NEWS
author haftmann
Fri Apr 20 11:21:33 2007 +0200 (2007-04-20)
changeset 22735 cf627add250a
parent 22450 51ee032f9591
child 22754 9947ae4792a0
permissions -rw-r--r--
code generator changes
     1 Isabelle NEWS -- history user-relevant changes
     2 ==============================================
     3 
     4 New in this Isabelle version
     5 ----------------------------
     6 
     7 *** General ***
     8 
     9 * Theory syntax: the header format ``theory A = B + C:'' has been
    10 discontinued in favour of ``theory A imports B C begin''.  Use isatool
    11 fixheaders to convert existing theory files.  INCOMPATIBILITY.
    12 
    13 * Theory syntax: the old non-Isar theory file format has been
    14 discontinued altogether.  Note that ML proof scripts may still be used
    15 with Isar theories; migration is usually quite simple with the ML
    16 function use_legacy_bindings.  INCOMPATIBILITY.
    17 
    18 * Theory syntax: some popular names (e.g. "class", "fun", "help",
    19 "if") are now keywords.  INCOMPATIBILITY, use double quotes.
    20 
    21 * Legacy goal package: reduced interface to the bare minimum required
    22 to keep existing proof scripts running.  Most other user-level
    23 functions are now part of the OldGoals structure, which is *not* open
    24 by default (consider isatool expandshort before open OldGoals).
    25 Removed top_sg, prin, printyp, pprint_term/typ altogether, because
    26 these tend to cause confusion about the actual goal (!) context being
    27 used here, which is not necessarily the same as the_context().
    28 
    29 * Command 'find_theorems': support "*" wildcard in "name:" criterion.
    30 
    31 * The ``prems limit'' option (cf. ProofContext.prems_limit) is now -1
    32 by default, which means that "prems" (and also "fixed variables") are
    33 suppressed from proof state output.  Note that the ProofGeneral
    34 settings mechanism allows to change and save options persistently, but
    35 older versions of Isabelle will fail to start up if a negative prems
    36 limit is imposed.
    37 
    38 * Local theory targets may be specified by non-nested blocks of
    39 ``context/locale/class ... begin'' followed by ``end''.  The body may
    40 contain definitions, theorems etc., including any derived mechanism
    41 that has been implemented on top of these primitives.  This concept
    42 generalizes the existing ``theorem (in ...)'' towards more versatility
    43 and scalability.
    44 
    45 * Proof General interface: proper undo of final 'end' command;
    46 discontinued Isabelle/classic mode (ML proof scripts).
    47 
    48 
    49 *** Document preparation ***
    50 
    51 * Added antiquotation @{theory name} which prints the given name,
    52 after checking that it refers to a valid ancestor theory in the
    53 current context.
    54 
    55 * Added antiquotations @{ML_type text} and @{ML_struct text} which
    56 check the given source text as ML type/structure, printing verbatim.
    57 
    58 * Added antiquotation @{abbrev "c args"} which prints the abbreviation
    59 "c args == rhs" given in the current context.  (Any number of
    60 arguments may be given on the LHS.)
    61 
    62 
    63 
    64 *** Pure ***
    65 
    66 * code generator: 
    67   - Isar "definition"s and primitive instance definitions are added explicitly
    68     to the table of defining equations
    69   - primitive definitions are not unfolded by default any longer
    70   - defining equations are now definitly restricted to meta "==" and object
    71         equality "="
    72   - HOL theories have been adopted accordingly
    73 
    74 * class_package.ML offers a combination of axclasses and locales to
    75 achieve Haskell-like type classes in Isabelle. See
    76 HOL/ex/Classpackage.thy for examples.
    77 
    78 * Yet another code generator framework allows to generate executable
    79 code for ML and Haskell (including "class"es). A short usage sketch:
    80 
    81     internal compilation:
    82         code_gen <list of constants (term syntax)> (SML #)
    83     writing SML code to a file:
    84         code_gen <list of constants (term syntax)> (SML <filename>)
    85     writing OCaml code to a file:
    86         code_gen <list of constants (term syntax)> (OCaml <filename>)
    87     writing Haskell code to a bunch of files:
    88         code_gen <list of constants (term syntax)> (Haskell <filename>)
    89 
    90 Reasonable default setup of framework in HOL/Main.
    91 
    92 Theorem attributs for selecting and transforming function equations theorems:
    93 
    94     [code fun]:       select a theorem as function equation for a specific constant
    95     [code nofun]:     deselect a theorem as function equation for a specific constant
    96     [code inline]:    select an equation theorem for unfolding (inlining) in place
    97     [code noinline]:  deselect an equation theorem for unfolding (inlining) in place
    98 
    99 User-defined serializations (target in {SML, OCaml, Haskell}):
   100 
   101     code_const <and-list of constants (term syntax)>
   102       {(target) <and-list of const target syntax>}+
   103 
   104     code_type <and-list of type constructors>
   105       {(target) <and-list of type target syntax>}+
   106 
   107     code_instance <and-list of instances>
   108       {(target)}+
   109         where instance ::= <type constructor> :: <class>
   110 
   111     code_class <and_list of classes>
   112       {(target) <and-list of class target syntax>}+
   113         where class target syntax ::= <class name> {where {<classop> == <target syntax>}+}?
   114 
   115 code_instance and code_class only apply to target Haskell.
   116 
   117 See HOL theories and HOL/ex/Codegenerator*.thy for usage examples.
   118 Doc/Isar/Advanced/Codegen/ provides a tutorial.
   119 
   120 * Command 'no_translations' removes translation rules from theory
   121 syntax.
   122 
   123 * Overloaded definitions are now actually checked for acyclic
   124 dependencies.  The overloading scheme is slightly more general than
   125 that of Haskell98, although Isabelle does not demand an exact
   126 correspondence to type class and instance declarations.
   127 INCOMPATIBILITY, use ``defs (unchecked overloaded)'' to admit more
   128 exotic versions of overloading -- at the discretion of the user!
   129 
   130 Polymorphic constants are represented via type arguments, i.e. the
   131 instantiation that matches an instance against the most general
   132 declaration given in the signature.  For example, with the declaration
   133 c :: 'a => 'a => 'a, an instance c :: nat => nat => nat is represented
   134 as c(nat).  Overloading is essentially simultaneous structural
   135 recursion over such type arguments.  Incomplete specification patterns
   136 impose global constraints on all occurrences, e.g. c('a * 'a) on the
   137 LHS means that more general c('a * 'b) will be disallowed on any RHS.
   138 Command 'print_theory' outputs the normalized system of recursive
   139 equations, see section "definitions".
   140 
   141 * Isar: improper proof element 'guess' is like 'obtain', but derives
   142 the obtained context from the course of reasoning!  For example:
   143 
   144   assume "EX x y. A x & B y"   -- "any previous fact"
   145   then guess x and y by clarify
   146 
   147 This technique is potentially adventurous, depending on the facts and
   148 proof tools being involved here.
   149 
   150 * Isar: known facts from the proof context may be specified as literal
   151 propositions, using ASCII back-quote syntax.  This works wherever
   152 named facts used to be allowed so far, in proof commands, proof
   153 methods, attributes etc.  Literal facts are retrieved from the context
   154 according to unification of type and term parameters.  For example,
   155 provided that "A" and "A ==> B" and "!!x. P x ==> Q x" are known
   156 theorems in the current context, then these are valid literal facts:
   157 `A` and `A ==> B` and `!!x. P x ==> Q x" as well as `P a ==> Q a` etc.
   158 
   159 There is also a proof method "fact" which does the same composition
   160 for explicit goal states, e.g. the following proof texts coincide with
   161 certain special cases of literal facts:
   162 
   163   have "A" by fact                 ==  note `A`
   164   have "A ==> B" by fact           ==  note `A ==> B`
   165   have "!!x. P x ==> Q x" by fact  ==  note `!!x. P x ==> Q x`
   166   have "P a ==> Q a" by fact       ==  note `P a ==> Q a`
   167 
   168 * Isar: ":" (colon) is no longer a symbolic identifier character in
   169 outer syntax.  Thus symbolic identifiers may be used without
   170 additional white space in declarations like this: ``assume *: A''.
   171 
   172 * Isar: 'print_facts' prints all local facts of the current context,
   173 both named and unnamed ones.
   174 
   175 * Isar: 'def' now admits simultaneous definitions, e.g.:
   176 
   177   def x == "t" and y == "u"
   178 
   179 * Isar: added command 'unfolding', which is structurally similar to
   180 'using', but affects both the goal state and facts by unfolding given
   181 rewrite rules.  Thus many occurrences of the 'unfold' method or
   182 'unfolded' attribute may be replaced by first-class proof text.
   183 
   184 * Isar: methods 'unfold' / 'fold', attributes 'unfolded' / 'folded',
   185 and command 'unfolding' now all support object-level equalities
   186 (potentially conditional).  The underlying notion of rewrite rule is
   187 analogous to the 'rule_format' attribute, but *not* that of the
   188 Simplifier (which is usually more generous).
   189 
   190 * Isar: the goal restriction operator [N] (default N = 1) evaluates a
   191 method expression within a sandbox consisting of the first N
   192 sub-goals, which need to exist.  For example, ``simp_all [3]''
   193 simplifies the first three sub-goals, while (rule foo, simp_all)[]
   194 simplifies all new goals that emerge from applying rule foo to the
   195 originally first one.
   196 
   197 * Isar: schematic goals are no longer restricted to higher-order
   198 patterns; e.g. ``lemma "?P(?x)" by (rule TrueI)'' now works as
   199 expected.
   200 
   201 * Isar: the conclusion of a long theorem statement is now either
   202 'shows' (a simultaneous conjunction, as before), or 'obtains'
   203 (essentially a disjunction of cases with local parameters and
   204 assumptions).  The latter allows to express general elimination rules
   205 adequately; in this notation common elimination rules look like this:
   206 
   207   lemma exE:    -- "EX x. P x ==> (!!x. P x ==> thesis) ==> thesis"
   208     assumes "EX x. P x"
   209     obtains x where "P x"
   210 
   211   lemma conjE:  -- "A & B ==> (A ==> B ==> thesis) ==> thesis"
   212     assumes "A & B"
   213     obtains A and B
   214 
   215   lemma disjE:  -- "A | B ==> (A ==> thesis) ==> (B ==> thesis) ==> thesis"
   216     assumes "A | B"
   217     obtains
   218       A
   219     | B
   220 
   221 The subsequent classical rules even refer to the formal "thesis"
   222 explicitly:
   223 
   224   lemma classical:     -- "(~ thesis ==> thesis) ==> thesis"
   225     obtains "~ thesis"
   226 
   227   lemma Peirce's_Law:  -- "((thesis ==> something) ==> thesis) ==> thesis"
   228     obtains "thesis ==> something"
   229 
   230 The actual proof of an 'obtains' statement is analogous to that of the
   231 Isar proof element 'obtain', only that there may be several cases.
   232 Optional case names may be specified in parentheses; these will be
   233 available both in the present proof and as annotations in the
   234 resulting rule, for later use with the 'cases' method (cf. attribute
   235 case_names).
   236 
   237 * Isar: the assumptions of a long theorem statement are available as
   238 "assms" fact in the proof context.  This is more appropriate than the
   239 (historical) "prems", which refers to all assumptions of the current
   240 context, including those from the target locale, proof body etc.
   241 
   242 * Isar: 'print_statement' prints theorems from the current theory or
   243 proof context in long statement form, according to the syntax of a
   244 top-level lemma.
   245 
   246 * Isar: 'obtain' takes an optional case name for the local context
   247 introduction rule (default "that").
   248 
   249 * Isar: removed obsolete 'concl is' patterns.  INCOMPATIBILITY, use
   250 explicit (is "_ ==> ?foo") in the rare cases where this still happens
   251 to occur.
   252 
   253 * Pure: syntax "CONST name" produces a fully internalized constant
   254 according to the current context.  This is particularly useful for
   255 syntax translations that should refer to internal constant
   256 representations independently of name spaces.
   257 
   258 * Pure: syntax constant for foo (binder "FOO ") is called "foo_binder"
   259 instead of "FOO ". This allows multiple binder declarations to coexist
   260 in the same context.  INCOMPATIBILITY.
   261 
   262 * Isar/locales: 'notation' provides a robust interface to the 'syntax'
   263 primitive that also works in a locale context (both for constants and
   264 fixed variables).  Type declaration and internal syntactic
   265 representation of given constants retrieved from the context.
   266 
   267 * Isar/locales: new derived specification elements 'axiomatization',
   268 'definition', 'abbreviation', which support type-inference, admit
   269 object-level specifications (equality, equivalence).  See also the
   270 isar-ref manual.  Examples:
   271 
   272   axiomatization
   273     eq  (infix "===" 50) where
   274     eq_refl: "x === x" and eq_subst: "x === y ==> P x ==> P y"
   275 
   276   definition "f x y = x + y + 1"
   277   definition g where "g x = f x x"
   278 
   279   abbreviation
   280     neq  (infix "=!=" 50) where
   281     "x =!= y == ~ (x === y)"
   282 
   283 These specifications may be also used in a locale context.  Then the
   284 constants being introduced depend on certain fixed parameters, and the
   285 constant name is qualified by the locale base name.  An internal
   286 abbreviation takes care for convenient input and output, making the
   287 parameters implicit and using the original short name.  See also
   288 HOL/ex/Abstract_NAT.thy for an example of deriving polymorphic
   289 entities from a monomorphic theory.
   290 
   291 Presently, abbreviations are only available 'in' a target locale, but
   292 not inherited by general import expressions.  Also note that
   293 'abbreviation' may be used as a type-safe replacement for 'syntax' +
   294 'translations' in common applications.
   295 
   296 Concrete syntax is attached to specified constants in internal form,
   297 independently of name spaces.  The parse tree representation is
   298 slightly different -- use 'notation' instead of raw 'syntax', and
   299 'translations' with explicit "CONST" markup to accommodate this.
   300 
   301 * Pure: command 'print_abbrevs' prints all constant abbreviations of
   302 the current context.  Print mode "no_abbrevs" prevents inversion of
   303 abbreviations on output.
   304 
   305 * Isar/locales: improved parameter handling:
   306 - use of locales "var" and "struct" no longer necessary;
   307 - parameter renamings are no longer required to be injective.
   308   This enables, for example, to define a locale for endomorphisms thus:
   309   locale endom = homom mult mult h.
   310 
   311 * Isar/locales: changed the way locales with predicates are defined.
   312 Instead of accumulating the specification, the imported expression is
   313 now an interpretation.  INCOMPATIBILITY: different normal form of
   314 locale expressions.  In particular, in interpretations of locales with
   315 predicates, goals repesenting already interpreted fragments are not
   316 removed automatically.  Use methods `intro_locales' and
   317 `unfold_locales'; see below.
   318 
   319 * Isar/locales: new methods `intro_locales' and `unfold_locales'
   320 provide backward reasoning on locales predicates.  The methods are
   321 aware of interpretations and discharge corresponding goals.
   322 `intro_locales' is less aggressive then `unfold_locales' and does not
   323 unfold predicates to assumptions.
   324 
   325 * Isar/locales: the order in which locale fragments are accumulated
   326 has changed.  This enables to override declarations from fragments due
   327 to interpretations -- for example, unwanted simp rules.
   328 
   329 * Provers/induct: improved internal context management to support
   330 local fixes and defines on-the-fly.  Thus explicit meta-level
   331 connectives !! and ==> are rarely required anymore in inductive goals
   332 (using object-logic connectives for this purpose has been long
   333 obsolete anyway).  The subsequent proof patterns illustrate advanced
   334 techniques of natural induction; general datatypes and inductive sets
   335 work analogously (see also src/HOL/Lambda for realistic examples).
   336 
   337 (1) This is how to ``strengthen'' an inductive goal wrt. certain
   338 parameters:
   339 
   340   lemma
   341     fixes n :: nat and x :: 'a
   342     assumes a: "A n x"
   343     shows "P n x"
   344     using a                     -- {* make induct insert fact a *}
   345   proof (induct n arbitrary: x) -- {* generalize goal to "!!x. A n x ==> P n x" *}
   346     case 0
   347     show ?case sorry
   348   next
   349     case (Suc n)
   350     note `!!x. A n x ==> P n x` -- {* induction hypothesis, according to induction rule *}
   351     note `A (Suc n) x`          -- {* induction premise, stemming from fact a *}
   352     show ?case sorry
   353   qed
   354 
   355 (2) This is how to perform induction over ``expressions of a certain
   356 form'', using a locally defined inductive parameter n == "a x"
   357 together with strengthening (the latter is usually required to get
   358 sufficiently flexible induction hypotheses):
   359 
   360   lemma
   361     fixes a :: "'a => nat"
   362     assumes a: "A (a x)"
   363     shows "P (a x)"
   364     using a
   365   proof (induct n == "a x" arbitrary: x)
   366     ...
   367 
   368 See also HOL/Isar_examples/Puzzle.thy for an application of the this
   369 particular technique.
   370 
   371 (3) This is how to perform existential reasoning ('obtains' or
   372 'obtain') by induction, while avoiding explicit object-logic
   373 encodings:
   374 
   375   lemma
   376     fixes n :: nat
   377     obtains x :: 'a where "P n x" and "Q n x"
   378   proof (induct n arbitrary: thesis)
   379     case 0
   380     obtain x where "P 0 x" and "Q 0 x" sorry
   381     then show thesis by (rule 0)
   382   next
   383     case (Suc n)
   384     obtain x where "P n x" and "Q n x" by (rule Suc.hyps)
   385     obtain x where "P (Suc n) x" and "Q (Suc n) x" sorry
   386     then show thesis by (rule Suc.prems)
   387   qed
   388 
   389 Here the 'arbitrary: thesis' specification essentially modifies the
   390 scope of the formal thesis parameter, in order to the get the whole
   391 existence statement through the induction as expected.
