paulson@4879: * in let x=t in u(x), neither t nor u(x) has to be an FOL term. oheimb@4649: paulson@5142: * calling (stac rew i) now fails if "rew" has no effect on the goal paulson@5142: [previously, this check worked only if the rewrite rule was unconditional] paulson@5308: Now rew can involve either definitions or equalities (either == or =). paulson@5142: paulson@5160: * case_tac provided for compatibility with HOL paulson@5160: (like the old excluded_middle_tac, but with subgoals swapped) paulson@5160: wenzelm@4842: wenzelm@5127: *** Internal programming interfaces *** wenzelm@5002: wenzelm@5251: * Pure: several new basic modules made available for general use, see wenzelm@5251: also src/Pure/README; wenzelm@5207: wenzelm@5008: * improved the theory data mechanism to support encapsulation (data wenzelm@5008: kind name replaced by private Object.kind, acting as authorization wenzelm@5373: key); new type-safe user interface via functor TheoryDataFun; generic wenzelm@5373: print_data function becomes basically useless; wenzelm@5002: wenzelm@5251: * removed global_names compatibility flag -- all theory declarations wenzelm@5251: are qualified by default; wenzelm@5251: wenzelm@5085: * module Pure/Syntax now offers quote / antiquote translation wenzelm@5085: functions (useful for Hoare logic etc. with implicit dependencies); wenzelm@5373: see HOL/ex/Antiquote for an example use; wenzelm@5085: wenzelm@5127: * Simplifier now offers conversions (asm_)(full_)rewrite: simpset -> wenzelm@5127: cterm -> thm; wenzelm@5127: wenzelm@5207: * new tactical CHANGED_GOAL for checking that a tactic modifies a wenzelm@5207: subgoal; paulson@5142: wenzelm@5251: * Display.print_goals function moved to Locale.print_goals; wenzelm@5251: wenzelm@5731: * standard print function for goals supports current_goals_markers wenzelm@5731: variable for marking begin of proof, end of proof, start of goal; the wenzelm@5731: default is ("", "", ""); setting current_goals_markers := ("", wenzelm@5731: "", "") causes SGML like tagged proof state printing, wenzelm@5731: for example; wenzelm@5731: wenzelm@5002: wenzelm@5002: wenzelm@4410: New in Isabelle98 (January 1998) wenzelm@4410: -------------------------------- wenzelm@4410: wenzelm@4410: *** Overview of INCOMPATIBILITIES (see below for more details) *** wenzelm@4410: wenzelm@4410: * changed lexical syntax of terms / types: dots made part of long wenzelm@4410: identifiers, e.g. "%x.x" no longer possible, should be "%x. x"; wenzelm@4410: wenzelm@4410: * simpset (and claset) reference variable replaced by functions wenzelm@4410: simpset / simpset_ref; wenzelm@4410: wenzelm@4410: * no longer supports theory aliases (via merge) and non-trivial wenzelm@4410: implicit merge of thms' signatures; wenzelm@4410: wenzelm@4410: * most internal names of constants changed due to qualified names; wenzelm@4410: wenzelm@4410: * changed Pure/Sequence interface (see Pure/seq.ML); wenzelm@4410: wenzelm@3454: paulson@3715: *** General Changes *** paulson@3715: wenzelm@4174: * hierachically structured name spaces (for consts, types, axms, thms wenzelm@3943: etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of wenzelm@4108: old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY: wenzelm@4108: isatool fixdots ensures space after dots (e.g. "%x. x"); set wenzelm@4174: long_names for fully qualified output names; NOTE: ML programs wenzelm@4174: (special tactics, packages etc.) referring to internal names may have wenzelm@4174: to be adapted to cope with fully qualified names; in case of severe wenzelm@4174: backward campatibility problems try setting 'global_names' at compile wenzelm@4174: time to have enrything declared within a flat name space; one may also wenzelm@4174: fine tune name declarations in theories via the 'global' and 'local' wenzelm@4174: section; wenzelm@4108: wenzelm@4108: * reimplemented the implicit simpset and claset using the new anytype wenzelm@4108: data filed in signatures; references simpset:simpset ref etc. are wenzelm@4108: replaced by functions simpset:unit->simpset and wenzelm@4108: simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp wenzelm@4108: to patch your ML files accordingly; paulson@3856: wenzelm@3857: * HTML output now includes theory graph data for display with Java wenzelm@3857: applet or isatool