src/HOL/Tools/ATP/atp_translate.ML
Thu, 28 Jul 2011 11:43:45 +0200 fixed lambda concealing
Tue, 26 Jul 2011 22:53:06 +0200 renamed "preds" encodings to "guards"
Tue, 26 Jul 2011 14:53:00 +0200 further worked around LEO-II parser limitation, with eta-expansion
Tue, 26 Jul 2011 14:53:00 +0200 no need for existential witnesses for sorts in TFF and THF formats
Tue, 26 Jul 2011 14:53:00 +0200 mangle "undefined"
Mon, 25 Jul 2011 14:10:12 +0200 declare "undefined" constant
Mon, 25 Jul 2011 14:10:12 +0200 avoid needless type args for lifted-lambdas
Thu, 21 Jul 2011 21:29:10 +0200 make "concealed" lambda translation sound
Wed, 20 Jul 2011 23:47:27 +0200 use a more robust naming convention for "polymorphic" frees -- the check is an overapproximation but that's fine as far as soundness is concerned
Wed, 20 Jul 2011 00:37:42 +0200 pass type arguments to lambda-lifted Frees, to account for polymorphism
Wed, 20 Jul 2011 00:37:42 +0200 generate slightly less type information -- this should be sound since type arguments should keep things cleanly apart
Wed, 20 Jul 2011 00:37:42 +0200 avoid calling "Term.is_first_order" (indirectly) on a term with loose de Bruijns -- this is not necessary anyway because of the Abs check in "simple_translate_lambdas"
Sun, 17 Jul 2011 14:21:19 +0200 fixed lambda-liftg: must ensure the formulas are in close form
Sun, 17 Jul 2011 14:12:45 +0200 ensure that the lambda translation procedure is called only once with all the facts, which is necessary for soundness of lambda-lifting (freshness of new names)
Sun, 17 Jul 2011 14:11:35 +0200 pass kind to lambda-translation function
Sun, 17 Jul 2011 14:11:35 +0200 more refactoring of preprocessing
Sun, 17 Jul 2011 14:11:35 +0200 more refactoring of preprocessing, so as to be able to centralize it
Sun, 17 Jul 2011 14:11:35 +0200 renamed internal data structure
Sun, 17 Jul 2011 14:11:35 +0200 simplify code -- there are no lambdas in helpers anyway
Sun, 17 Jul 2011 14:11:35 +0200 added lambda-lifting to Sledgehammer (rough)
Sun, 17 Jul 2011 14:11:34 +0200 move more lambda-handling logic to Sledgehammer, from ATP module, for formal dependency reasons
Thu, 14 Jul 2011 17:29:30 +0200 move error logic closer to user
Thu, 14 Jul 2011 16:50:05 +0200 move lambda translation option from ATP to Sledgehammer, to avoid accidentally breaking Metis (its reconstruction code can only deal with combinators)
Thu, 14 Jul 2011 16:50:05 +0200 added option to control which lambda translation to use (for experiments)
Wed, 06 Jul 2011 17:19:34 +0100 make SML/NJ happier
Wed, 06 Jul 2011 17:19:34 +0100 make SML/NJ happy + tuning
Tue, 05 Jul 2011 17:09:59 +0100 improved translation of lambdas in THF
Tue, 05 Jul 2011 17:09:59 +0100 added generation of lambdas in THF
Tue, 05 Jul 2011 17:09:59 +0100 add support for lambdas in TPTP THF generator + killed an unsound type encoding (because the monotonicity calculus assumes first-order)
Fri, 01 Jul 2011 15:53:38 +0200 further repair "mangled_tags", now that tags are also mangled
Fri, 01 Jul 2011 15:53:38 +0200 renamed "type_sys" to "type_enc", which is more accurate
Fri, 01 Jul 2011 15:53:37 +0200 cleaner handling of higher-order simple types, so that it's also possible to use first-order simple types with LEO-II and company
Fri, 01 Jul 2011 15:53:37 +0200 mangle "ti" tags
Fri, 01 Jul 2011 15:53:37 +0200 tuning
Mon, 27 Jun 2011 14:56:28 +0200 added "sound" option to force Sledgehammer to be pedantically sound
Tue, 21 Jun 2011 17:17:39 +0200 don't change the way helpers are generated for the exporter's sake
Tue, 21 Jun 2011 17:17:39 +0200 remove historical bloat -- another benefit of merging Metis's and Sledgehammer's translations
Tue, 21 Jun 2011 17:17:39 +0200 avoid double ASCII-fication
Tue, 21 Jun 2011 17:17:39 +0200 generate type predicates for existentials/skolems, otherwise some problems might not be provable
Thu, 16 Jun 2011 13:50:35 +0200 gave up an optimization that sometimes lead to unsound proofs -- in short, facts talking about a schematic type variable can encode a cardinality constraint and be consistent with HOL, e.g. "card (UNIV::?'a set) = 1 ==> ALL x y. x = y"
Thu, 16 Jun 2011 13:50:35 +0200 fixed soundness bug related to extensionality
Wed, 15 Jun 2011 14:36:41 +0200 fixed soundness bug made more visible by previous change
Wed, 15 Jun 2011 14:36:41 +0200 type arguments now (unlike back when fa2cf11d6351 was done) normally carry enough information to reconstruct the type of an applied constant, so no need to constraint the argument types in those cases
Fri, 10 Jun 2011 17:52:09 +0200 name tuning
Fri, 10 Jun 2011 17:52:09 +0200 revived the lightweight "poly_tags_{query,bang}" type encodings by fixing their soundness bug
Thu, 09 Jun 2011 16:34:49 +0200 discontinued Name.variant to emphasize that this is old-style / indirect;
Thu, 09 Jun 2011 00:16:28 +0200 cleaner fact freshening, which also works in corner cases, e.g. if two backquoted facts have the same name (but have different variable indices)
Wed, 08 Jun 2011 22:13:49 +0200 merged
Wed, 08 Jun 2011 17:49:01 +0200 updated headers;
Wed, 08 Jun 2011 15:56:57 +0200 more robust exception pattern General.Subscript;
Wed, 08 Jun 2011 16:20:19 +0200 pass Metis facts and negated conjecture as facts, with (almost) correctly set localities, so that the correct encoding is used for nonmonotonic occurrences of infinite types
Wed, 08 Jun 2011 16:20:18 +0200 made "query" type systes a bit more sound -- local facts, e.g. the negated conjecture, may make invalid the infinity check, e.g. if we are proving that there exists two values of an infinite type, we can use the negated conjecture that there is only one value to derive unsound proofs unless the type is properly encoded
Wed, 08 Jun 2011 16:20:18 +0200 fixed format selection logic for Waldmeister
Wed, 08 Jun 2011 08:47:43 +0200 minor optimization
Wed, 08 Jun 2011 08:47:43 +0200 don't needlessly extensionalize
Wed, 08 Jun 2011 08:47:43 +0200 don't needlessly presimplify -- makes ATP problem preparation much faster
Wed, 08 Jun 2011 08:47:43 +0200 tuned
Wed, 08 Jun 2011 08:47:43 +0200 killed "explicit_apply" option in Sledgehammer -- the "smart" default is about as lightweight as "false" and just as complete as "true"
Wed, 08 Jun 2011 08:47:43 +0200 slightly faster/cleaner accumulation of polymorphic consts
Tue, 07 Jun 2011 14:17:35 +0200 more suitable implementation of "schematic_consts_of" for monomorphizer, for ATPs