   392 
   393 * Provers/induct: mutual induction rules are now specified as a list
   394 of rule sharing the same induction cases.  HOL packages usually
   395 provide foo_bar.inducts for mutually defined items foo and bar
   396 (e.g. inductive sets or datatypes).  INCOMPATIBILITY, users need to
   397 specify mutual induction rules differently, i.e. like this:
   398 
   399   (induct rule: foo_bar.inducts)
   400   (induct set: foo bar)
   401   (induct type: foo bar)
   402 
   403 The ML function ProjectRule.projections turns old-style rules into the
   404 new format.
   405 
   406 * Provers/induct: improved handling of simultaneous goals.  Instead of
   407 introducing object-level conjunction, the statement is now split into
   408 several conclusions, while the corresponding symbolic cases are
   409 nested accordingly.  INCOMPATIBILITY, proofs need to be structured
   410 explicitly.  For example:
   411 
   412   lemma
   413     fixes n :: nat
   414     shows "P n" and "Q n"
   415   proof (induct n)
   416     case 0 case 1
   417     show "P 0" sorry
   418   next
   419     case 0 case 2
   420     show "Q 0" sorry
   421   next
   422     case (Suc n) case 1
   423     note `P n` and `Q n`
   424     show "P (Suc n)" sorry
   425   next
   426     case (Suc n) case 2
   427     note `P n` and `Q n`
   428     show "Q (Suc n)" sorry
   429   qed
   430 
   431 The split into subcases may be deferred as follows -- this is
   432 particularly relevant for goal statements with local premises.
   433 
   434   lemma
   435     fixes n :: nat
   436     shows "A n ==> P n" and "B n ==> Q n"
   437   proof (induct n)
   438     case 0
   439     {
   440       case 1
   441       note `A 0`
   442       show "P 0" sorry
   443     next
   444       case 2
   445       note `B 0`
   446       show "Q 0" sorry
   447     }
   448   next
   449     case (Suc n)
   450     note `A n ==> P n` and `B n ==> Q n`
   451     {
   452       case 1
   453       note `A (Suc n)`
   454       show "P (Suc n)" sorry
   455     next
   456       case 2
   457       note `B (Suc n)`
   458       show "Q (Suc n)" sorry
   459     }
   460   qed
   461 
   462 If simultaneous goals are to be used with mutual rules, the statement
   463 needs to be structured carefully as a two-level conjunction, using
   464 lists of propositions separated by 'and':
   465 
   466   lemma
   467     shows "a : A ==> P1 a"
   468           "a : A ==> P2 a"
   469       and "b : B ==> Q1 b"
   470           "b : B ==> Q2 b"
   471           "b : B ==> Q3 b"
   472   proof (induct set: A B)
   473 
   474 * Provers/induct: support coinduction as well.  See
   475 src/HOL/Library/Coinductive_List.thy for various examples.
   476 
   477 * Attribute "symmetric" produces result with standardized schematic
   478 variables (index 0).  Potential INCOMPATIBILITY.
   479 
   480 * Simplifier: by default the simplifier trace only shows top level
   481 rewrites now. That is, trace_simp_depth_limit is set to 1 by
   482 default. Thus there is less danger of being flooded by the trace. The
   483 trace indicates where parts have been suppressed.
   484   
   485 * Provers/classical: removed obsolete classical version of elim_format
   486 attribute; classical elim/dest rules are now treated uniformly when
   487 manipulating the claset.
   488 
   489 * Provers/classical: stricter checks to ensure that supplied intro,
   490 dest and elim rules are well-formed; dest and elim rules must have at
   491 least one premise.
   492 
   493 * Provers/classical: attributes dest/elim/intro take an optional
   494 weight argument for the rule (just as the Pure versions).  Weights are
   495 ignored by automated tools, but determine the search order of single
   496 rule steps.
   497 
   498 * Syntax: input syntax now supports dummy variable binding "%_. b",
   499 where the body does not mention the bound variable.  Note that dummy
   500 patterns implicitly depend on their context of bounds, which makes
   501 "{_. _}" match any set comprehension as expected.  Potential
   502 INCOMPATIBILITY -- parse translations need to cope with syntactic
   503 constant "_idtdummy" in the binding position.
   504 
   505 * Syntax: removed obsolete syntactic constant "_K" and its associated
   506 parse translation.  INCOMPATIBILITY -- use dummy abstraction instead,
   507 for example "A -> B" => "Pi A (%_. B)".
   508 
   509 * Pure: 'class_deps' command visualizes the subclass relation, using
   510 the graph browser tool.
   511 
   512 * Pure: 'print_theory' now suppresses entities with internal name
   513 (trailing "_") by default; use '!' option for full details.
   514 
   515 
   516 *** HOL ***
   517 
   518 * Library/Commutative_Ring.thy: switched from recdef to function package;
   519     constants add, mul, pow now curried.  Infix syntax for algebraic operations.
   520 
   521 * Some steps towards more uniform lattice theory development in HOL.
   522 
   523     constants "meet" and "join" now named "inf" and "sup"
   524     constant "Meet" now named "Inf"
   525 
   526     classes "meet_semilorder" and "join_semilorder" now named
   527       "lower_semilattice" and "upper_semilattice"
   528     class "lorder" now named "lattice"
   529     class "comp_lat" now named "complete_lattice"
   530 
   531     Instantiation of lattice classes allows explicit definitions
   532     for "inf" and "sup" operations.
   533 
   534   INCOMPATIBILITY. Theorem renames:
   535 
   536     meet_left_le            ~> inf_le1
   537     meet_right_le           ~> inf_le2
   538     join_left_le            ~> sup_ge1
   539     join_right_le           ~> sup_ge2
   540     meet_join_le            ~> inf_sup_ord
   541     le_meetI                ~> le_infI
   542     join_leI                ~> le_supI
   543     le_meet                 ~> le_inf_iff
   544     le_join                 ~> ge_sup_conv
   545     meet_idempotent         ~> inf_idem
   546     join_idempotent         ~> sup_idem
   547     meet_comm               ~> inf_commute
   548     join_comm               ~> sup_commute
   549     meet_leI1               ~> le_infI1
   550     meet_leI2               ~> le_infI2
   551     le_joinI1               ~> le_supI1
   552     le_joinI2               ~> le_supI2
   553     meet_assoc              ~> inf_assoc
   554     join_assoc              ~> sup_assoc
   555     meet_left_comm          ~> inf_left_commute
   556     meet_left_idempotent    ~> inf_left_idem
   557     join_left_comm          ~> sup_left_commute
   558     join_left_idempotent    ~> sup_left_idem
   559     meet_aci                ~> inf_aci
   560     join_aci                ~> sup_aci
   561     le_def_meet             ~> le_iff_inf
   562     le_def_join             ~> le_iff_sup
   563     join_absorp2            ~> sup_absorb2
   564     join_absorp1            ~> sup_absorb1
   565     meet_absorp1            ~> inf_absorb1
   566     meet_absorp2            ~> inf_absorb2
   567     meet_join_absorp        ~> inf_sup_absorb
   568     join_meet_absorp        ~> sup_inf_absorb
   569     distrib_join_le         ~> distrib_sup_le
   570     distrib_meet_le         ~> distrib_inf_le
   571 
   572     add_meet_distrib_left   ~> add_inf_distrib_left
   573     add_join_distrib_left   ~> add_sup_distrib_left
   574     is_join_neg_meet        ~> is_join_neg_inf
   575     is_meet_neg_join        ~> is_meet_neg_sup
   576     add_meet_distrib_right  ~> add_inf_distrib_right
   577     add_join_distrib_right  ~> add_sup_distrib_right
   578     add_meet_join_distribs  ~> add_sup_inf_distribs
   579     join_eq_neg_meet        ~> sup_eq_neg_inf
   580     meet_eq_neg_join        ~> inf_eq_neg_sup
   581     add_eq_meet_join        ~> add_eq_inf_sup
   582     meet_0_imp_0            ~> inf_0_imp_0
   583     join_0_imp_0            ~> sup_0_imp_0
   584     meet_0_eq_0             ~> inf_0_eq_0
   585     join_0_eq_0             ~> sup_0_eq_0
   586     neg_meet_eq_join        ~> neg_inf_eq_sup
   587     neg_join_eq_meet        ~> neg_sup_eq_inf
   588     join_eq_if              ~> sup_eq_if
   589 
   590     mono_meet               ~> mono_inf
   591     mono_join               ~> mono_sup
   592     meet_bool_eq            ~> inf_bool_eq
   593     join_bool_eq            ~> sup_bool_eq
   594     meet_fun_eq             ~> inf_fun_eq
   595     join_fun_eq             ~> sup_fun_eq
   596     meet_set_eq             ~> inf_set_eq
   597     join_set_eq             ~> sup_set_eq
   598     meet1_iff               ~> inf1_iff
   599     meet2_iff               ~> inf2_iff
   600     meet1I                  ~> inf1I
   601     meet2I                  ~> inf2I
   602     meet1D1                 ~> inf1D1
   603     meet2D1                 ~> inf2D1
   604     meet1D2                 ~> inf1D2
   605     meet2D2                 ~> inf2D2
   606     meet1E                  ~> inf1E
   607     meet2E                  ~> inf2E
   608     join1_iff               ~> sup1_iff
   609     join2_iff               ~> sup2_iff
   610     join1I1                 ~> sup1I1
   611     join2I1                 ~> sup2I1
   612     join1I1                 ~> sup1I1
   613     join2I2                 ~> sup1I2
   614     join1CI                 ~> sup1CI
   615     join2CI                 ~> sup2CI
   616     join1E                  ~> sup1E
   617     join2E                  ~> sup2E
   618 
   619     is_meet_Meet            ~> is_meet_Inf
   620     Meet_bool_def           ~> Inf_bool_def
   621     Meet_fun_def            ~> Inf_fun_def
   622     Meet_greatest           ~> Inf_greatest
   623     Meet_lower              ~> Inf_lower
   624     Meet_set_def            ~> Inf_set_def
   625 
   626     listsp_meetI            ~> listsp_infI
   627     listsp_meet_eq          ~> listsp_inf_eq
   628 
   629     meet_min                ~> inf_min
   630     join_max                ~> sup_max
   631 
   632 * Class (axclass + locale) "preorder" and "linear": facts "refl", "trans" and
   633 "cases" renamed ro "order_refl", "order_trans" and "linorder_cases", to
   634 avoid clashes with HOL "refl" and "trans". INCOMPATIBILITY.
   635 
   636 * Addes class (axclass + locale) "preorder" as superclass of "order";
   637 potential INCOMPATIBILITY: order of proof goals in order/linorder instance
   638 proofs changed.
   639 
   640 * Dropped lemma duplicate def_imp_eq in favor of meta_eq_to_obj_eq.
   641 INCOMPATIBILITY.
   642 
   643 * Dropped lemma duplicate if_def2 in favor of if_bool_eq_conj.
   644 INCOMPATIBILITY.
   645 
   646 * Added syntactic class "size"; overloaded constant "size" now has
   647 type "'a::size ==> bool"
   648 
   649 * Renamed constants "Divides.op div", "Divides.op mod" and "Divides.op
   650 dvd" to "Divides.div", "Divides.mod" and "Divides.dvd". INCOMPATIBILITY.
   651 
   652 * Added method "lexicographic_order" automatically synthesizes
   653 termination relations as lexicographic combinations of size measures
   654 -- 'function' package.
   655 
   656 * HOL/records: generalised field-update to take a function on the
   657 field rather than the new value: r(|A := x|) is translated to A_update
   658 (K x) r The K-combinator that is internally used is called K_record.
   659 INCOMPATIBILITY: Usage of the plain update functions has to be
   660 adapted.
   661  
   662 * axclass "semiring_0" now contains annihilation axioms x * 0 = 0 and
   663 0 * x = 0, which are required for a semiring.  Richer structures do
   664 not inherit from semiring_0 anymore, because this property is a
   665 theorem there, not an axiom.  INCOMPATIBILITY: In instances of
   666 semiring_0, there is more to prove, but this is mostly trivial.
   667 
   668 * axclass "recpower" was generalized to arbitrary monoids, not just
   669 commutative semirings.  INCOMPATIBILITY: If you use recpower and need
   670 commutativity or a semiring property, add the corresponding classes.
   671 
   672 * Unified locale partial_order with class definition (cf. theory
   673 Orderings), added parameter ``less''.  INCOMPATIBILITY.
   674 
   675 * Constant "List.list_all2" in List.thy now uses authentic syntax.
   676 INCOMPATIBILITY: translations containing list_all2 may go wrong.  On
   677 Isar level, use abbreviations instead.
   678 
   679 * Renamed constant "List.op mem" to "List.memberl" INCOMPATIBILITY:
   680 rarely occuring name references (e.g. ``List.op mem.simps'') require
   681 renaming (e.g. ``List.memberl.simps'').
   682 
   683 * Renamed constants "0" to "HOL.zero" and "1" to "HOL.one".
   684 INCOMPATIBILITY.
   685 
   686 * Added theory Code_Generator providing class 'eq', allowing for code
   687 generation with polymorphic equality.
   688 
   689 * Numeral syntax: type 'bin' which was a mere type copy of 'int' has
   690 been abandoned in favour of plain 'int'. INCOMPATIBILITY --
   691 significant changes for setting up numeral syntax for types:
   692 
   693   - new constants Numeral.pred and Numeral.succ instead
   694       of former Numeral.bin_pred and Numeral.bin_succ.
   695   - Use integer operations instead of bin_add, bin_mult and so on.
   696   - Numeral simplification theorems named Numeral.numeral_simps instead of Bin_simps.
   697   - ML structure Bin_Simprocs now named Int_Numeral_Base_Simprocs.
   698 
   699 See HOL/Integ/IntArith.thy for an example setup.
   700 
   701 * New top level command 'normal_form' computes the normal form of a
   702 term that may contain free variables. For example ``normal_form
   703 "rev[a,b,c]"'' produces ``[b,c,a]'' (without proof).  This command is
   704 suitable for heavy-duty computations because the functions are
   705 compiled to ML first.
   706 
   707 * Alternative iff syntax "A <-> B" for equality on bool (with priority
   708 25 like -->); output depends on the "iff" print_mode, the default is
   709 "A = B" (with priority 50).
   710 
   711 * Renamed constants in HOL.thy and Orderings.thy:
   712     op +   ~> HOL.plus
   713     op -   ~> HOL.minus
   714     uminus ~> HOL.uminus
   715     op *   ~> HOL.times
   716     op <   ~> Orderings.less
   717     op <=  ~> Orderings.less_eq
   718 
   719 Adaptions may be required in the following cases:
   720 
   721 a) User-defined constants using any of the names "plus", "minus", "times",
   722 "less" or "less_eq". The standard syntax translations for "+", "-" and "*"
   723 may go wrong.
   724 INCOMPATIBILITY: use more specific names.
   725 
   726 b) Variables named "plus", "minus", "times", "less", "less_eq"
   727 INCOMPATIBILITY: use more specific names.
   728 
   729 c) Permutative equations (e.g. "a + b = b + a")
   730 Since the change of names also changes the order of terms, permutative
   731 rewrite rules may get applied in a different order. Experience shows that
   732 this is rarely the case (only two adaptions in the whole Isabelle
   733 distribution).
   734 INCOMPATIBILITY: rewrite proofs
   735 
   736 d) ML code directly refering to constant names
   737 This in general only affects hand-written proof tactics, simprocs and so on.
   738 INCOMPATIBILITY: grep your sourcecode and replace names.
   739 
   740 * Relations less (<) and less_eq (<=) are also available on type bool.
   741 Modified syntax to disallow nesting without explicit parentheses,
   742 e.g. "(x < y) < z" or "x < (y < z)", but NOT "x < y < z".
   743 
   744 * "LEAST x:A. P" expands to "LEAST x. x:A & P" (input only).
   745 
   746 * Relation composition operator "op O" now has precedence 75 and binds
   747 stronger than union and intersection. INCOMPATIBILITY.
   748 
   749 * The old set interval syntax "{m..n(}" (and relatives) has been
   750 removed.  Use "{m..<n}" (and relatives) instead.
   751 
   752 * In the context of the assumption "~(s = t)" the Simplifier rewrites
   753 "t = s" to False (by simproc "neq_simproc").  For backward
   754 compatibility this can be disabled by ML "reset use_neq_simproc".
   755 
   756 * "m dvd n" where m and n are numbers is evaluated to True/False by
   757 simp.
   758 
   759 * Theorem Cons_eq_map_conv no longer declared as ``simp''.
   760 
   761 * Theorem setsum_mult renamed to setsum_right_distrib.
   762 
   763 * Prefer ex1I over ex_ex1I in single-step reasoning, e.g. by the
   764 ``rule'' method.
   765 
   766 * Reimplemented methods ``sat'' and ``satx'', with several
   767 improvements: goals no longer need to be stated as "<prems> ==>
   768 False", equivalences (i.e. "=" on type bool) are handled, variable
   769 names of the form "lit_<n>" are no longer reserved, significant
   770 speedup.