browser; data generated automatically via isatool wenzelm@3901: usedir (see -i option, ISABELLE_USEDIR_OPTIONS); wenzelm@3857: paulson@3856: * defs may now be conditional; improved rewrite_goals_tac to handle paulson@3856: conditional equations; paulson@3856: wenzelm@4174: * defs now admits additional type arguments, using TYPE('a) syntax; wenzelm@4174: wenzelm@3901: * theory aliases via merge (e.g. M=A+B+C) no longer supported, always wenzelm@3901: creates a new theory node; implicit merge of thms' signatures is wenzelm@4112: restricted to 'trivial' ones; COMPATIBILITY: one may have to use wenzelm@3901: transfer:theory->thm->thm in (rare) cases; wenzelm@3901: wenzelm@3968: * improved handling of draft signatures / theories; draft thms (and wenzelm@3968: ctyps, cterms) are automatically promoted to real ones; wenzelm@3968: wenzelm@3901: * slightly changed interfaces for oracles: admit many per theory, named wenzelm@3901: (e.g. oracle foo = mlfun), additional name argument for invoke_oracle; wenzelm@3901: wenzelm@3901: * print_goals: optional output of const types (set show_consts and wenzelm@3901: show_types); wenzelm@3851: wenzelm@4388: * improved output of warnings (###) and errors (***); wenzelm@3697: paulson@4178: * subgoal_tac displays a warning if the new subgoal has type variables; paulson@4178: paulson@3715: * removed old README and Makefiles; wenzelm@3697: paulson@3856: * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn; wenzelm@3670: paulson@3715: * removed obsolete init_pps and init_database; paulson@3715: paulson@3715: * deleted the obsolete tactical STATE, which was declared by paulson@3715: fun STATE tacfun st = tacfun st st; paulson@3715: wenzelm@4388: * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~ wenzelm@4388: (which abbreviates $HOME); wenzelm@4269: wenzelm@4269: * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY: wenzelm@4269: use isatool fixseq to adapt your ML programs (this works for fully wenzelm@4269: qualified references to the Sequence structure only!); wenzelm@4269: wenzelm@4381: * use_thy no longer requires writable current directory; it always wenzelm@4381: reloads .ML *and* .thy file, if either one is out of date; wenzelm@4269: paulson@3715: paulson@3715: *** Classical Reasoner *** paulson@3715: wenzelm@3744: * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new wenzelm@3744: tactics that use classical reasoning to simplify a subgoal without wenzelm@3744: splitting it into several subgoals; paulson@3715: paulson@3719: * Safe_tac: like safe_tac but uses the default claset; paulson@3719: paulson@3715: paulson@3715: *** Simplifier *** paulson@3715: paulson@3715: * added simplification meta rules: paulson@3715: (asm_)(full_)simplify: simpset -> thm -> thm; paulson@3715: paulson@3715: * simplifier.ML no longer part of Pure -- has to be loaded by object paulson@3715: logics (again); paulson@3715: paulson@3715: * added prems argument to simplification procedures; paulson@3715: paulson@4325: * HOL, FOL, ZF: added infix function `addsplits': paulson@4325: instead of ` setloop (split_tac )' paulson@4325: you can simply write ` addsplits ' paulson@4325: paulson@3715: paulson@3715: *** Syntax *** paulson@3715: wenzelm@4174: * TYPE('a) syntax for type reflection terms; wenzelm@4174: wenzelm@3985: * no longer handles consts with name "" -- declare as 'syntax' instead; paulson@3856: paulson@3856: * pretty printer: changed order of mixfix annotation preference (again!); wenzelm@3846: paulson@3715: * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories; paulson@3715: paulson@3715: paulson@3715: *** HOL *** paulson@3715: wenzelm@5726: * HOL: there is a new splitter `split_asm_tac' that can be used e.g. oheimb@4189: with `addloop' of the simplifier to faciliate case splitting in premises. oheimb@4189: nipkow@4035: * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions; wenzelm@3985: wenzelm@3985: * HOL/Auth: new protocol proofs including some for the Internet nipkow@4035: protocol TLS; wenzelm@3985: wenzelm@4125: * HOL/Map: new theory of `maps' a la VDM; nipkow@3982: wenzelm@4335: * HOL/simplifier: simplification procedures nat_cancel_sums for wenzelm@4335: cancelling out common nat summands from =, <, <= (in)equalities, or wenzelm@4335: differences; simplification procedures nat_cancel_factor for wenzelm@4335: cancelling common factor from =, <, <= (in)equalities over natural wenzelm@4373: sums; nat_cancel contains both kinds of procedures, it is installed by wenzelm@4373: default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL; wenzelm@4335: nipkow@3580: * HOL/simplifier: terms of the form paulson@4325: `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)' (or t=x) nipkow@3580: are rewritten to nipkow@4035: `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)', nipkow@4035: and those of the form paulson@4325: `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)' (or t=x) nipkow@4035: are rewritten to nipkow@4035: `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)', nipkow@4035: nipkow@4035: * HOL/datatype nipkow@4035: Each datatype `t' now comes with a theorem `split_t_case' of the form nipkow@3580: nipkow@4035: P(t_case f1 ... fn x) = nipkow@4035: ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) & nipkow@4035: ... oheimb@4189: (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn)) nipkow@4035: ) nipkow@4035: oheimb@4930: and a theorem `split_t_case_asm' of the form oheimb@4189: oheimb@4189: P(t_case f1 ... fn x) = oheimb@4189: ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) | oheimb@4189: ... oheimb@4189: (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn)) oheimb@4189: ) oheimb@4930: which can be added to a simpset via `addsplits'. The existing theorems oheimb@4930: expand_list_case and expand_option_case have been renamed to oheimb@4930: split_list_case and split_option_case. oheimb@4189: nipkow@4361: * HOL/Arithmetic: nipkow@4361: - `pred n' is automatically converted to `n-1'. nipkow@4361: Users are strongly encouraged not to use `pred' any longer, nipkow@4361: because it will disappear altogether at some point. nipkow@4361: - Users are strongly encouraged to write "0 < n" rather than nipkow@4361: "n ~= 0". Theorems and proof tools have been modified towards this nipkow@4361: `standard'. nipkow@4357: nipkow@4502: * HOL/Lists: nipkow@4502: the function "set_of_list" has been renamed "set" (and its theorems too); nipkow@4502: the function "nth" now takes its arguments in the reverse order and nipkow@4502: has acquired the infix notation "!" as in "xs!n". paulson@3570: paulson@4154: * HOL/Set: UNIV is now a constant and is no longer translated to Compl{}; paulson@4154: paulson@4154: * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its paulson@4154: specialist theorems (like UN1_I) are gone. Similarly for (INT x.B x); paulson@4154: wenzelm@4575: * HOL/record: extensible records with schematic structural subtyping wenzelm@4575: (single inheritance); EXPERIMENTAL version demonstrating the encoding, wenzelm@4575: still lacks various theorems and concrete record syntax; wenzelm@4575: wenzelm@4125: paulson@3715: *** HOLCF *** wenzelm@3535: wenzelm@4125: * removed "axioms" and "generated by" sections; wenzelm@4125: oheimb@4123: * replaced "ops" section by extended "consts" section, which is capable of wenzelm@4125: handling the continuous function space "->" directly; wenzelm@4125: wenzelm@4125: * domain package: wenzelm@4125: . proves theorems immediately and stores them in the theory, wenzelm@4125: . creates hierachical name space, wenzelm@4125: . now uses normal mixfix annotations (instead of cinfix...), wenzelm@4125: . minor changes to some names and values (for consistency), wenzelm@4125: . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas, wenzelm@4125: . separator between mutual domain defs: changed "," to "and", wenzelm@4125: . improved handling of sort constraints; now they have to wenzelm@4125: appear on the left-hand side of the equations only; oheimb@4123: oheimb@4123: * fixed LAM .b syntax; wenzelm@3567: wenzelm@3744: * added extended adm_tac to simplifier in HOLCF -- can now discharge wenzelm@3744: adm (%x. P (t x)), where P is chainfinite and t continuous; wenzelm@3579: wenzelm@3579: paulson@3719: *** FOL and ZF *** paulson@3719: wenzelm@5726: * FOL: there is a new splitter `split_asm_tac' that can be used e.g. oheimb@4189: with `addloop' of the simplifier to faciliate case splitting in premises. oheimb@4189: wenzelm@3744: * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as wenzelm@3744: in HOL, they strip ALL and --> from proved theorems; wenzelm@3744: paulson@3719: wenzelm@3579: wenzelm@3006: New in Isabelle94-8 (May 1997) wenzelm@3006: ------------------------------ wenzelm@2654: paulson@3002: *** General Changes *** paulson@3002: paulson@3002: * new utilities to build / run / maintain Isabelle etc. (in parts paulson@3002: still somewhat experimental); old Makefiles etc. still functional; wenzelm@2971: wenzelm@3205: * new 'Isabelle System Manual'; wenzelm@3205: wenzelm@2825: * INSTALL text, together with ./configure and ./build scripts; wenzelm@2773: wenzelm@3006: * reimplemented type inference for greater efficiency, better error wenzelm@3006: messages and clean internal interface; paulson@3002: paulson@3002: * prlim command for dealing with lots of subgoals (an easier way of paulson@3002: setting goals_limit); paulson@3002: wenzelm@3006: wenzelm@3006: *** Syntax *** paulson@3002: wenzelm@3116: * supports alternative (named) syntax tables (parser and pretty wenzelm@3116: printer); internal interface is provided by add_modesyntax(_i); wenzelm@3116: paulson@3002: * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to paulson@3002: be used in conjunction with the Isabelle symbol font; uses the paulson@3002: "symbols" syntax table; paulson@3002: wenzelm@2705: * added token_translation interface (may translate name tokens in wenzelm@2756: arbitrary ways, dependent on their type (free, bound, tfree, ...) and wenzelm@3116: the current print_mode); IMPORTANT: user print translation functions wenzelm@3116: are responsible for marking newly introduced bounds wenzelm@3116: (Syntax.mark_boundT); wenzelm@2705: wenzelm@2730: * token translations for modes "xterm" and "xterm_color" that display wenzelm@3006: names in bold, underline etc. or colors (which requires a color wenzelm@3006: version of xterm); wenzelm@2730: paulson@3002: * infixes may now be declared with names independent of their syntax; paulson@3002: paulson@3002: * added typed_print_translation (like print_translation, but may paulson@3002: access type of constant); paulson@3002: wenzelm@3006: paulson@3002: *** Classical Reasoner *** paulson@3002: paulson@3002: Blast_tac: a new tactic! It is often more powerful than fast_tac, but has paulson@3002: some limitations. Blast_tac... paulson@3002: + ignores addss, addbefore, addafter; this restriction is intrinsic paulson@3002: + ignores elimination rules that don't have the correct format wenzelm@5726: (the conclusion MUST be a formula variable) paulson@3002: + ignores types, which can make HOL proofs fail paulson@3002: + rules must not require higher-order unification, e.g. apply_type in ZF paulson@3002: [message "Function Var's argument not a bound variable" relates to this] paulson@3002: + its proof strategy is more general but can actually be slower paulson@3002: wenzelm@3107: * substitution with equality assumptions no longer permutes other wenzelm@3107: assumptions; paulson@3002: paulson@3002: * minor changes in semantics of addafter (now called addaltern); renamed paulson@3002: setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper wenzelm@3107: (and access functions for it); paulson@3002: wenzelm@5726: * improved combination of classical reasoner and simplifier: oheimb@3317: + functions for handling clasimpsets oheimb@3317: + improvement of addss: now the simplifier is called _after_ the oheimb@3317: safe steps. oheimb@3317: + safe variant of addss called addSss: uses safe simplifications wenzelm@5726: _during_ the safe steps. It is more complete as it allows multiple oheimb@3317: instantiations of unknowns (e.g. with slow_tac). wenzelm@3006: paulson@3002: *** Simplifier *** paulson@3002: wenzelm@3006: * added interface for simplification procedures (functions that wenzelm@3006: produce *proven* rewrite rules on the fly, depending on current wenzelm@3006: redex); wenzelm@3006: wenzelm@3006: * ordering on terms as parameter (used for ordered rewriting); wenzelm@3006: wenzelm@3107: * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss; paulson@3002: paulson@3002: * the solver is now split into a safe and an unsafe part. paulson@3002: This should be invisible for the normal user, except that the paulson@3002: functions setsolver and addsolver have been renamed to setSolver and wenzelm@3107: addSolver; added safe_asm_full_simp_tac; paulson@3002: paulson@3002: paulson@3002: *** HOL *** paulson@3002: nipkow@3042: * a generic