   771 
   772 * Methods ``sat'' and ``satx'' can now replay MiniSat proof traces.
   773 zChaff is still supported as well.
   774 
   775 * 'inductive' and 'datatype': provide projections of mutual rules,
   776 bundled as foo_bar.inducts;
   777 
   778 * Library: moved theories Parity, GCD, Binomial, Infinite_Set to
   779 Library.
   780 
   781 * Library: moved theory Accessible_Part to main HOL.
   782 
   783 * Library: added theory Coinductive_List of potentially infinite lists
   784 as greatest fixed-point.
   785 
   786 * Library: added theory AssocList which implements (finite) maps as
   787 association lists.
   788 
   789 * Added proof method ``evaluation'' for efficiently solving a goal
   790 (i.e. a boolean expression) by compiling it to ML. The goal is
   791 "proved" (via an oracle) if it evaluates to True.
   792 
   793 * Linear arithmetic now splits certain operators (e.g. min, max, abs)
   794 also when invoked by the simplifier.  This results in the simplifier
   795 being more powerful on arithmetic goals.  INCOMPATIBILITY.  Set
   796 fast_arith_split_limit to 0 to obtain the old behavior.
   797 
   798 * Support for hex (0x20) and binary (0b1001) numerals.
   799 
   800 * New method: reify eqs (t), where eqs are equations for an
   801 interpretation I :: 'a list => 'b => 'c and t::'c is an optional
   802 parameter, computes a term s::'b and a list xs::'a list and proves the
   803 theorem I xs s = t. This is also known as reification or quoting. The
   804 resulting theorem is applied to the subgoal to substitute t with I xs
   805 s.  If t is omitted, the subgoal itself is reified.
   806 
   807 * New method: reflection corr_thm eqs (t). The parameters eqs and (t)
   808 are as explained above. corr_thm is a theorem for I vs (f t) = I vs t,
   809 where f is supposed to be a computable function (in the sense of code
   810 generattion). The method uses reify to compute s and xs as above then
   811 applies corr_thm and uses normalization by evaluation to "prove" f s =
   812 r and finally gets the theorem t = r, which is again applied to the
   813 subgoal. An Example is available in HOL/ex/ReflectionEx.thy.
   814 
   815 * Reflection: Automatic refification now handels binding, an example
   816 is available in HOL/ex/ReflectionEx.thy
   817 
   818 
   819 *** HOL-Algebra ***
   820 
   821 * Formalisation of ideals and the quotient construction over rings.
   822 
   823 * Order and lattice theory no longer based on records.
   824 INCOMPATIBILITY.
   825 
   826 * Renamed lemmas least_carrier -> least_closed and greatest_carrier ->
   827 greatest_closed.  INCOMPATIBILITY.
   828 
   829 * Method algebra is now set up via an attribute.  For examples see
   830 Ring.thy.  INCOMPATIBILITY: the method is now weaker on combinations
   831 of algebraic structures.
   832 
   833 * Renamed theory CRing to Ring.
   834 
   835 
   836 *** HOL-Complex ***
   837 
   838 * Theory Real: new method ferrack implements quantifier elimination
   839 for linear arithmetic over the reals. The quantifier elimination
   840 feature is used only for decision, for compatibility with arith. This
   841 means a goal is either solved or left unchanged, no simplification.
   842 
   843 * Real: New axiomatic classes formalize real normed vector spaces and
   844 algebras, using new overloaded constants scaleR :: real => 'a => 'a
   845 and norm :: 'a => real.
   846 
   847 * Real: New constant of_real :: real => 'a::real_algebra_1 injects
   848 from reals into other types. The overloaded constant Reals :: 'a set
   849 is now defined as range of_real; potential INCOMPATIBILITY.
   850 
   851 * Hyperreal: Several constants that previously worked only for the
   852 reals have been generalized, so they now work over arbitrary vector
   853 spaces. Type annotations may need to be added in some cases; potential
   854 INCOMPATIBILITY.
   855 
   856   Infinitesimal  :: ('a::real_normed_vector) star set"
   857   HFinite        :: ('a::real_normed_vector) star set"
   858   HInfinite      :: ('a::real_normed_vector) star set"
   859   approx         :: ('a::real_normed_vector) star => 'a star => bool
   860   monad          :: ('a::real_normed_vector) star => 'a star set
   861   galaxy         :: ('a::real_normed_vector) star => 'a star set
   862   (NS)LIMSEQ     :: [nat, 'a::real_normed_vector, 'a] => bool
   863   (NS)convergent :: (nat => 'a::real_normed_vector) => bool
   864   (NS)Bseq       :: (nat => 'a::real_normed_vector) => bool
   865   (NS)Cauchy     :: (nat => 'a::real_normed_vector) => bool
   866   (NS)LIM        :: ['a::real_normed_vector => 'b::real_normed_vector, 'a, 'b] => bool
   867   is(NS)Cont     :: ['a::real_normed_vector => 'b::real_normed_vector, 'a] => bool
   868   deriv          :: ['a::real_normed_field => 'a, 'a, 'a] => bool
   869 
   870 * Complex: Some complex-specific constants are now abbreviations for
   871 overloaded ones: complex_of_real = of_real, cmod = norm, hcmod =
   872 hnorm.  Other constants have been entirely removed in favor of the
   873 polymorphic versions (INCOMPATIBILITY):
   874 
   875   approx        <-- capprox
   876   HFinite       <-- CFinite
   877   HInfinite     <-- CInfinite
   878   Infinitesimal <-- CInfinitesimal
   879   monad         <-- cmonad
   880   galaxy        <-- cgalaxy
   881   (NS)LIM       <-- (NS)CLIM, (NS)CRLIM
   882   is(NS)Cont    <-- is(NS)Contc, is(NS)contCR
   883   (ns)deriv     <-- (ns)cderiv
   884 
   885 
   886 *** ML ***
   887 
   888 * ML within Isar: antiquotations allow to embed statically-checked
   889 formal entities in the source, referring to the context available at
   890 compile-time.  For example:
   891 
   892 ML {* @{typ "'a => 'b"} *}
   893 ML {* @{term "%x. x"} *}
   894 ML {* @{prop "x == y"} *}
   895 ML {* @{ctyp "'a => 'b"} *}
   896 ML {* @{cterm "%x. x"} *}
   897 ML {* @{cprop "x == y"} *}
   898 ML {* @{thm asm_rl} *}
   899 ML {* @{thms asm_rl} *}
   900 ML {* @{const_name c} *}
   901 ML {* @{const_syntax c} *}
   902 ML {* @{context} *}
   903 ML {* @{theory} *}
   904 ML {* @{theory Pure} *}
   905 ML {* @{simpset} *}
   906 ML {* @{claset} *}
   907 ML {* @{clasimpset} *}
   908 
   909 The same works for sources being ``used'' within an Isar context.
   910 
   911 * ML in Isar: improved error reporting; extra verbosity with
   912 Toplevel.debug enabled.
   913 
   914 * Pure/library:
   915 
   916   val burrow: ('a list -> 'b list) -> 'a list list -> 'b list list
   917   val fold_burrow: ('a list -> 'c -> 'b list * 'd) -> 'a list list -> 'c -> 'b list list * 'd
   918 
   919 The semantics of "burrow" is: "take a function with *simulatanously*
   920 transforms a list of value, and apply it *simulatanously* to a list of
   921 list of values of the appropriate type". Compare this with "map" which
   922 would *not* apply its argument function simulatanously but in
   923 sequence; "fold_burrow" has an additional context.
   924 
   925 * Pure/library: functions map2 and fold2 with curried syntax for
   926 simultanous mapping and folding:
   927 
   928     val map2: ('a -> 'b -> 'c) -> 'a list -> 'b list -> 'c list
   929     val fold2: ('a -> 'b -> 'c -> 'c) -> 'a list -> 'b list -> 'c -> 'c
   930 
   931 * Pure/library: indexed lists - some functions in the Isabelle library
   932 treating lists over 'a as finite mappings from [0...n] to 'a have been
   933 given more convenient names and signatures reminiscent of similar
   934 functions for alists, tables, etc:
   935 
   936   val nth: 'a list -> int -> 'a 
   937   val nth_map: int -> ('a -> 'a) -> 'a list -> 'a list
   938   val fold_index: (int * 'a -> 'b -> 'b) -> 'a list -> 'b -> 'b
   939 
   940 Note that fold_index starts counting at index 0, not 1 like foldln
   941 used to.
   942 
   943 * Pure/library: added general ``divide_and_conquer'' combinator on
   944 lists.
   945 
   946 * Pure/General/table.ML: the join operations now works via exceptions
   947 DUP/SAME instead of type option.  This is simpler in simple cases, and
   948 admits slightly more efficient complex applications.
   949 
   950 * Pure: datatype Context.generic joins theory/Proof.context and
   951 provides some facilities for code that works in either kind of
   952 context, notably GenericDataFun for uniform theory and proof data.
   953 
   954 * Pure: 'advanced' translation functions (parse_translation etc.) now
   955 use Context.generic instead of just theory.
   956 
   957 * Pure: simplified internal attribute type, which is now always
   958 Context.generic * thm -> Context.generic * thm.  Global (theory)
   959 vs. local (Proof.context) attributes have been discontinued, while
   960 minimizing code duplication.  Thm.rule_attribute and
   961 Thm.declaration_attribute build canonical attributes; see also
   962 structure Context for further operations on Context.generic, notably
   963 GenericDataFun.  INCOMPATIBILITY, need to adapt attribute type
   964 declarations and definitions.
   965 
   966 * Pure/kernel: consts certification ignores sort constraints given in
   967 signature declarations.  (This information is not relevant to the
   968 logic, but only for type inference.)  IMPORTANT INTERNAL CHANGE,
   969 potential INCOMPATIBILITY.
   970 
   971 * Pure: axiomatic type classes are now purely definitional, with
   972 explicit proofs of class axioms and super class relations performed
   973 internally.  See Pure/axclass.ML for the main internal interfaces --
   974 notably AxClass.define_class supercedes AxClass.add_axclass, and
   975 AxClass.axiomatize_class/classrel/arity supercede
   976 Sign.add_classes/classrel/arities.
   977 
   978 * Pure/Isar: Args/Attrib parsers operate on Context.generic --
   979 global/local versions on theory vs. Proof.context have been
   980 discontinued; Attrib.syntax and Method.syntax have been adapted
   981 accordingly.  INCOMPATIBILITY, need to adapt parser expressions for
   982 attributes, methods, etc.
   983 
   984 * Pure: several functions of signature "... -> theory -> theory * ..."
   985 have been reoriented to "... -> theory -> ... * theory" in order to
   986 allow natural usage in combination with the ||>, ||>>, |-> and
   987 fold_map combinators.
   988 
   989 * Pure: official theorem names (closed derivations) and additional
   990 comments (tags) are now strictly separate.  Name hints -- which are
   991 maintained as tags -- may be attached any time without affecting the
   992 derivation.
   993 
   994 * Pure: primitive rule lift_rule now takes goal cterm instead of an
   995 actual goal state (thm).  Use Thm.lift_rule (Thm.cprem_of st i) to
   996 achieve the old behaviour.
   997 
   998 * Pure: the "Goal" constant is now called "prop", supporting a
   999 slightly more general idea of ``protecting'' meta-level rule
  1000 statements.
  1001 
  1002 * Pure: Logic.(un)varify only works in a global context, which is now
  1003 enforced instead of silently assumed.  INCOMPATIBILITY, may use
  1004 Logic.legacy_(un)varify as temporary workaround.
  1005 
  1006 * Pure: structure Name provides scalable operations for generating
  1007 internal variable names, notably Name.variants etc.  This replaces
  1008 some popular functions from term.ML:
  1009 
  1010   Term.variant		->  Name.variant
  1011   Term.variantlist	->  Name.variant_list  (*canonical argument order*)
  1012   Term.invent_names	->  Name.invent_list
  1013 
  1014 Note that low-level renaming rarely occurs in new code -- operations
  1015 from structure Variable are used instead (see below).
  1016 
  1017 * Pure: structure Variable provides fundamental operations for proper
  1018 treatment of fixed/schematic variables in a context.  For example,
  1019 Variable.import introduces fixes for schematics of given facts and
  1020 Variable.export reverses the effect (up to renaming) -- this replaces
  1021 various freeze_thaw operations.
  1022 
  1023 * Pure: structure Goal provides simple interfaces for
  1024 init/conclude/finish and tactical prove operations (replacing former
  1025 Tactic.prove).  Goal.prove is the canonical way to prove results
  1026 within a given context; Goal.prove_global is a degraded version for
  1027 theory level goals, including a global Drule.standard.  Note that
  1028 OldGoals.prove_goalw_cterm has long been obsolete, since it is
  1029 ill-behaved in a local proof context (e.g. with local fixes/assumes or
  1030 in a locale context).
  1031 
  1032 * Isar: simplified treatment of user-level errors, using exception
  1033 ERROR of string uniformly.  Function error now merely raises ERROR,
  1034 without any side effect on output channels.  The Isar toplevel takes
  1035 care of proper display of ERROR exceptions.  ML code may use plain
  1036 handle/can/try; cat_error may be used to concatenate errors like this:
  1037 
  1038   ... handle ERROR msg => cat_error msg "..."
  1039 
  1040 Toplevel ML code (run directly or through the Isar toplevel) may be
  1041 embedded into the Isar toplevel with exception display/debug like
  1042 this:
  1043 
  1044   Isar.toplevel (fn () => ...)
  1045 
  1046 INCOMPATIBILITY, removed special transform_error facilities, removed
  1047 obsolete variants of user-level exceptions (ERROR_MESSAGE,
  1048 Context.PROOF, ProofContext.CONTEXT, Proof.STATE, ProofHistory.FAIL)
  1049 -- use plain ERROR instead.
  1050 
  1051 * Isar: theory setup now has type (theory -> theory), instead of a
  1052 list.  INCOMPATIBILITY, may use #> to compose setup functions.
  1053 
  1054 * Isar: installed ML toplevel pretty printer for type Proof.context,
  1055 subject to ProofContext.debug/verbose flags.
  1056 
  1057 * Isar: Toplevel.theory_to_proof admits transactions that modify the
  1058 theory before entering a proof state.  Transactions now always see a
  1059 quasi-functional intermediate checkpoint, both in interactive and
  1060 batch mode.
  1061 
  1062 * Simplifier: the simpset of a running simplification process now
  1063 contains a proof context (cf. Simplifier.the_context), which is the
  1064 very context that the initial simpset has been retrieved from (by
  1065 simpset_of/local_simpset_of).  Consequently, all plug-in components
  1066 (solver, looper etc.) may depend on arbitrary proof data.
  1067 
  1068 * Simplifier.inherit_context inherits the proof context (plus the
  1069 local bounds) of the current simplification process; any simproc
  1070 etc. that calls the Simplifier recursively should do this!  Removed
  1071 former Simplifier.inherit_bounds, which is already included here --
  1072 INCOMPATIBILITY.  Tools based on low-level rewriting may even have to
  1073 specify an explicit context using Simplifier.context/theory_context.
  1074 
  1075 * Simplifier/Classical Reasoner: more abstract interfaces
  1076 change_simpset/claset for modifying the simpset/claset reference of a
  1077 theory; raw versions simpset/claset_ref etc. have been discontinued --
  1078 INCOMPATIBILITY.
  1079 
  1080 * Provers: more generic wrt. syntax of object-logics, avoid hardwired
  1081 "Trueprop" etc.
  1082 
  1083 
  1084 *** System ***
  1085 
  1086 * settings: ML_IDENTIFIER -- which is appended to user specific heap
  1087 locations -- now includes the Isabelle version identifier as well.
  1088 This simplifies use of multiple Isabelle installations.
  1089 
  1090 * isabelle-process: option -S (secure mode) disables some critical
  1091 operations, notably runtime compilation and evaluation of ML source
  1092 code.
  1093 
  1094 
  1095 New in Isabelle2005 (October 2005)
  1096 ----------------------------------
  1097 
  1098 *** General ***
  1099 
  1100 * Theory headers: the new header syntax for Isar theories is
  1101 
  1102   theory <name>
  1103   imports <theory1> ... <theoryN>
  1104   uses <file1> ... <fileM>
  1105   begin
  1106 
  1107 where the 'uses' part is optional.  The previous syntax
  1108 
  1109   theory <name> = <theory1> + ... + <theoryN>:
  1110 
  1111 will disappear in the next release.  Use isatool fixheaders to convert
  1112 existing theory files.  Note that there is no change in ancient
  1113 non-Isar theories now, but these will disappear soon.
  1114 
  1115 * Theory loader: parent theories can now also be referred to via
  1116 relative and absolute paths.
  1117 
  1118 * Command 'find_theorems' searches for a list of criteria instead of a
  1119 list of constants. Known criteria are: intro, elim, dest, name:string,
  1120 simp:term, and any term. Criteria can be preceded by '-' to select
  1121 theorems that do not match. Intro, elim, dest select theorems that
  1122 match the current goal, name:s selects theorems whose fully qualified
  1123 name contain s, and simp:term selects all simplification rules whose
  1124 lhs match term.  Any other term is interpreted as pattern and selects
  1125 all theorems matching the pattern. Available in ProofGeneral under
  1126 'ProofGeneral -> Find Theorems' or C-c C-f.  Example:
  1127 
  1128   C-c C-f (100) "(_::nat) + _ + _" intro -name: "HOL."
  1129 
  1130 prints the last 100 theorems matching the pattern "(_::nat) + _ + _",
  1131 matching the current goal as introduction rule and not having "HOL."
  1132 in their name (i.e. not being defined in theory HOL).
  1133 
  1134 * Command 'thms_containing' has been discontinued in favour of
  1135 'find_theorems'; INCOMPATIBILITY.
  1136 
  1137 * Communication with Proof General is now 8bit clean, which means that
  1138 Unicode text in UTF-8 encoding may be used within theory texts (both
  1139 formal and informal parts).  Cf. option -U of the Isabelle Proof
  1140 General interface.  Here are some simple examples (cf. src/HOL/ex):
  1141 
  1142   http://isabelle.in.tum.de/library/HOL/ex/Hebrew.html
  1143   http://isabelle.in.tum.de/library/HOL/ex/Chinese.html
  1144 
  1145 * Improved efficiency of the Simplifier and, to a lesser degree, the
  1146 Classical Reasoner.  Typical big applications run around 2 times
  1147 faster.
  1148 
  1149 
  1150 *** Document preparation ***
  1151 
  1152 * Commands 'display_drafts' and 'print_drafts' perform simple output
  1153 of raw sources.  Only those symbols that do not require additional
  1154 LaTeX packages (depending on comments in isabellesym.sty) are
  1155 displayed properly, everything else is left verbatim.  isatool display
  1156 and isatool print are used as front ends (these are subject to the
  1157 DVI/PDF_VIEWER and PRINT_COMMAND settings, respectively).
  1158 
  1159 * Command tags control specific markup of certain regions of text,
  1160 notably folding and hiding.  Predefined tags include "theory" (for
  1161 theory begin and end), "proof" for proof commands, and "ML" for
  1162 commands involving ML code; the additional tags "visible" and
  1163 "invisible" are unused by default.  Users may give explicit tag
  1164 specifications in the text, e.g. ''by %invisible (auto)''.  The
  1165 interpretation of tags is determined by the LaTeX job during document
  1166 preparation: see option -V of isatool usedir, or options -n and -t of
  1167 isatool document, or even the LaTeX macros \isakeeptag, \isafoldtag,
  1168 \isadroptag.
  1169 
  1170 Several document versions may be produced at the same time via isatool
  1171 usedir (the generated index.html will link all of them).  Typical
  1172 specifications include ''-V document=theory,proof,ML'' to present
  1173 theory/proof/ML parts faithfully, ''-V outline=/proof,/ML'' to fold
  1174 proof and ML commands, and ''-V mutilated=-theory,-proof,-ML'' to omit
  1175 these parts without any formal replacement text.  The Isabelle site
  1176 default settings produce ''document'' and ''outline'' versions as
  1177 specified above.
  1178 
  1179 * Several new antiquotations:
  1180 
  1181   @{term_type term} prints a term with its type annotated;
  1182 
  1183   @{typeof term} prints the type of a term;
  1184 
  1185   @{const const} is the same as @{term const}, but checks that the
  1186   argument is a known logical constant;
  1187 
  1188   @{term_style style term} and @{thm_style style thm} print a term or
  1189   theorem applying a "style" to it
  1190 
  1191   @{ML text}
  1192 
  1193 Predefined styles are 'lhs' and 'rhs' printing the lhs/rhs of
  1194 definitions, equations, inequations etc., 'concl' printing only the
  1195 conclusion of a meta-logical statement theorem, and 'prem1' .. 'prem19'
  1196 to print the specified premise.  TermStyle.add_style provides an ML
  1197 interface for introducing further styles.  See also the "LaTeX Sugar"
  1198 document practical applications.  The ML antiquotation prints
  1199 type-checked ML expressions verbatim.
  1200 
  1201 * Markup commands 'chapter', 'section', 'subsection', 'subsubsection',
  1202 and 'text' support optional locale specification '(in loc)', which
  1203 specifies the default context for interpreting antiquotations.  For
  1204 example: 'text (in lattice) {* @{thm inf_assoc}*}'.
  1205 
  1206 * Option 'locale=NAME' of antiquotations specifies an alternative
  1207 context interpreting the subsequent argument.  For example: @{thm
  1208 [locale=lattice] inf_assoc}.
  1209 
  1210 * Proper output of proof terms (@{prf ...} and @{full_prf ...}) within
  1211 a proof context.
  1212 
  1213 * Proper output of antiquotations for theory commands involving a
  1214 proof context (such as 'locale' or 'theorem (in loc) ...').
  1215 
  1216 * Delimiters of outer tokens (string etc.) now produce separate LaTeX
  1217 macros (\isachardoublequoteopen, isachardoublequoteclose etc.).
  1218 
  1219 * isatool usedir: new option -C (default true) controls whether option
  1220 -D should include a copy of the original document directory; -C false
  1221 prevents unwanted effects such as copying of administrative CVS data.
  1222 
  1223 
  1224 *** Pure ***
  1225 
  1226 * Considerably improved version of 'constdefs' command.  Now performs
  1227 automatic type-inference of declared constants; additional support for
  1228 local structure declarations (cf. locales and HOL records), see also
  1229 isar-ref manual.  Potential INCOMPATIBILITY: need to observe strictly
  1230 sequential dependencies of definitions within a single 'constdefs'
  1231 section; moreover, the declared name needs to be an identifier.  If
  1232 all fails, consider to fall back on 'consts' and 'defs' separately.
  1233 
  1234 * Improved indexed syntax and implicit structures.  First of all,
  1235 indexed syntax provides a notational device for subscripted
  1236 application, using the new syntax \<^bsub>term\<^esub> for arbitrary
  1237 expressions.  Secondly, in a local context with structure
  1238 declarations, number indexes \<^sub>n or the empty index (default
  1239 number 1) refer to a certain fixed variable implicitly; option
  1240 show_structs controls printing of implicit structures.  Typical
  1241 applications of these concepts involve record types and locales.
  1242 
  1243 * New command 'no_syntax' removes grammar declarations (and
  1244 translations) resulting from the given syntax specification, which is
  1245 interpreted in the same manner as for the 'syntax' command.
  1246 
  1247 * 'Advanced' translation functions (parse_translation etc.) may depend
  1248 on the signature of the theory context being presently used for
  1249 parsing/printing, see also isar-ref manual.
  1250 
  1251 * Improved 'oracle' command provides a type-safe interface to turn an
  1252 ML expression of type theory -> T -> term into a primitive rule of
  1253 type theory -> T -> thm (i.e. the functionality of Thm.invoke_oracle
  1254 is already included here); see also FOL/ex/IffExample.thy;
  1255 INCOMPATIBILITY.
  1256 
  1257 * axclass: name space prefix for class "c" is now "c_class" (was "c"
  1258 before); "cI" is no longer bound, use "c.intro" instead.
  1259 INCOMPATIBILITY.  This change avoids clashes of fact bindings for
  1260 axclasses vs. locales.
  1261 
  1262 * Improved internal renaming of symbolic identifiers -- attach primes
  1263 instead of base 26 numbers.
  1264 
  1265 * New flag show_question_marks controls printing of leading question
  1266 marks in schematic variable names.
  1267 
  1268 * In schematic variable names, *any* symbol following \<^isub> or
  1269 \<^isup> is now treated as part of the base name.  For example, the
  1270 following works without printing of awkward ".0" indexes:
  1271 
  1272   lemma "x\<^isub>1 = x\<^isub>2 ==> x\<^isub>2 = x\<^isub>1"
  1273     by simp
  1274 
  1275 * Inner syntax includes (*(*nested*) comments*).
  1276 
  1277 * Pretty printer now supports unbreakable blocks, specified in mixfix
  1278 annotations as "(00...)".
  1279 
  1280 * Clear separation of logical types and nonterminals, where the latter
  1281 may only occur in 'syntax' specifications or type abbreviations.
  1282 Before that distinction was only partially implemented via type class
  1283 "logic" vs. "{}".  Potential INCOMPATIBILITY in rare cases of improper
  1284 use of 'types'/'consts' instead of 'nonterminals'/'syntax'.  Some very
  1285 exotic syntax specifications may require further adaption
  1286 (e.g. Cube/Cube.thy).
  1287 
  1288 * Removed obsolete type class "logic", use the top sort {} instead.
  1289 Note that non-logical types should be declared as 'nonterminals'
  1290 rather than 'types'.  INCOMPATIBILITY for new object-logic
  1291 specifications.
  1292 
  1293 * Attributes 'induct' and 'cases': type or set names may now be
  1294 locally fixed variables as well.
  1295 
  1296 * Simplifier: can now control the depth to which conditional rewriting
  1297 is traced via the PG menu Isabelle -> Settings -> Trace Simp Depth
  1298 Limit.
  1299 
  1300 * Simplifier: simplification procedures may now take the current
  1301 simpset into account (cf. Simplifier.simproc(_i) / mk_simproc
  1302 interface), which is very useful for calling the Simplifier
  1303 recursively.  Minor INCOMPATIBILITY: the 'prems' argument of simprocs
  1304 is gone -- use prems_of_ss on the simpset instead.  Moreover, the
  1305 low-level mk_simproc no longer applies Logic.varify internally, to
  1306 allow for use in a context of fixed variables.
  1307 
  1308 * thin_tac now works even if the assumption being deleted contains !!
  1309 or ==>.  More generally, erule now works even if the major premise of
  1310 the elimination rule contains !! or ==>.
  1311 
  1312 * Method 'rules' has been renamed to 'iprover'. INCOMPATIBILITY.
  1313 
  1314 * Reorganized bootstrapping of the Pure theories; CPure is now derived
  1315 from Pure, which contains all common declarations already.  Both
  1316 theories are defined via plain Isabelle/Isar .thy files.
  1317 INCOMPATIBILITY: elements of CPure (such as the CPure.intro /
  1318 CPure.elim / CPure.dest attributes) now appear in the Pure name space;
  1319 use isatool fixcpure to adapt your theory and ML sources.
  1320 
  1321 * New syntax 'name(i-j, i-, i, ...)' for referring to specific
  1322 selections of theorems in named facts via index ranges.
  1323 
  1324 * 'print_theorems': in theory mode, really print the difference
  1325 wrt. the last state (works for interactive theory development only),
  1326 in proof mode print all local facts (cf. 'print_facts');
  1327 
  1328 * 'hide': option '(open)' hides only base names.
  1329 
  1330 * More efficient treatment of intermediate checkpoints in interactive
  1331 theory development.
  1332 
  1333 * Code generator is now invoked via code_module (incremental code
  1334 generation) and code_library (modular code generation, ML structures
  1335 for each theory).  INCOMPATIBILITY: new keywords 'file' and 'contains'
  1336 must be quoted when used as identifiers.
  1337 
  1338 * New 'value' command for reading, evaluating and printing terms using
  1339 the code generator.  INCOMPATIBILITY: command keyword 'value' must be
  1340 quoted when used as identifier.
  1341 
  1342 
  1343 *** Locales ***
  1344 
  1345 * New commands for the interpretation of locale expressions in
  1346 theories (1), locales (2) and proof contexts (3).  These generate
  1347 proof obligations from the expression specification.  After the
  1348 obligations have been discharged, theorems of the expression are added
  1349 to the theory, target locale or proof context.  The synopsis of the
  1350 commands is a follows:
  1351 
  1352   (1) interpretation expr inst
  1353   (2) interpretation target < expr
  1354   (3) interpret expr inst
  1355 
  1356 Interpretation in theories and proof contexts require a parameter
  1357 instantiation of terms from the current context.  This is applied to
  1358 specifications and theorems of the interpreted expression.
  1359 Interpretation in locales only permits parameter renaming through the
  1360 locale expression.  Interpretation is smart in that interpretations
  1361 that are active already do not occur in proof obligations, neither are
  1362 instantiated theorems stored in duplicate.  Use 'print_interps' to
  1363 inspect active interpretations of a particular locale.  For details,
  1364 see the Isar Reference manual.  Examples can be found in
  1365 HOL/Finite_Set.thy and HOL/Algebra/UnivPoly.thy.
  1366 
  1367 INCOMPATIBILITY: former 'instantiate' has been withdrawn, use
  1368 'interpret' instead.
  1369 
  1370 * New context element 'constrains' for adding type constraints to
  1371 parameters.
  1372 
  1373 * Context expressions: renaming of parameters with syntax
  1374 redeclaration.
  1375 
  1376 * Locale declaration: 'includes' disallowed.
  1377 
  1378 * Proper static binding of attribute syntax -- i.e. types / terms /
  1379 facts mentioned as arguments are always those of the locale definition
  1380 context, independently of the context of later invocations.  Moreover,
  1381 locale operations (renaming and type / term instantiation) are applied
  1382 to attribute arguments as expected.
  1383 
  1384 INCOMPATIBILITY of the ML interface: always pass Attrib.src instead of
  1385 actual attributes; rare situations may require Attrib.attribute to
  1386 embed those attributes into Attrib.src that lack concrete syntax.
  1387 Attribute implementations need to cooperate properly with the static
  1388 binding mechanism.  Basic parsers Args.XXX_typ/term/prop and
  1389 Attrib.XXX_thm etc. already do the right thing without further
  1390 intervention.  Only unusual applications -- such as "where" or "of"
  1391 (cf. src/Pure/Isar/attrib.ML), which process arguments depending both
  1392 on the context and the facts involved -- may have to assign parsed
  1393 values to argument tokens explicitly.
  1394 
  1395 * Changed parameter management in theorem generation for long goal
  1396 statements with 'includes'.  INCOMPATIBILITY: produces a different
  1397 theorem statement in rare situations.
  1398 
  1399 * Locale inspection command 'print_locale' omits notes elements.  Use
  1400 'print_locale!' to have them included in the output.
  1401 
  1402 
  1403 *** Provers ***
  1404 
  1405 * Provers/hypsubst.ML: improved version of the subst method, for
  1406 single-step rewriting: it now works in bound variable contexts. New is
  1407 'subst (asm)', for rewriting an assumption.  INCOMPATIBILITY: may
  1408 rewrite a different subterm than the original subst method, which is
  1409 still available as 'simplesubst'.
  1410 
  1411 * Provers/quasi.ML: new transitivity reasoners for transitivity only
  1412 and quasi orders.
  1413 
  1414 * Provers/trancl.ML: new transitivity reasoner for transitive and
  1415 reflexive-transitive closure of relations.
  1416 
  1417 * Provers/blast.ML: new reference depth_limit to make blast's depth
  1418 limit (previously hard-coded with a value of 20) user-definable.
  1419 
  1420 * Provers/simplifier.ML has been moved to Pure, where Simplifier.setup
  1421 is peformed already.  Object-logics merely need to finish their
  1422 initial simpset configuration as before.  INCOMPATIBILITY.
  1423 
  1424 
  1425 *** HOL ***
  1426 
  1427 * Symbolic syntax of Hilbert Choice Operator is now as follows:
  1428 
  1429   syntax (epsilon)
  1430     "_Eps" :: "[pttrn, bool] => 'a"    ("(3\<some>_./ _)" [0, 10] 10)
  1431 
  1432 The symbol \<some> is displayed as the alternative epsilon of LaTeX
  1433 and x-symbol; use option '-m epsilon' to get it actually printed.
  1434 Moreover, the mathematically important symbolic identifier \<epsilon>
  1435 becomes available as variable, constant etc.  INCOMPATIBILITY,
  1436 
  1437 * "x > y" abbreviates "y < x" and "x >= y" abbreviates "y <= x".
  1438 Similarly for all quantifiers: "ALL x > y" etc.  The x-symbol for >=
  1439 is \<ge>. New transitivity rules have been added to HOL/Orderings.thy to
  1440 support corresponding Isar calculations.
  1441 
  1442 * "{x:A. P}" abbreviates "{x. x:A & P}", and similarly for "\<in>"
  1443 instead of ":".
  1444 
  1445 * theory SetInterval: changed the syntax for open intervals:
  1446 
  1447   Old       New
  1448   {..n(}    {..<n}
  1449   {)n..}    {n<..}
  1450   {m..n(}   {m..<n}
  1451   {)m..n}   {m<..n}
  1452   {)m..n(}  {m<..<n}
  1453 
  1454 The old syntax is still supported but will disappear in the next
  1455 release.  For conversion use the following Emacs search and replace
  1456 patterns (these are not perfect but work quite well):
  1457 
  1458   {)\([^\.]*\)\.\.  ->  {\1<\.\.}
  1459   \.\.\([^(}]*\)(}  ->  \.\.<\1}
  1460 
  1461 * Theory Commutative_Ring (in Library): method comm_ring for proving
  1462 equalities in commutative rings; method 'algebra' provides a generic
  1463 interface.
  1464 
  1465 * Theory Finite_Set: changed the syntax for 'setsum', summation over
  1466 finite sets: "setsum (%x. e) A", which used to be "\<Sum>x:A. e", is
  1467 now either "SUM x:A. e" or "\<Sum>x \<in> A. e". The bound variable can
  1468 be a tuple pattern.
  1469 
  1470 Some new syntax forms are available:
  1471 
  1472   "\<Sum>x | P. e"      for     "setsum (%x. e) {x. P}"
  1473   "\<Sum>x = a..b. e"   for     "setsum (%x. e) {a..b}"
  1474   "\<Sum>x = a..<b. e"  for     "setsum (%x. e) {a..<b}"
  1475   "\<Sum>x < k. e"      for     "setsum (%x. e) {..<k}"
  1476 
  1477 The latter form "\<Sum>x < k. e" used to be based on a separate
  1478 function "Summation", which has been discontinued.