induction tactic `induct_tac' which works for all datatypes and wenzelm@3107: also for type `nat'; nipkow@3042: nipkow@3316: * a generic case distinction tactic `exhaust_tac' which works for all nipkow@3316: datatypes and also for type `nat'; nipkow@3316: nipkow@3316: * each datatype comes with a function `size'; nipkow@3316: paulson@3002: * patterns in case expressions allow tuple patterns as arguments to wenzelm@3107: constructors, for example `case x of [] => ... | (x,y,z)#ps => ...'; paulson@3002: paulson@3002: * primrec now also works with type nat; paulson@3002: paulson@3338: * recdef: a new declaration form, allows general recursive functions to be paulson@3338: defined in theory files. See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify. paulson@3338: paulson@3002: * the constant for negation has been renamed from "not" to "Not" to wenzelm@3107: harmonize with FOL, ZF, LK, etc.; paulson@3002: wenzelm@3107: * HOL/ex/LFilter theory of a corecursive "filter" functional for wenzelm@3107: infinite lists; paulson@3002: wenzelm@3227: * HOL/Modelcheck demonstrates invocation of model checker oracle; wenzelm@3227: paulson@3002: * HOL/ex/Ring.thy declares cring_simp, which solves equational paulson@3002: problems in commutative rings, using axiomatic type classes for + and *; paulson@3002: paulson@3002: * more examples in HOL/MiniML and HOL/Auth; paulson@3002: paulson@3002: * more default rewrite rules for quantifiers, union/intersection; paulson@3002: nipkow@3321: * a new constant `arbitrary == @x.False'; nipkow@3321: wenzelm@3107: * HOLCF/IOA replaces old HOL/IOA; wenzelm@3107: wenzelm@5726: * HOLCF changes: derived all rules and arities wenzelm@5726: + axiomatic type classes instead of classes slotosch@2653: + typedef instead of faking type definitions paulson@2747: + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc. wenzelm@2730: + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po slotosch@2653: + eliminated the types void, one, tr slotosch@2653: + use unit lift and bool lift (with translations) instead of one and tr slotosch@2653: + eliminated blift from Lift3.thy (use Def instead of blift) wenzelm@3107: all eliminated rules are derived as theorems --> no visible changes ; oheimb@2649: wenzelm@3006: paulson@3002: *** ZF *** wenzelm@2553: paulson@2865: * ZF now has Fast_tac, Simp_tac and Auto_tac. Union_iff is a now a default paulson@2865: rewrite rule; this may affect some proofs. eq_cs is gone but can be put back paulson@2865: as ZF_cs addSIs [equalityI]; wenzelm@2553: wenzelm@2554: wenzelm@2732: wenzelm@2553: New in Isabelle94-7 (November 96) wenzelm@2553: --------------------------------- wenzelm@2553: wenzelm@2553: * allowing negative levels (as offsets) in prlev and choplev; wenzelm@2553: wenzelm@2554: * super-linear speedup for large simplifications; wenzelm@2554: wenzelm@2554: * FOL, ZF and HOL now use miniscoping: rewriting pushes wenzelm@2554: quantifications in as far as possible (COULD MAKE EXISTING PROOFS wenzelm@2554: FAIL); can suppress it using the command Delsimps (ex_simps @ wenzelm@2554: all_simps); De Morgan laws are also now included, by default; wenzelm@2554: wenzelm@2554: * improved printing of ==> : ~: wenzelm@2554: wenzelm@2554: * new object-logic "Sequents" adds linear logic, while replacing LK wenzelm@2554: and Modal (thanks to Sara Kalvala); wenzelm@2554: wenzelm@2554: * HOL/Auth: correctness proofs for authentication protocols; wenzelm@2554: wenzelm@2554: * HOL: new auto_tac combines rewriting and classical reasoning (many wenzelm@2554: examples on HOL/Auth); wenzelm@2554: wenzelm@2554: * HOL: new command AddIffs for declaring theorems of the form P=Q to wenzelm@2554: the rewriter and classical reasoner simultaneously; wenzelm@2554: wenzelm@2554: * function uresult no longer returns theorems in "standard" format; wenzelm@2554: regain previous version by: val uresult = standard o uresult; wenzelm@2554: wenzelm@2554: wenzelm@2554: wenzelm@2554: New in Isabelle94-6 wenzelm@2554: ------------------- wenzelm@2554: wenzelm@2554: * oracles -- these establish an interface between Isabelle and trusted wenzelm@2554: external reasoners, which may deliver results as theorems; wenzelm@2554: wenzelm@2554: * proof objects (in particular record all uses of oracles); wenzelm@2554: wenzelm@2554: * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset; wenzelm@2554: wenzelm@2554: * "constdefs" section in theory files; wenzelm@2554: wenzelm@2554: * "primrec" section (HOL) no longer requires names; wenzelm@2554: wenzelm@2554: * internal type "tactic" now simply "thm -> thm Sequence.seq"; wenzelm@2554: wenzelm@2554: wenzelm@2554: wenzelm@2554: New in Isabelle94-5 wenzelm@2554: ------------------- wenzelm@2554: wenzelm@2554: * reduced space requirements; wenzelm@2554: wenzelm@2554: * automatic HTML generation from theories; wenzelm@2554: wenzelm@2554: * theory files no longer require "..." (quotes) around most types; wenzelm@2554: wenzelm@2554: * new examples, including two proofs of the Church-Rosser theorem; wenzelm@2554: wenzelm@2554: * non-curried (1994) version of HOL is no longer distributed; wenzelm@2554: wenzelm@2553: paulson@2557: paulson@2557: New in Isabelle94-4 paulson@2557: ------------------- paulson@2557: paulson@2747: * greatly reduced space requirements; paulson@2557: paulson@2557: * theory files (.thy) no longer require \...\ escapes at line breaks; paulson@2557: wenzelm@5726: * searchable theorem database (see the section "Retrieving theorems" on paulson@2557: page 8 of the Reference Manual); paulson@2557: paulson@2557: * new examples, including Grabczewski's monumental case study of the paulson@2557: Axiom of Choice; paulson@2557: paulson@2557: * The previous version of HOL renamed to Old_HOL; paulson@2557: wenzelm@5726: * The new version of HOL (previously called CHOL) uses a curried syntax paulson@2557: for functions. Application looks like f a b instead of f(a,b); paulson@2557: paulson@2557: * Mutually recursive inductive definitions finally work in HOL; paulson@2557: paulson@2557: * In ZF, pattern-matching on tuples is now available in all abstractions and paulson@2557: translates to the operator "split"; paulson@2557: paulson@2557: paulson@2557: paulson@2557: New in Isabelle94-3 paulson@2557: ------------------- paulson@2557: wenzelm@5726: * new infix operator, addss, allowing the classical reasoner to paulson@2557: perform simplification at each step of its search. Example: wenzelm@5726: fast_tac (cs addss ss) paulson@2557: wenzelm@5726: * a new logic, CHOL, the same as HOL, but with a curried syntax wenzelm@5726: for functions. Application looks like f a b instead of f(a,b). Also pairs paulson@2557: look like (a,b) instead of ; paulson@2557: paulson@2557: * PLEASE NOTE: CHOL will eventually replace HOL! paulson@2557: paulson@2557: * In CHOL, pattern-matching on tuples is now available in all abstractions. paulson@2557: It translates to the operator "split". A new theory of integers is available; paulson@2557: paulson@2557: * In ZF, integer numerals now denote two's-complement binary integers. paulson@2557: Arithmetic operations can be performed by rewriting. See ZF/ex/Bin.ML; paulson@2557: wenzelm@5726: * Many new examples: I/O automata, Church-Rosser theorem, equivalents paulson@2557: of the Axiom of Choice; paulson@2557: paulson@2557: paulson@2557: paulson@2557: New in Isabelle94-2 paulson@2557: ------------------- paulson@2557: wenzelm@5726: * Significantly faster resolution; paulson@2557: paulson@2557: * the different sections in a .thy file can now be mixed and repeated paulson@2557: freely; paulson@2557: paulson@2557: * Database of theorems for FOL, HOL and ZF. New paulson@2557: commands including qed, qed_goal and bind_thm store theorems in the database. paulson@2557: paulson@2557: * Simple database queries: return a named theorem (get_thm) or all theorems of paulson@2557: a given theory (thms_of), or find out what theory a theorem was proved in paulson@2557: (theory_of_thm); paulson@2557: paulson@2557: * Bugs fixed in the inductive definition and datatype packages; paulson@2557: paulson@2557: * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs paulson@2557: and HOL_dup_cs obsolete; paulson@2557: paulson@2557: * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1 paulson@2557: have been removed; paulson@2557: paulson@2557: * Simpler definition of function space in ZF; paulson@2557: paulson@2557: * new results about cardinal and ordinal arithmetic in ZF; paulson@2557: paulson@2557: * 'subtype' facility in HOL for introducing new types as subsets of existing paulson@2557: types; paulson@2557: paulson@2557: wenzelm@2553: $Id$