  1479 
  1480 * theory Finite_Set: in structured induction proofs, the insert case
  1481 is now 'case (insert x F)' instead of the old counterintuitive 'case
  1482 (insert F x)'.
  1483 
  1484 * The 'refute' command has been extended to support a much larger
  1485 fragment of HOL, including axiomatic type classes, constdefs and
  1486 typedefs, inductive datatypes and recursion.
  1487 
  1488 * New tactics 'sat' and 'satx' to prove propositional tautologies.
  1489 Requires zChaff with proof generation to be installed.  See
  1490 HOL/ex/SAT_Examples.thy for examples.
  1491 
  1492 * Datatype induction via method 'induct' now preserves the name of the
  1493 induction variable. For example, when proving P(xs::'a list) by
  1494 induction on xs, the induction step is now P(xs) ==> P(a#xs) rather
  1495 than P(list) ==> P(a#list) as previously.  Potential INCOMPATIBILITY
  1496 in unstructured proof scripts.
  1497 
  1498 * Reworked implementation of records.  Improved scalability for
  1499 records with many fields, avoiding performance problems for type
  1500 inference. Records are no longer composed of nested field types, but
  1501 of nested extension types. Therefore the record type only grows linear
  1502 in the number of extensions and not in the number of fields.  The
  1503 top-level (users) view on records is preserved.  Potential
  1504 INCOMPATIBILITY only in strange cases, where the theory depends on the
  1505 old record representation. The type generated for a record is called
  1506 <record_name>_ext_type.
  1507 
  1508 Flag record_quick_and_dirty_sensitive can be enabled to skip the
  1509 proofs triggered by a record definition or a simproc (if
  1510 quick_and_dirty is enabled).  Definitions of large records can take
  1511 quite long.
  1512 
  1513 New simproc record_upd_simproc for simplification of multiple record
  1514 updates enabled by default.  Moreover, trivial updates are also
  1515 removed: r(|x := x r|) = r.  INCOMPATIBILITY: old proofs break
  1516 occasionally, since simplification is more powerful by default.
  1517 
  1518 * typedef: proper support for polymorphic sets, which contain extra
  1519 type-variables in the term.
  1520 
  1521 * Simplifier: automatically reasons about transitivity chains
  1522 involving "trancl" (r^+) and "rtrancl" (r^*) by setting up tactics
  1523 provided by Provers/trancl.ML as additional solvers.  INCOMPATIBILITY:
  1524 old proofs break occasionally as simplification may now solve more
  1525 goals than previously.
  1526 
  1527 * Simplifier: converts x <= y into x = y if assumption y <= x is
  1528 present.  Works for all partial orders (class "order"), in particular
  1529 numbers and sets.  For linear orders (e.g. numbers) it treats ~ x < y
  1530 just like y <= x.
  1531 
  1532 * Simplifier: new simproc for "let x = a in f x".  If a is a free or
  1533 bound variable or a constant then the let is unfolded.  Otherwise
  1534 first a is simplified to b, and then f b is simplified to g. If
  1535 possible we abstract b from g arriving at "let x = b in h x",
  1536 otherwise we unfold the let and arrive at g.  The simproc can be
  1537 enabled/disabled by the reference use_let_simproc.  Potential
  1538 INCOMPATIBILITY since simplification is more powerful by default.
  1539 
  1540 * Classical reasoning: the meson method now accepts theorems as arguments.
  1541 
  1542 * Prover support: pre-release of the Isabelle-ATP linkup, which runs background
  1543 jobs to provide advice on the provability of subgoals.
  1544 
  1545 * Theory OrderedGroup and Ring_and_Field: various additions and
  1546 improvements to faciliate calculations involving equalities and
  1547 inequalities.
  1548 
  1549 The following theorems have been eliminated or modified
  1550 (INCOMPATIBILITY):
  1551 
  1552   abs_eq             now named abs_of_nonneg
  1553   abs_of_ge_0        now named abs_of_nonneg
  1554   abs_minus_eq       now named abs_of_nonpos
  1555   imp_abs_id         now named abs_of_nonneg
  1556   imp_abs_neg_id     now named abs_of_nonpos
  1557   mult_pos           now named mult_pos_pos
  1558   mult_pos_le        now named mult_nonneg_nonneg
  1559   mult_pos_neg_le    now named mult_nonneg_nonpos
  1560   mult_pos_neg2_le   now named mult_nonneg_nonpos2
  1561   mult_neg           now named mult_neg_neg
  1562   mult_neg_le        now named mult_nonpos_nonpos
  1563 
  1564 * Theory Parity: added rules for simplifying exponents.
  1565 
  1566 * Theory List:
  1567 
  1568 The following theorems have been eliminated or modified
  1569 (INCOMPATIBILITY):
  1570 
  1571   list_all_Nil       now named list_all.simps(1)
  1572   list_all_Cons      now named list_all.simps(2)
  1573   list_all_conv      now named list_all_iff
  1574   set_mem_eq         now named mem_iff
  1575 
  1576 * Theories SetsAndFunctions and BigO (see HOL/Library) support
  1577 asymptotic "big O" calculations.  See the notes in BigO.thy.
  1578 
  1579 
  1580 *** HOL-Complex ***
  1581 
  1582 * Theory RealDef: better support for embedding natural numbers and
  1583 integers in the reals.
  1584 
  1585 The following theorems have been eliminated or modified
  1586 (INCOMPATIBILITY):
  1587 
  1588   exp_ge_add_one_self  now requires no hypotheses
  1589   real_of_int_add      reversed direction of equality (use [symmetric])
  1590   real_of_int_minus    reversed direction of equality (use [symmetric])
  1591   real_of_int_diff     reversed direction of equality (use [symmetric])
  1592   real_of_int_mult     reversed direction of equality (use [symmetric])
  1593 
  1594 * Theory RComplete: expanded support for floor and ceiling functions.
  1595 
  1596 * Theory Ln is new, with properties of the natural logarithm
  1597 
  1598 * Hyperreal: There is a new type constructor "star" for making
  1599 nonstandard types.  The old type names are now type synonyms:
  1600 
  1601   hypreal = real star
  1602   hypnat = nat star
  1603   hcomplex = complex star
  1604 
  1605 * Hyperreal: Many groups of similarly-defined constants have been
  1606 replaced by polymorphic versions (INCOMPATIBILITY):
  1607 
  1608   star_of <-- hypreal_of_real, hypnat_of_nat, hcomplex_of_complex
  1609 
  1610   starset      <-- starsetNat, starsetC
  1611   *s*          <-- *sNat*, *sc*
  1612   starset_n    <-- starsetNat_n, starsetC_n
  1613   *sn*         <-- *sNatn*, *scn*
  1614   InternalSets <-- InternalNatSets, InternalCSets
  1615 
  1616   starfun      <-- starfun{Nat,Nat2,C,RC,CR}
  1617   *f*          <-- *fNat*, *fNat2*, *fc*, *fRc*, *fcR*
  1618   starfun_n    <-- starfun{Nat,Nat2,C,RC,CR}_n
  1619   *fn*         <-- *fNatn*, *fNat2n*, *fcn*, *fRcn*, *fcRn*
  1620   InternalFuns <-- InternalNatFuns, InternalNatFuns2, Internal{C,RC,CR}Funs
  1621 
  1622 * Hyperreal: Many type-specific theorems have been removed in favor of
  1623 theorems specific to various axiomatic type classes (INCOMPATIBILITY):
  1624 
  1625   add_commute <-- {hypreal,hypnat,hcomplex}_add_commute
  1626   add_assoc   <-- {hypreal,hypnat,hcomplex}_add_assocs
  1627   OrderedGroup.add_0 <-- {hypreal,hypnat,hcomplex}_add_zero_left
  1628   OrderedGroup.add_0_right <-- {hypreal,hcomplex}_add_zero_right
  1629   right_minus <-- hypreal_add_minus
  1630   left_minus <-- {hypreal,hcomplex}_add_minus_left
  1631   mult_commute <-- {hypreal,hypnat,hcomplex}_mult_commute
  1632   mult_assoc <-- {hypreal,hypnat,hcomplex}_mult_assoc
  1633   mult_1_left <-- {hypreal,hypnat}_mult_1, hcomplex_mult_one_left
  1634   mult_1_right <-- hcomplex_mult_one_right
  1635   mult_zero_left <-- hcomplex_mult_zero_left
  1636   left_distrib <-- {hypreal,hypnat,hcomplex}_add_mult_distrib
  1637   right_distrib <-- hypnat_add_mult_distrib2
  1638   zero_neq_one <-- {hypreal,hypnat,hcomplex}_zero_not_eq_one
  1639   right_inverse <-- hypreal_mult_inverse
  1640   left_inverse <-- hypreal_mult_inverse_left, hcomplex_mult_inv_left
  1641   order_refl <-- {hypreal,hypnat}_le_refl
  1642   order_trans <-- {hypreal,hypnat}_le_trans
  1643   order_antisym <-- {hypreal,hypnat}_le_anti_sym
  1644   order_less_le <-- {hypreal,hypnat}_less_le
  1645   linorder_linear <-- {hypreal,hypnat}_le_linear
  1646   add_left_mono <-- {hypreal,hypnat}_add_left_mono
  1647   mult_strict_left_mono <-- {hypreal,hypnat}_mult_less_mono2
  1648   add_nonneg_nonneg <-- hypreal_le_add_order
  1649 
  1650 * Hyperreal: Separate theorems having to do with type-specific
  1651 versions of constants have been merged into theorems that apply to the
  1652 new polymorphic constants (INCOMPATIBILITY):
  1653 
  1654   STAR_UNIV_set <-- {STAR_real,NatStar_real,STARC_complex}_set
  1655   STAR_empty_set <-- {STAR,NatStar,STARC}_empty_set
  1656   STAR_Un <-- {STAR,NatStar,STARC}_Un
  1657   STAR_Int <-- {STAR,NatStar,STARC}_Int
  1658   STAR_Compl <-- {STAR,NatStar,STARC}_Compl
  1659   STAR_subset <-- {STAR,NatStar,STARC}_subset
  1660   STAR_mem <-- {STAR,NatStar,STARC}_mem
  1661   STAR_mem_Compl <-- {STAR,STARC}_mem_Compl
  1662   STAR_diff <-- {STAR,STARC}_diff
  1663   STAR_star_of_image_subset <-- {STAR_hypreal_of_real, NatStar_hypreal_of_real,
  1664     STARC_hcomplex_of_complex}_image_subset
  1665   starset_n_Un <-- starset{Nat,C}_n_Un
  1666   starset_n_Int <-- starset{Nat,C}_n_Int
  1667   starset_n_Compl <-- starset{Nat,C}_n_Compl
  1668   starset_n_diff <-- starset{Nat,C}_n_diff
  1669   InternalSets_Un <-- Internal{Nat,C}Sets_Un
  1670   InternalSets_Int <-- Internal{Nat,C}Sets_Int
  1671   InternalSets_Compl <-- Internal{Nat,C}Sets_Compl
  1672   InternalSets_diff <-- Internal{Nat,C}Sets_diff
  1673   InternalSets_UNIV_diff <-- Internal{Nat,C}Sets_UNIV_diff
  1674   InternalSets_starset_n <-- Internal{Nat,C}Sets_starset{Nat,C}_n
  1675   starset_starset_n_eq <-- starset{Nat,C}_starset{Nat,C}_n_eq
  1676   starset_n_starset <-- starset{Nat,C}_n_starset{Nat,C}
  1677   starfun_n_starfun <-- starfun{Nat,Nat2,C,RC,CR}_n_starfun{Nat,Nat2,C,RC,CR}
  1678   starfun <-- starfun{Nat,Nat2,C,RC,CR}
  1679   starfun_mult <-- starfun{Nat,Nat2,C,RC,CR}_mult
  1680   starfun_add <-- starfun{Nat,Nat2,C,RC,CR}_add
  1681   starfun_minus <-- starfun{Nat,Nat2,C,RC,CR}_minus
  1682   starfun_diff <-- starfun{C,RC,CR}_diff
  1683   starfun_o <-- starfun{NatNat2,Nat2,_stafunNat,C,C_starfunRC,_starfunCR}_o
  1684   starfun_o2 <-- starfun{NatNat2,_stafunNat,C,C_starfunRC,_starfunCR}_o2
  1685   starfun_const_fun <-- starfun{Nat,Nat2,C,RC,CR}_const_fun
  1686   starfun_inverse <-- starfun{Nat,C,RC,CR}_inverse
  1687   starfun_eq <-- starfun{Nat,Nat2,C,RC,CR}_eq
  1688   starfun_eq_iff <-- starfun{C,RC,CR}_eq_iff
  1689   starfun_Id <-- starfunC_Id
  1690   starfun_approx <-- starfun{Nat,CR}_approx
  1691   starfun_capprox <-- starfun{C,RC}_capprox
  1692   starfun_abs <-- starfunNat_rabs
  1693   starfun_lambda_cancel <-- starfun{C,CR,RC}_lambda_cancel
  1694   starfun_lambda_cancel2 <-- starfun{C,CR,RC}_lambda_cancel2
  1695   starfun_mult_HFinite_approx <-- starfunCR_mult_HFinite_capprox
  1696   starfun_mult_CFinite_capprox <-- starfun{C,RC}_mult_CFinite_capprox
  1697   starfun_add_capprox <-- starfun{C,RC}_add_capprox
  1698   starfun_add_approx <-- starfunCR_add_approx
  1699   starfun_inverse_inverse <-- starfunC_inverse_inverse
  1700   starfun_divide <-- starfun{C,CR,RC}_divide
  1701   starfun_n <-- starfun{Nat,C}_n
  1702   starfun_n_mult <-- starfun{Nat,C}_n_mult
  1703   starfun_n_add <-- starfun{Nat,C}_n_add
  1704   starfun_n_add_minus <-- starfunNat_n_add_minus
  1705   starfun_n_const_fun <-- starfun{Nat,C}_n_const_fun
  1706   starfun_n_minus <-- starfun{Nat,C}_n_minus
  1707   starfun_n_eq <-- starfun{Nat,C}_n_eq
  1708 
  1709   star_n_add <-- {hypreal,hypnat,hcomplex}_add
  1710   star_n_minus <-- {hypreal,hcomplex}_minus
  1711   star_n_diff <-- {hypreal,hcomplex}_diff
  1712   star_n_mult <-- {hypreal,hcomplex}_mult
  1713   star_n_inverse <-- {hypreal,hcomplex}_inverse
  1714   star_n_le <-- {hypreal,hypnat}_le
  1715   star_n_less <-- {hypreal,hypnat}_less
  1716   star_n_zero_num <-- {hypreal,hypnat,hcomplex}_zero_num
  1717   star_n_one_num <-- {hypreal,hypnat,hcomplex}_one_num
  1718   star_n_abs <-- hypreal_hrabs
  1719   star_n_divide <-- hcomplex_divide
  1720 
  1721   star_of_add <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_add
  1722   star_of_minus <-- {hypreal_of_real,hcomplex_of_complex}_minus
  1723   star_of_diff <-- hypreal_of_real_diff
  1724   star_of_mult <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_mult
  1725   star_of_one <-- {hypreal_of_real,hcomplex_of_complex}_one
  1726   star_of_zero <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_zero
  1727   star_of_le <-- {hypreal_of_real,hypnat_of_nat}_le_iff
  1728   star_of_less <-- {hypreal_of_real,hypnat_of_nat}_less_iff
  1729   star_of_eq <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_eq_iff
  1730   star_of_inverse <-- {hypreal_of_real,hcomplex_of_complex}_inverse
  1731   star_of_divide <-- {hypreal_of_real,hcomplex_of_complex}_divide
  1732   star_of_of_nat <-- {hypreal_of_real,hcomplex_of_complex}_of_nat
  1733   star_of_of_int <-- {hypreal_of_real,hcomplex_of_complex}_of_int
  1734   star_of_number_of <-- {hypreal,hcomplex}_number_of
  1735   star_of_number_less <-- number_of_less_hypreal_of_real_iff
  1736   star_of_number_le <-- number_of_le_hypreal_of_real_iff
  1737   star_of_eq_number <-- hypreal_of_real_eq_number_of_iff
  1738   star_of_less_number <-- hypreal_of_real_less_number_of_iff
  1739   star_of_le_number <-- hypreal_of_real_le_number_of_iff
  1740   star_of_power <-- hypreal_of_real_power
  1741   star_of_eq_0 <-- hcomplex_of_complex_zero_iff
  1742 
  1743 * Hyperreal: new method "transfer" that implements the transfer
  1744 principle of nonstandard analysis. With a subgoal that mentions
  1745 nonstandard types like "'a star", the command "apply transfer"
  1746 replaces it with an equivalent one that mentions only standard types.
  1747 To be successful, all free variables must have standard types; non-
  1748 standard variables must have explicit universal quantifiers.
  1749 
  1750 * Hyperreal: A theory of Taylor series.
  1751 
  1752 
  1753 *** HOLCF ***
  1754 
  1755 * Discontinued special version of 'constdefs' (which used to support
  1756 continuous functions) in favor of the general Pure one with full
  1757 type-inference.
  1758 
  1759 * New simplification procedure for solving continuity conditions; it
  1760 is much faster on terms with many nested lambda abstractions (cubic
  1761 instead of exponential time).
  1762 
  1763 * New syntax for domain package: selector names are now optional.
  1764 Parentheses should be omitted unless argument is lazy, for example:
  1765 
  1766   domain 'a stream = cons "'a" (lazy "'a stream")
  1767 
  1768 * New command 'fixrec' for defining recursive functions with pattern
  1769 matching; defining multiple functions with mutual recursion is also
  1770 supported.  Patterns may include the constants cpair, spair, up, sinl,
  1771 sinr, or any data constructor defined by the domain package. The given
  1772 equations are proven as rewrite rules. See HOLCF/ex/Fixrec_ex.thy for
  1773 syntax and examples.
  1774 
  1775 * New commands 'cpodef' and 'pcpodef' for defining predicate subtypes
  1776 of cpo and pcpo types. Syntax is exactly like the 'typedef' command,
  1777 but the proof obligation additionally includes an admissibility
  1778 requirement. The packages generate instances of class cpo or pcpo,
  1779 with continuity and strictness theorems for Rep and Abs.
  1780 
  1781 * HOLCF: Many theorems have been renamed according to a more standard naming
  1782 scheme (INCOMPATIBILITY):
  1783 
  1784   foo_inject:  "foo$x = foo$y ==> x = y"
  1785   foo_eq:      "(foo$x = foo$y) = (x = y)"
  1786   foo_less:    "(foo$x << foo$y) = (x << y)"
  1787   foo_strict:  "foo$UU = UU"
  1788   foo_defined: "... ==> foo$x ~= UU"
  1789   foo_defined_iff: "(foo$x = UU) = (x = UU)"
  1790 
  1791 
  1792 *** ZF ***
  1793 
  1794 * ZF/ex: theories Group and Ring provide examples in abstract algebra,
  1795 including the First Isomorphism Theorem (on quotienting by the kernel
  1796 of a homomorphism).
  1797 
  1798 * ZF/Simplifier: install second copy of type solver that actually
  1799 makes use of TC rules declared to Isar proof contexts (or locales);
  1800 the old version is still required for ML proof scripts.
  1801 
  1802 
  1803 *** Cube ***
  1804 
  1805 * Converted to Isar theory format; use locales instead of axiomatic
  1806 theories.
  1807 
  1808 
  1809 *** ML ***
  1810 
  1811 * Pure/library.ML: added ##>, ##>>, #>> -- higher-order counterparts
  1812 for ||>, ||>>, |>>,
  1813 
  1814 * Pure/library.ML no longer defines its own option datatype, but uses
  1815 that of the SML basis, which has constructors NONE and SOME instead of
  1816 None and Some, as well as exception Option.Option instead of OPTION.
  1817 The functions the, if_none, is_some, is_none have been adapted
  1818 accordingly, while Option.map replaces apsome.
  1819 
  1820 * Pure/library.ML: the exception LIST has been given up in favour of
  1821 the standard exceptions Empty and Subscript, as well as
  1822 Library.UnequalLengths.  Function like Library.hd and Library.tl are
  1823 superceded by the standard hd and tl functions etc.
  1824 
  1825 A number of basic list functions are no longer exported to the ML
  1826 toplevel, as they are variants of predefined functions.  The following
  1827 suggests how one can translate existing code:
  1828 
  1829     rev_append xs ys = List.revAppend (xs, ys)
  1830     nth_elem (i, xs) = List.nth (xs, i)
  1831     last_elem xs = List.last xs
  1832     flat xss = List.concat xss
  1833     seq fs = List.app fs
  1834     partition P xs = List.partition P xs
  1835     mapfilter f xs = List.mapPartial f xs
  1836 
  1837 * Pure/library.ML: several combinators for linear functional
  1838 transformations, notably reverse application and composition:
  1839 
  1840   x |> f                f #> g
  1841   (x, y) |-> f          f #-> g
  1842 
  1843 * Pure/library.ML: introduced/changed precedence of infix operators:
  1844 
  1845   infix 1 |> |-> ||> ||>> |>> |>>> #> #->;
  1846   infix 2 ?;
  1847   infix 3 o oo ooo oooo;
  1848   infix 4 ~~ upto downto;
  1849 
  1850 Maybe INCOMPATIBILITY when any of those is used in conjunction with other
  1851 infix operators.
  1852 
  1853 * Pure/library.ML: natural list combinators fold, fold_rev, and
  1854 fold_map support linear functional transformations and nesting.  For
  1855 example:
  1856 
  1857   fold f [x1, ..., xN] y =
  1858     y |> f x1 |> ... |> f xN
  1859 
  1860   (fold o fold) f [xs1, ..., xsN] y =
  1861     y |> fold f xs1 |> ... |> fold f xsN
  1862 
  1863   fold f [x1, ..., xN] =
  1864     f x1 #> ... #> f xN
  1865 
  1866   (fold o fold) f [xs1, ..., xsN] =
  1867     fold f xs1 #> ... #> fold f xsN
  1868 
  1869 * Pure/library.ML: the following selectors on type 'a option are
  1870 available:
  1871 
  1872   the:               'a option -> 'a  (*partial*)
  1873   these:             'a option -> 'a  where 'a = 'b list
  1874   the_default: 'a -> 'a option -> 'a
  1875   the_list:          'a option -> 'a list
  1876 
  1877 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  1878 basic operations for association lists, following natural argument
  1879 order; moreover the explicit equality predicate passed here avoids
  1880 potentially expensive polymorphic runtime equality checks.
  1881 The old functions may be expressed as follows:
  1882 
  1883   assoc = uncurry (AList.lookup (op =))
  1884   assocs = these oo AList.lookup (op =)
  1885   overwrite = uncurry (AList.update (op =)) o swap
  1886 
  1887 * Pure/General: structure AList (cf. Pure/General/alist.ML) provides
  1888 
  1889   val make: ('a -> 'b) -> 'a list -> ('a * 'b) list
  1890   val find: ('a * 'b -> bool) -> ('c * 'b) list -> 'a -> 'c list
  1891 
  1892 replacing make_keylist and keyfilter (occassionally used)
  1893 Naive rewrites:
  1894 
  1895   make_keylist = AList.make
  1896   keyfilter = AList.find (op =)
  1897 
  1898 * eq_fst and eq_snd now take explicit equality parameter, thus
  1899   avoiding eqtypes. Naive rewrites:
  1900 
  1901     eq_fst = eq_fst (op =)
  1902     eq_snd = eq_snd (op =)
  1903 
  1904 * Removed deprecated apl and apr (rarely used).
  1905   Naive rewrites:
  1906 
  1907     apl (n, op) =>>= curry op n
  1908     apr (op, m) =>>= fn n => op (n, m)
  1909 
  1910 * Pure/General: structure OrdList (cf. Pure/General/ord_list.ML)
  1911 provides a reasonably efficient light-weight implementation of sets as
  1912 lists.
  1913 
  1914 * Pure/General: generic tables (cf. Pure/General/table.ML) provide a
  1915 few new operations; existing lookup and update are now curried to
  1916 follow natural argument order (for use with fold etc.);
  1917 INCOMPATIBILITY, use (uncurry Symtab.lookup) etc. as last resort.
  1918 
  1919 * Pure/General: output via the Isabelle channels of
  1920 writeln/warning/error etc. is now passed through Output.output, with a
  1921 hook for arbitrary transformations depending on the print_mode
  1922 (cf. Output.add_mode -- the first active mode that provides a output
  1923 function wins).  Already formatted output may be embedded into further
  1924 text via Output.raw; the result of Pretty.string_of/str_of and derived
  1925 functions (string_of_term/cterm/thm etc.) is already marked raw to
  1926 accommodate easy composition of diagnostic messages etc.  Programmers
  1927 rarely need to care about Output.output or Output.raw at all, with
  1928 some notable exceptions: Output.output is required when bypassing the
  1929 standard channels (writeln etc.), or in token translations to produce
  1930 properly formatted results; Output.raw is required when capturing
  1931 already output material that will eventually be presented to the user
  1932 a second time.  For the default print mode, both Output.output and
  1933 Output.raw have no effect.
  1934 
  1935 * Pure/General: Output.time_accumulator NAME creates an operator ('a
  1936 -> 'b) -> 'a -> 'b to measure runtime and count invocations; the
  1937 cumulative results are displayed at the end of a batch session.
  1938 
  1939 * Pure/General: File.sysify_path and File.quote_sysify path have been
  1940 replaced by File.platform_path and File.shell_path (with appropriate
  1941 hooks).  This provides a clean interface for unusual systems where the
  1942 internal and external process view of file names are different.
  1943 
  1944 * Pure: more efficient orders for basic syntactic entities: added
  1945 fast_string_ord, fast_indexname_ord, fast_term_ord; changed sort_ord
  1946 and typ_ord to use fast_string_ord and fast_indexname_ord (term_ord is
  1947 NOT affected); structures Symtab, Vartab, Typtab, Termtab use the fast
  1948 orders now -- potential INCOMPATIBILITY for code that depends on a
  1949 particular order for Symtab.keys, Symtab.dest, etc. (consider using
  1950 Library.sort_strings on result).
  1951 
  1952 * Pure/term.ML: combinators fold_atyps, fold_aterms, fold_term_types,
  1953 fold_types traverse types/terms from left to right, observing natural
  1954 argument order.  Supercedes previous foldl_XXX versions, add_frees,
  1955 add_vars etc. have been adapted as well: INCOMPATIBILITY.
  1956 
  1957 * Pure: name spaces have been refined, with significant changes of the
  1958 internal interfaces -- INCOMPATIBILITY.  Renamed cond_extern(_table)
  1959 to extern(_table).  The plain name entry path is superceded by a
  1960 general 'naming' context, which also includes the 'policy' to produce
  1961 a fully qualified name and external accesses of a fully qualified
  1962 name; NameSpace.extend is superceded by context dependent
  1963 Sign.declare_name.  Several theory and proof context operations modify
  1964 the naming context.  Especially note Theory.restore_naming and
  1965 ProofContext.restore_naming to get back to a sane state; note that
  1966 Theory.add_path is no longer sufficient to recover from
  1967 Theory.absolute_path in particular.
  1968 
  1969 * Pure: new flags short_names (default false) and unique_names
  1970 (default true) for controlling output of qualified names.  If
  1971 short_names is set, names are printed unqualified.  If unique_names is
  1972 reset, the name prefix is reduced to the minimum required to achieve
  1973 the original result when interning again, even if there is an overlap
  1974 with earlier declarations.
  1975 
  1976 * Pure/TheoryDataFun: change of the argument structure; 'prep_ext' is
  1977 now 'extend', and 'merge' gets an additional Pretty.pp argument
  1978 (useful for printing error messages).  INCOMPATIBILITY.
  1979 
  1980 * Pure: major reorganization of the theory context.  Type Sign.sg and
  1981 Theory.theory are now identified, referring to the universal
  1982 Context.theory (see Pure/context.ML).  Actual signature and theory
  1983 content is managed as theory data.  The old code and interfaces were
  1984 spread over many files and structures; the new arrangement introduces
  1985 considerable INCOMPATIBILITY to gain more clarity:
  1986 
  1987   Context -- theory management operations (name, identity, inclusion,
  1988     parents, ancestors, merge, etc.), plus generic theory data;
  1989 
  1990   Sign -- logical signature and syntax operations (declaring consts,
  1991     types, etc.), plus certify/read for common entities;
  1992 
  1993   Theory -- logical theory operations (stating axioms, definitions,
  1994     oracles), plus a copy of logical signature operations (consts,
  1995     types, etc.); also a few basic management operations (Theory.copy,
  1996     Theory.merge, etc.)
  1997 
  1998 The most basic sign_of operations (Theory.sign_of, Thm.sign_of_thm
  1999 etc.) as well as the sign field in Thm.rep_thm etc. have been retained
  2000 for convenience -- they merely return the theory.
  2001 
  2002 * Pure: type Type.tsig is superceded by theory in most interfaces.
  2003 
  2004 * Pure: the Isar proof context type is already defined early in Pure
  2005 as Context.proof (note that ProofContext.context and Proof.context are
  2006 aliases, where the latter is the preferred name).  This enables other
  2007 Isabelle components to refer to that type even before Isar is present.
  2008 
  2009 * Pure/sign/theory: discontinued named name spaces (i.e. classK,
  2010 typeK, constK, axiomK, oracleK), but provide explicit operations for
  2011 any of these kinds.  For example, Sign.intern typeK is now
  2012 Sign.intern_type, Theory.hide_space Sign.typeK is now
  2013 Theory.hide_types.  Also note that former
  2014 Theory.hide_classes/types/consts are now
  2015 Theory.hide_classes_i/types_i/consts_i, while the non '_i' versions
  2016 internalize their arguments!  INCOMPATIBILITY.
  2017 
  2018 * Pure: get_thm interface (of PureThy and ProofContext) expects
  2019 datatype thmref (with constructors Name and NameSelection) instead of
  2020 plain string -- INCOMPATIBILITY;
  2021 
  2022 * Pure: cases produced by proof methods specify options, where NONE
  2023 means to remove case bindings -- INCOMPATIBILITY in
  2024 (RAW_)METHOD_CASES.
  2025 
  2026 * Pure: the following operations retrieve axioms or theorems from a
  2027 theory node or theory hierarchy, respectively:
  2028 
  2029   Theory.axioms_of: theory -> (string * term) list
  2030   Theory.all_axioms_of: theory -> (string * term) list
  2031   PureThy.thms_of: theory -> (string * thm) list
  2032   PureThy.all_thms_of: theory -> (string * thm) list
  2033 
  2034 * Pure: print_tac now outputs the goal through the trace channel.
  2035 
  2036 * Isar toplevel: improved diagnostics, mostly for Poly/ML only.
  2037 Reference Toplevel.debug (default false) controls detailed printing
  2038 and tracing of low-level exceptions; Toplevel.profiling (default 0)
  2039 controls execution profiling -- set to 1 for time and 2 for space
  2040 (both increase the runtime).
  2041 
  2042 * Isar session: The initial use of ROOT.ML is now always timed,
  2043 i.e. the log will show the actual process times, in contrast to the
  2044 elapsed wall-clock time that the outer shell wrapper produces.
  2045 
  2046 * Simplifier: improved handling of bound variables (nameless
  2047 representation, avoid allocating new strings).  Simprocs that invoke
  2048 the Simplifier recursively should use Simplifier.inherit_bounds to
  2049 avoid local name clashes.  Failure to do so produces warnings
  2050 "Simplifier: renamed bound variable ..."; set Simplifier.debug_bounds
  2051 for further details.
  2052 
  2053 * ML functions legacy_bindings and use_legacy_bindings produce ML fact
  2054 bindings for all theorems stored within a given theory; this may help
  2055 in porting non-Isar theories to Isar ones, while keeping ML proof
  2056 scripts for the time being.
  2057 
  2058 * ML operator HTML.with_charset specifies the charset begin used for
  2059 generated HTML files.  For example:
  2060 
  2061   HTML.with_charset "utf-8" use_thy "Hebrew";
  2062   HTML.with_charset "utf-8" use_thy "Chinese";
  2063 
  2064 
  2065 *** System ***
  2066 
  2067 * Allow symlinks to all proper Isabelle executables (Isabelle,
  2068 isabelle, isatool etc.).
  2069 
  2070 * ISABELLE_DOC_FORMAT setting specifies preferred document format (for
  2071 isatool doc, isatool mkdir, display_drafts etc.).
  2072 
  2073 * isatool usedir: option -f allows specification of the ML file to be
  2074 used by Isabelle; default is ROOT.ML.
  2075 
  2076 * New isatool version outputs the version identifier of the Isabelle
  2077 distribution being used.
  2078 
  2079 * HOL: new isatool dimacs2hol converts files in DIMACS CNF format
  2080 (containing Boolean satisfiability problems) into Isabelle/HOL
  2081 theories.
  2082 
  2083 
  2084 
  2085 New in Isabelle2004 (April 2004)
  2086 --------------------------------
  2087 
  2088 *** General ***
  2089 
  2090 * Provers/order.ML:  new efficient reasoner for partial and linear orders.
  2091   Replaces linorder.ML.
  2092 
  2093 * Pure: Greek letters (except small lambda, \<lambda>), as well as Gothic
  2094   (\<aa>...\<zz>\<AA>...\<ZZ>), calligraphic (\<A>...\<Z>), and Euler
  2095   (\<a>...\<z>), are now considered normal letters, and can therefore
  2096   be used anywhere where an ASCII letter (a...zA...Z) has until
  2097   now. COMPATIBILITY: This obviously changes the parsing of some
  2098   terms, especially where a symbol has been used as a binder, say
  2099   '\<Pi>x. ...', which is now a type error since \<Pi>x will be parsed
  2100   as an identifier.  Fix it by inserting a space around former
  2101   symbols.  Call 'isatool fixgreek' to try to fix parsing errors in
  2102   existing theory and ML files.
  2103 
  2104 * Pure: Macintosh and Windows line-breaks are now allowed in theory files.
  2105 
  2106 * Pure: single letter sub/superscripts (\<^isub> and \<^isup>) are now
  2107   allowed in identifiers. Similar to Greek letters \<^isub> is now considered
  2108   a normal (but invisible) letter. For multiple letter subscripts repeat
  2109   \<^isub> like this: x\<^isub>1\<^isub>2.
  2110 
  2111 * Pure: There are now sub-/superscripts that can span more than one
  2112   character. Text between \<^bsub> and \<^esub> is set in subscript in
  2113   ProofGeneral and LaTeX, text between \<^bsup> and \<^esup> in
  2114   superscript. The new control characters are not identifier parts.
  2115 
  2116 * Pure: Control-symbols of the form \<^raw:...> will literally print the
  2117   content of "..." to the latex file instead of \isacntrl... . The "..."
  2118   may consist of any printable characters excluding the end bracket >.
  2119 
  2120 * Pure: Using new Isar command "finalconsts" (or the ML functions
  2121   Theory.add_finals or Theory.add_finals_i) it is now possible to
  2122   declare constants "final", which prevents their being given a definition
  2123   later.  It is useful for constants whose behaviour is fixed axiomatically
  2124   rather than definitionally, such as the meta-logic connectives.
  2125 
  2126 * Pure: 'instance' now handles general arities with general sorts
  2127   (i.e. intersections of classes),
  2128 
  2129 * Presentation: generated HTML now uses a CSS style sheet to make layout
  2130   (somewhat) independent of content. It is copied from lib/html/isabelle.css.
  2131   It can be changed to alter the colors/layout of generated pages.
  2132 
  2133 
  2134 *** Isar ***
  2135 
  2136 * Tactic emulation methods rule_tac, erule_tac, drule_tac, frule_tac,
  2137   cut_tac, subgoal_tac and thin_tac:
  2138   - Now understand static (Isar) contexts.  As a consequence, users of Isar
  2139     locales are no longer forced to write Isar proof scripts.
  2140     For details see Isar Reference Manual, paragraph 4.3.2: Further tactic
  2141     emulations.
  2142   - INCOMPATIBILITY: names of variables to be instantiated may no
  2143     longer be enclosed in quotes.  Instead, precede variable name with `?'.
  2144     This is consistent with the instantiation attribute "where".
  2145 
  2146 * Attributes "where" and "of":
  2147   - Now take type variables of instantiated theorem into account when reading
  2148     the instantiation string.  This fixes a bug that caused instantiated
  2149     theorems to have too special types in some circumstances.
  2150   - "where" permits explicit instantiations of type variables.
  2151 
  2152 * Calculation commands "moreover" and "also" no longer interfere with
  2153   current facts ("this"), admitting arbitrary combinations with "then"
  2154   and derived forms.
  2155 
  2156 * Locales:
  2157   - Goal statements involving the context element "includes" no longer
  2158     generate theorems with internal delta predicates (those ending on
  2159     "_axioms") in the premise.
  2160     Resolve particular premise with <locale>.intro to obtain old form.
  2161   - Fixed bug in type inference ("unify_frozen") that prevented mix of target
  2162     specification and "includes" elements in goal statement.
  2163   - Rule sets <locale>.intro and <locale>.axioms no longer declared as
  2164     [intro?] and [elim?] (respectively) by default.
  2165   - Experimental command for instantiation of locales in proof contexts:
  2166         instantiate <label>[<attrs>]: <loc>
  2167     Instantiates locale <loc> and adds all its theorems to the current context
  2168     taking into account their attributes.  Label and attrs are optional
  2169     modifiers, like in theorem declarations.  If present, names of
  2170     instantiated theorems are qualified with <label>, and the attributes
  2171     <attrs> are applied after any attributes these theorems might have already.
  2172       If the locale has assumptions, a chained fact of the form
  2173     "<loc> t1 ... tn" is expected from which instantiations of the parameters
  2174     are derived.  The command does not support old-style locales declared
  2175     with "locale (open)".
  2176       A few (very simple) examples can be found in FOL/ex/LocaleInst.thy.
  2177 
  2178 * HOL: Tactic emulation methods induct_tac and case_tac understand static
  2179   (Isar) contexts.
  2180 
  2181 
  2182 *** HOL ***
  2183 
  2184 * Proof import: new image HOL4 contains the imported library from
  2185   the HOL4 system with about 2500 theorems. It is imported by
  2186   replaying proof terms produced by HOL4 in Isabelle. The HOL4 image
  2187   can be used like any other Isabelle image.  See
  2188   HOL/Import/HOL/README for more information.
  2189 
  2190 * Simplifier:
  2191   - Much improved handling of linear and partial orders.
  2192     Reasoners for linear and partial orders are set up for type classes
  2193     "linorder" and "order" respectively, and are added to the default simpset
  2194     as solvers.  This means that the simplifier can build transitivity chains
  2195     to solve goals from the assumptions.
  2196   - INCOMPATIBILITY: old proofs break occasionally.  Typically, applications
  2197     of blast or auto after simplification become unnecessary because the goal
  2198     is solved by simplification already.
  2199 
  2200 * Numerics: new theory Ring_and_Field contains over 250 basic numerical laws,
  2201     all proved in axiomatic type classes for semirings, rings and fields.
  2202 
  2203 * Numerics:
  2204   - Numeric types (nat, int, and in HOL-Complex rat, real, complex, etc.) are
  2205     now formalized using the Ring_and_Field theory mentioned above.
  2206   - INCOMPATIBILITY: simplification and arithmetic behaves somewhat differently
  2207     than before, because now they are set up once in a generic manner.
  2208   - INCOMPATIBILITY: many type-specific arithmetic laws have gone.
  2209     Look for the general versions in Ring_and_Field (and Power if they concern
  2210     exponentiation).
  2211 
  2212 * Type "rat" of the rational numbers is now available in HOL-Complex.
  2213 
  2214 * Records:
  2215   - Record types are now by default printed with their type abbreviation
  2216     instead of the list of all field types. This can be configured via
  2217     the reference "print_record_type_abbr".
  2218   - Simproc "record_upd_simproc" for simplification of multiple updates added
  2219     (not enabled by default).
  2220   - Simproc "record_ex_sel_eq_simproc" to simplify EX x. sel r = x resp.
  2221     EX x. x = sel r to True (not enabled by default).
  2222   - Tactic "record_split_simp_tac" to split and simplify records added.
  2223 
  2224 * 'specification' command added, allowing for definition by
  2225   specification.  There is also an 'ax_specification' command that
  2226   introduces the new constants axiomatically.
  2227 
  2228 * arith(_tac) is now able to generate counterexamples for reals as well.
  2229 
  2230 * HOL-Algebra: new locale "ring" for non-commutative rings.
  2231 
  2232 * HOL-ex: InductiveInvariant_examples illustrates advanced recursive function
  2233   definitions, thanks to Sava Krsti\'{c} and John Matthews.
  2234 
  2235 * HOL-Matrix: a first theory for matrices in HOL with an application of
  2236   matrix theory to linear programming.
  2237 
  2238 * Unions and Intersections:
  2239   The latex output syntax of UN and INT has been changed
  2240   from "\Union x \in A. B" to "\Union_{x \in A} B"
  2241   i.e. the index formulae has become a subscript.
  2242   Similarly for "\Union x. B", and for \Inter instead of \Union.
  2243 
  2244 * Unions and Intersections over Intervals:
  2245   There is new short syntax "UN i<=n. A" for "UN i:{0..n}. A". There is
  2246   also an x-symbol version with subscripts "\<Union>\<^bsub>i <= n\<^esub>. A"
  2247   like in normal math, and corresponding versions for < and for intersection.
  2248 
  2249 * HOL/List: Ordering "lexico" is renamed "lenlex" and the standard
  2250   lexicographic dictonary ordering has been added as "lexord".
  2251 
  2252 * ML: the legacy theory structures Int and List have been removed. They had
  2253   conflicted with ML Basis Library structures having the same names.
  2254 
  2255 * 'refute' command added to search for (finite) countermodels.  Only works
  2256   for a fragment of HOL.  The installation of an external SAT solver is
  2257   highly recommended.  See "HOL/Refute.thy" for details.
  2258 
  2259 * 'quickcheck' command: Allows to find counterexamples by evaluating
  2260   formulae under an assignment of free variables to random values.
  2261   In contrast to 'refute', it can deal with inductive datatypes,
  2262   but cannot handle quantifiers. See "HOL/ex/Quickcheck_Examples.thy"
  2263   for examples.
  2264 
  2265 
  2266 *** HOLCF ***
  2267 
  2268 * Streams now come with concatenation and are part of the HOLCF image
  2269 
  2270 
  2271 
  2272 New in Isabelle2003 (May 2003)
  2273 ------------------------------
  2274 
  2275 *** General ***
  2276 
  2277 * Provers/simplifier:
  2278 
  2279   - Completely reimplemented method simp (ML: Asm_full_simp_tac):
  2280     Assumptions are now subject to complete mutual simplification,
  2281     not just from left to right. The simplifier now preserves
  2282     the order of assumptions.
  2283 
  2284     Potential INCOMPATIBILITY:
  2285 
  2286     -- simp sometimes diverges where the old version did
  2287        not, e.g. invoking simp on the goal
  2288 
  2289         [| P (f x); y = x; f x = f y |] ==> Q
  2290 
  2291        now gives rise to the infinite reduction sequence
  2292 
  2293         P(f x) --(f x = f y)--> P(f y) --(y = x)--> P(f x) --(f x = f y)--> ...
  2294 
  2295        Using "simp (asm_lr)" (ML: Asm_lr_simp_tac) instead often solves this
  2296        kind of problem.
  2297 
  2298     -- Tactics combining classical reasoner and simplification (such as auto)
  2299        are also affected by this change, because many of them rely on
  2300        simp. They may sometimes diverge as well or yield a different numbers
  2301        of subgoals. Try to use e.g. force, fastsimp, or safe instead of auto
  2302        in case of problems. Sometimes subsequent calls to the classical
  2303        reasoner will fail because a preceeding call to the simplifier too
  2304        eagerly simplified the goal, e.g. deleted redundant premises.
  2305 
  2306   - The simplifier trace now shows the names of the applied rewrite rules
  2307 
  2308   - You can limit the number of recursive invocations of the simplifier
  2309     during conditional rewriting (where the simplifie tries to solve the
  2310     conditions before applying the rewrite rule):
  2311     ML "simp_depth_limit := n"
  2312     where n is an integer. Thus you can force termination where previously
  2313     the simplifier would diverge.
  2314 
  2315   - Accepts free variables as head terms in congruence rules.  Useful in Isar.
  2316 
  2317   - No longer aborts on failed congruence proof.  Instead, the
  2318     congruence is ignored.
  2319 
  2320 * Pure: New generic framework for extracting programs from constructive
  2321   proofs. See HOL/Extraction.thy for an example instantiation, as well
  2322   as HOL/Extraction for some case studies.
  2323 
  2324 * Pure: The main goal of the proof state is no longer shown by default, only
  2325 the subgoals. This behaviour is controlled by a new flag.
  2326    PG menu: Isabelle/Isar -> Settings -> Show Main Goal
  2327 (ML: Proof.show_main_goal).
  2328 
  2329 * Pure: You can find all matching introduction rules for subgoal 1, i.e. all
  2330 rules whose conclusion matches subgoal 1:
  2331       PG menu: Isabelle/Isar -> Show me -> matching rules
  2332 The rules are ordered by how closely they match the subgoal.
  2333 In particular, rules that solve a subgoal outright are displayed first
  2334 (or rather last, the way they are printed).
  2335 (ML: ProofGeneral.print_intros())
  2336 
  2337 * Pure: New flag trace_unify_fail causes unification to print
  2338 diagnostic information (PG: in trace buffer) when it fails. This is
  2339 useful for figuring out why single step proofs like rule, erule or
  2340 assumption failed.
  2341 
  2342 * Pure: Locale specifications now produce predicate definitions
  2343 according to the body of text (covering assumptions modulo local
  2344 definitions); predicate "loc_axioms" covers newly introduced text,
  2345 while "loc" is cumulative wrt. all included locale expressions; the
  2346 latter view is presented only on export into the global theory
  2347 context; potential INCOMPATIBILITY, use "(open)" option to fall back
  2348 on the old view without predicates;
  2349 
  2350 * Pure: predefined locales "var" and "struct" are useful for sharing
  2351 parameters (as in CASL, for example); just specify something like
  2352 ``var x + var y + struct M'' as import;
  2353 
  2354 * Pure: improved thms_containing: proper indexing of facts instead of
  2355 raw theorems; check validity of results wrt. current name space;
  2356 include local facts of proof configuration (also covers active
  2357 locales), cover fixed variables in index; may use "_" in term
  2358 specification; an optional limit for the number of printed facts may
  2359 be given (the default is 40);
  2360 
  2361 * Pure: disallow duplicate fact bindings within new-style theory files
  2362 (batch-mode only);
  2363 
  2364 * Provers: improved induct method: assumptions introduced by case
  2365 "foo" are split into "foo.hyps" (from the rule) and "foo.prems" (from
  2366 the goal statement); "foo" still refers to all facts collectively;
  2367 
  2368 * Provers: the function blast.overloaded has been removed: all constants
  2369 are regarded as potentially overloaded, which improves robustness in exchange
  2370 for slight decrease in efficiency;
  2371 
  2372 * Provers/linorder: New generic prover for transitivity reasoning over
  2373 linear orders.  Note: this prover is not efficient!
  2374 
  2375 * Isar: preview of problems to finish 'show' now produce an error
  2376 rather than just a warning (in interactive mode);
  2377 
  2378 
  2379 *** HOL ***
  2380 
  2381 * arith(_tac)
  2382 
  2383  - Produces a counter example if it cannot prove a goal.
  2384    Note that the counter example may be spurious if the goal is not a formula
  2385    of quantifier-free linear arithmetic.
  2386    In ProofGeneral the counter example appears in the trace buffer.
  2387 
  2388  - Knows about div k and mod k where k is a numeral of type nat or int.
  2389 
  2390  - Calls full Presburger arithmetic (by Amine Chaieb) if quantifier-free
  2391    linear arithmetic fails. This takes account of quantifiers and divisibility.
  2392    Presburger arithmetic can also be called explicitly via presburger(_tac).
  2393 
  2394 * simp's arithmetic capabilities have been enhanced a bit: it now
  2395 takes ~= in premises into account (by performing a case split);
  2396 
  2397 * simp reduces "m*(n div m) + n mod m" to n, even if the two summands
  2398 are distributed over a sum of terms;
  2399 
  2400 * New tactic "trans_tac" and method "trans" instantiate
  2401 Provers/linorder.ML for axclasses "order" and "linorder" (predicates
  2402 "<=", "<" and "=").
  2403 
  2404 * function INCOMPATIBILITIES: Pi-sets have been redefined and moved from main
  2405 HOL to Library/FuncSet; constant "Fun.op o" is now called "Fun.comp";
  2406 
  2407 * 'typedef' command has new option "open" to suppress the set
  2408 definition;
  2409 
  2410 * functions Min and Max on finite sets have been introduced (theory
  2411 Finite_Set);
  2412 
  2413 * attribute [symmetric] now works for relations as well; it turns
  2414 (x,y) : R^-1 into (y,x) : R, and vice versa;
  2415 
  2416 * induct over a !!-quantified statement (say !!x1..xn):
  2417   each "case" automatically performs "fix x1 .. xn" with exactly those names.
  2418 
  2419 * Map: `empty' is no longer a constant but a syntactic abbreviation for
  2420 %x. None. Warning: empty_def now refers to the previously hidden definition
  2421 of the empty set.
  2422 
  2423 * Algebra: formalization of classical algebra.  Intended as base for
  2424 any algebraic development in Isabelle.  Currently covers group theory
  2425 (up to Sylow's theorem) and ring theory (Universal Property of
  2426 Univariate Polynomials).  Contributions welcome;
  2427 
  2428 * GroupTheory: deleted, since its material has been moved to Algebra;
  2429 
  2430 * Complex: new directory of the complex numbers with numeric constants,
  2431 nonstandard complex numbers, and some complex analysis, standard and
  2432 nonstandard (Jacques Fleuriot);
  2433 
  2434 * HOL-Complex: new image for analysis, replacing HOL-Real and HOL-Hyperreal;
  2435 
  2436 * Hyperreal: introduced Gauge integration and hyperreal logarithms (Jacques
  2437 Fleuriot);
  2438 
  2439 * Real/HahnBanach: updated and adapted to locales;
  2440 
  2441 * NumberTheory: added Gauss's law of quadratic reciprocity (by Avigad,
  2442 Gray and Kramer);
  2443 
  2444 * UNITY: added the Meier-Sanders theory of progress sets;
  2445 
  2446 * MicroJava: bytecode verifier and lightweight bytecode verifier
  2447 as abstract algorithms, instantiated to the JVM;
  2448 
  2449 * Bali: Java source language formalization. Type system, operational
  2450 semantics, axiomatic semantics. Supported language features:
  2451 classes, interfaces, objects,virtual methods, static methods,
  2452 static/instance fields, arrays, access modifiers, definite
  2453 assignment, exceptions.
  2454 
  2455 
  2456 *** ZF ***
  2457 
  2458 * ZF/Constructible: consistency proof for AC (Gdel's constructible
  2459 universe, etc.);
  2460 
  2461 * Main ZF: virtually all theories converted to new-style format;
  2462 
  2463 
  2464 *** ML ***
  2465 
  2466 * Pure: Tactic.prove provides sane interface for internal proofs;
  2467 omits the infamous "standard" operation, so this is more appropriate
  2468 than prove_goalw_cterm in many situations (e.g. in simprocs);
  2469 
  2470 * Pure: improved error reporting of simprocs;
  2471 
  2472 * Provers: Simplifier.simproc(_i) provides sane interface for setting
  2473 up simprocs;
  2474 
  2475 
  2476 *** Document preparation ***
  2477 
  2478 * uses \par instead of \\ for line breaks in theory text. This may
  2479 shift some page breaks in large documents. To get the old behaviour
  2480 use \renewcommand{\isanewline}{\mbox{}\\\mbox{}} in root.tex.
  2481 
  2482 * minimized dependencies of isabelle.sty and isabellesym.sty on
  2483 other packages
  2484 
  2485 * \<euro> now needs package babel/greek instead of marvosym (which
  2486 broke \Rightarrow)
  2487 
  2488 * normal size for \<zero>...\<nine> (uses \mathbf instead of
  2489 textcomp package)
  2490 
  2491 
  2492 
  2493 New in Isabelle2002 (March 2002)
  2494 --------------------------------
  2495 
  2496 *** Document preparation ***
  2497 
  2498 * greatly simplified document preparation setup, including more
  2499 graceful interpretation of isatool usedir -i/-d/-D options, and more
  2500 instructive isatool mkdir; users should basically be able to get
  2501 started with "isatool mkdir HOL Test && isatool make"; alternatively,
  2502 users may run a separate document processing stage manually like this:
  2503 "isatool usedir -D output HOL Test && isatool document Test/output";
  2504 
  2505 * theory dependency graph may now be incorporated into documents;
  2506 isatool usedir -g true will produce session_graph.eps/.pdf for use
  2507 with \includegraphics of LaTeX;
  2508 
  2509 * proper spacing of consecutive markup elements, especially text
  2510 blocks after section headings;
  2511 
  2512 * support bold style (for single symbols only), input syntax is like
  2513 this: "\<^bold>\<alpha>" or "\<^bold>A";
  2514 
  2515 * \<bullet> is now output as bold \cdot by default, which looks much
  2516 better in printed text;
  2517 
  2518 * added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
  2519 note that these symbols are currently unavailable in Proof General /
  2520 X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
  2521 
  2522 * isatool latex no longer depends on changed TEXINPUTS, instead
  2523 isatool document copies the Isabelle style files to the target
  2524 location;
  2525 
  2526 
  2527 *** Isar ***
  2528 
  2529 * Pure/Provers: improved proof by cases and induction;
  2530   - 'case' command admits impromptu naming of parameters (such as
  2531     "case (Suc n)");
  2532   - 'induct' method divinates rule instantiation from the inductive
  2533     claim; no longer requires excessive ?P bindings for proper
  2534     instantiation of cases;
  2535   - 'induct' method properly enumerates all possibilities of set/type
  2536     rules; as a consequence facts may be also passed through *type*
  2537     rules without further ado;
  2538   - 'induct' method now derives symbolic cases from the *rulified*
  2539     rule (before it used to rulify cases stemming from the internal
  2540     atomized version); this means that the context of a non-atomic
  2541     statement becomes is included in the hypothesis, avoiding the
  2542     slightly cumbersome show "PROP ?case" form;
  2543   - 'induct' may now use elim-style induction rules without chaining
  2544     facts, using ``missing'' premises from the goal state; this allows
  2545     rules stemming from inductive sets to be applied in unstructured
  2546     scripts, while still benefitting from proper handling of non-atomic
  2547     statements; NB: major inductive premises need to be put first, all
  2548     the rest of the goal is passed through the induction;
  2549   - 'induct' proper support for mutual induction involving non-atomic
  2550     rule statements (uses the new concept of simultaneous goals, see
  2551     below);
  2552   - append all possible rule selections, but only use the first
  2553     success (no backtracking);
  2554   - removed obsolete "(simplified)" and "(stripped)" options of methods;
  2555   - undeclared rule case names default to numbers 1, 2, 3, ...;
  2556   - added 'print_induct_rules' (covered by help item in recent Proof
  2557     General versions);
  2558   - moved induct/cases attributes to Pure, methods to Provers;
  2559   - generic method setup instantiated for FOL and HOL;
  2560 
  2561 * Pure: support multiple simultaneous goal statements, for example
  2562 "have a: A and b: B" (same for 'theorem' etc.); being a pure
  2563 meta-level mechanism, this acts as if several individual goals had
  2564 been stated separately; in particular common proof methods need to be
  2565 repeated in order to cover all claims; note that a single elimination
  2566 step is *not* sufficient to establish the two conjunctions, so this
  2567 fails:
  2568 
  2569   assume "A & B" then have A and B ..   (*".." fails*)
  2570 
  2571 better use "obtain" in situations as above; alternative refer to
  2572 multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
  2573 
  2574 * Pure: proper integration with ``locales''; unlike the original
  2575 version by Florian Kammller, Isar locales package high-level proof
  2576 contexts rather than raw logical ones (e.g. we admit to include
  2577 attributes everywhere); operations on locales include merge and
  2578 rename; support for implicit arguments (``structures''); simultaneous
  2579 type-inference over imports and text; see also HOL/ex/Locales.thy for
  2580 some examples;
  2581 
  2582 * Pure: the following commands have been ``localized'', supporting a
  2583 target locale specification "(in name)": 'lemma', 'theorem',
  2584 'corollary', 'lemmas', 'theorems', 'declare'; the results will be
  2585 stored both within the locale and at the theory level (exported and
  2586 qualified by the locale name);
  2587 
  2588 * Pure: theory goals may now be specified in ``long'' form, with
  2589 ad-hoc contexts consisting of arbitrary locale elements. for example
  2590 ``lemma foo: fixes x assumes "A x" shows "B x"'' (local syntax and
  2591 definitions may be given, too); the result is a meta-level rule with
  2592 the context elements being discharged in the obvious way;
  2593 
  2594 * Pure: new proof command 'using' allows to augment currently used
  2595 facts after a goal statement ('using' is syntactically analogous to
  2596 'apply', but acts on the goal's facts only); this allows chained facts
  2597 to be separated into parts given before and after a claim, as in
  2598 ``from a and b have C using d and e <proof>'';
  2599 
  2600 * Pure: renamed "antecedent" case to "rule_context";
  2601 
  2602 * Pure: new 'judgment' command records explicit information about the
  2603 object-logic embedding (used by several tools internally); no longer
  2604 use hard-wired "Trueprop";
  2605 
  2606 * Pure: added 'corollary' command;
  2607 
  2608 * Pure: fixed 'token_translation' command;
  2609 
  2610 * Pure: removed obsolete 'exported' attribute;
  2611 
  2612 * Pure: dummy pattern "_" in is/let is now automatically lifted over
  2613 bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
  2614 supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
  2615 
  2616 * Pure: method 'atomize' presents local goal premises as object-level
  2617 statements (atomic meta-level propositions); setup controlled via
  2618 rewrite rules declarations of 'atomize' attribute; example
  2619 application: 'induct' method with proper rule statements in improper
  2620 proof *scripts*;
  2621 
  2622 * Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
  2623 now consider the syntactic context of assumptions, giving a better
  2624 chance to get type-inference of the arguments right (this is
  2625 especially important for locales);
  2626 
  2627 * Pure: "sorry" no longer requires quick_and_dirty in interactive
  2628 mode;
  2629 
  2630 * Pure/obtain: the formal conclusion "thesis", being marked as
  2631 ``internal'', may no longer be reference directly in the text;
  2632 potential INCOMPATIBILITY, may need to use "?thesis" in rare
  2633 situations;
  2634 
  2635 * Pure: generic 'sym' attribute which declares a rule both as pure
  2636 'elim?' and for the 'symmetric' operation;
  2637 
  2638 * Pure: marginal comments ``--'' may now occur just anywhere in the
  2639 text; the fixed correlation with particular command syntax has been
  2640 discontinued;
  2641 
  2642 * Pure: new method 'rules' is particularly well-suited for proof
  2643 search in intuitionistic logic; a bit slower than 'blast' or 'fast',
  2644 but often produces more compact proof terms with less detours;
  2645 
  2646 * Pure/Provers/classical: simplified integration with pure rule
  2647 attributes and methods; the classical "intro?/elim?/dest?"
  2648 declarations coincide with the pure ones; the "rule" method no longer
  2649 includes classically swapped intros; "intro" and "elim" methods no
  2650 longer pick rules from the context; also got rid of ML declarations
  2651 AddXIs/AddXEs/AddXDs; all of this has some potential for
  2652 INCOMPATIBILITY;
  2653 
  2654 * Provers/classical: attribute 'swapped' produces classical inversions
  2655 of introduction rules;
  2656 
  2657 * Provers/simplifier: 'simplified' attribute may refer to explicit
  2658 rules instead of full simplifier context; 'iff' attribute handles
  2659 conditional rules;
  2660 
  2661 * HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
  2662 
  2663 * HOL: 'recdef' now fails on unfinished automated proofs, use
  2664 "(permissive)" option to recover old behavior;
  2665 
  2666 * HOL: 'inductive' no longer features separate (collective) attributes
  2667 for 'intros' (was found too confusing);
  2668 
  2669 * HOL: properly declared induction rules less_induct and
  2670 wf_induct_rule;
  2671 
  2672 
  2673 *** HOL ***
  2674 
  2675 * HOL: moved over to sane numeral syntax; the new policy is as
  2676 follows:
  2677 
  2678   - 0 and 1 are polymorphic constants, which are defined on any
  2679   numeric type (nat, int, real etc.);
  2680 
  2681   - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
  2682   binary representation internally;
  2683 
  2684   - type nat has special constructor Suc, and generally prefers Suc 0
  2685   over 1::nat and Suc (Suc 0) over 2::nat;
  2686 
  2687 This change may cause significant problems of INCOMPATIBILITY; here
  2688 are some hints on converting existing sources:
  2689 
  2690   - due to the new "num" token, "-0" and "-1" etc. are now atomic
  2691   entities, so expressions involving "-" (unary or binary minus) need
  2692   to be spaced properly;
  2693 
  2694   - existing occurrences of "1" may need to be constraint "1::nat" or
  2695   even replaced by Suc 0; similar for old "2";
  2696 
  2697   - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
  2698 
  2699   - remove all special provisions on numerals in proofs;
  2700 
  2701 * HOL: simp rules nat_number expand numerals on nat to Suc/0
  2702 representation (depends on bin_arith_simps in the default context);
  2703 
  2704 * HOL: symbolic syntax for x^2 (numeral 2);
  2705 
  2706 * HOL: the class of all HOL types is now called "type" rather than
  2707 "term"; INCOMPATIBILITY, need to adapt references to this type class
  2708 in axclass/classes, instance/arities, and (usually rare) occurrences
  2709 in typings (of consts etc.); internally the class is called
  2710 "HOL.type", ML programs should refer to HOLogic.typeS;
  2711 
  2712 * HOL/record package improvements:
  2713   - new derived operations "fields" to build a partial record section,
  2714     "extend" to promote a fixed record to a record scheme, and
  2715     "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
  2716     declared as simp by default;
  2717   - shared operations ("more", "fields", etc.) now need to be always
  2718     qualified) --- potential INCOMPATIBILITY;
  2719   - removed "make_scheme" operations (use "make" with "extend") --
  2720     INCOMPATIBILITY;
  2721   - removed "more" class (simply use "term") -- INCOMPATIBILITY;
  2722   - provides cases/induct rules for use with corresponding Isar
  2723     methods (for concrete records, record schemes, concrete more
  2724     parts, and schematic more parts -- in that order);
  2725   - internal definitions directly based on a light-weight abstract
  2726     theory of product types over typedef rather than datatype;
  2727 
  2728 * HOL: generic code generator for generating executable ML code from
  2729 specifications; specific support for HOL constructs such as inductive
  2730 datatypes and sets, as well as recursive functions; can be invoked
  2731 via 'generate_code' theory section;
  2732 
  2733 * HOL: canonical cases/induct rules for n-tuples (n = 3..7);
  2734 
  2735 * HOL: consolidated and renamed several theories.  In particular:
  2736         Ord.thy has been absorbed into HOL.thy
  2737         String.thy has been absorbed into List.thy
  2738 
  2739 * HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
  2740 (beware of argument permutation!);
  2741 
  2742 * HOL: linorder_less_split superseded by linorder_cases;
  2743 
  2744 * HOL/List: "nodups" renamed to "distinct";
  2745 
  2746 * HOL: added "The" definite description operator; move Hilbert's "Eps"
  2747 to peripheral theory "Hilbert_Choice"; some INCOMPATIBILITIES:
  2748   - Ex_def has changed, now need to use some_eq_ex
  2749 
  2750 * HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
  2751 in this (rare) case use:
  2752 
  2753   delSWrapper "split_all_tac"
  2754   addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
  2755 
  2756 * HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
  2757 MAY FAIL;
  2758 
  2759 * HOL: introduced f^n = f o ... o f; warning: due to the limits of
  2760 Isabelle's type classes, ^ on functions and relations has too general
  2761 a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
  2762 necessary to attach explicit type constraints;
  2763 
  2764 * HOL/Relation: the prefix name of the infix "O" has been changed from
  2765 "comp" to "rel_comp"; INCOMPATIBILITY: a few theorems have been
  2766 renamed accordingly (eg "compI" -> "rel_compI").
  2767 
  2768 * HOL: syntax translations now work properly with numerals and records
  2769 expressions;
  2770 
  2771 * HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
  2772 of "lam" -- INCOMPATIBILITY;
  2773 
  2774 * HOL: got rid of some global declarations (potential INCOMPATIBILITY
  2775 for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
  2776 renamed "Product_Type.unit";
  2777 
  2778 * HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
  2779 
  2780 * HOL: removed obsolete theorem "optionE" (use "option.exhaust", or
  2781 the "cases" method);
  2782 
  2783 * HOL/GroupTheory: group theory examples including Sylow's theorem (by
  2784 Florian Kammller);
  2785 
  2786 * HOL/IMP: updated and converted to new-style theory format; several
  2787 parts turned into readable document, with proper Isar proof texts and
  2788 some explanations (by Gerwin Klein);
  2789 
  2790 * HOL-Real: added Complex_Numbers (by Gertrud Bauer);
  2791 
  2792 * HOL-Hyperreal is now a logic image;
  2793 
  2794 
  2795 *** HOLCF ***
  2796 
  2797 * Isar: consts/constdefs supports mixfix syntax for continuous
  2798 operations;
  2799 
  2800 * Isar: domain package adapted to new-style theory format, e.g. see
  2801 HOLCF/ex/Dnat.thy;
  2802 
  2803 * theory Lift: proper use of rep_datatype lift instead of ML hacks --
  2804 potential INCOMPATIBILITY; now use plain induct_tac instead of former
  2805 lift.induct_tac, always use UU instead of Undef;
  2806 
  2807 * HOLCF/IMP: updated and converted to new-style theory;
  2808 
  2809 
  2810 *** ZF ***
  2811 
  2812 * Isar: proper integration of logic-specific tools and packages,
  2813 including theory commands '(co)inductive', '(co)datatype',
  2814 'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
  2815 'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
  2816 
  2817 * theory Main no longer includes AC; for the Axiom of Choice, base
  2818 your theory on Main_ZFC;
  2819 
  2820 * the integer library now covers quotients and remainders, with many
  2821 laws relating division to addition, multiplication, etc.;
  2822 
  2823 * ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
  2824 typeless version of the formalism;
  2825 
  2826 * ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory
  2827 format;
  2828 
  2829 * ZF/Induct: new directory for examples of inductive definitions,
  2830 including theory Multiset for multiset orderings; converted to
  2831 new-style theory format;
  2832 
  2833 * ZF: many new theorems about lists, ordinals, etc.;
  2834 
  2835 
  2836 *** General ***
  2837 
  2838 * Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
  2839 variable proof controls level of detail: 0 = no proofs (only oracle
  2840 dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
  2841 also ref manual for further ML interfaces;
  2842 
  2843 * Pure/axclass: removed obsolete ML interface
  2844 goal_subclass/goal_arity;
  2845 
  2846 * Pure/syntax: new token syntax "num" for plain numerals (without "#"
  2847 of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
  2848 separate tokens, so expressions involving minus need to be spaced
  2849 properly;
  2850 
  2851 * Pure/syntax: support non-oriented infixes, using keyword "infix"
  2852 rather than "infixl" or "infixr";
  2853 
  2854 * Pure/syntax: concrete syntax for dummy type variables admits genuine
  2855 sort constraint specifications in type inference; e.g. "x::_::foo"
  2856 ensures that the type of "x" is of sort "foo" (but not necessarily a
  2857 type variable);
  2858 
  2859 * Pure/syntax: print modes "type_brackets" and "no_type_brackets"
  2860 control output of nested => (types); the default behavior is
  2861 "type_brackets";
  2862 
  2863 * Pure/syntax: builtin parse translation for "_constify" turns valued
  2864 tokens into AST constants;
  2865 
  2866 * Pure/syntax: prefer later declarations of translations and print
  2867 translation functions; potential INCOMPATIBILITY: need to reverse
  2868 multiple declarations for same syntax element constant;
  2869 
  2870 * Pure/show_hyps reset by default (in accordance to existing Isar
  2871 practice);
  2872 
  2873 * Provers/classical: renamed addaltern to addafter, addSaltern to
  2874 addSafter;
  2875 
  2876 * Provers/clasimp: ``iff'' declarations now handle conditional rules
  2877 as well;
  2878 
  2879 * system: tested support for MacOS X; should be able to get Isabelle +
  2880 Proof General to work in a plain Terminal after installing Poly/ML
  2881 (e.g. from the Isabelle distribution area) and GNU bash alone
  2882 (e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
  2883 support requires further installations, e.g. from
  2884 http://fink.sourceforge.net/);
  2885 
  2886 * system: support Poly/ML 4.1.1 (able to manage larger heaps);
  2887 
  2888 * system: reduced base memory usage by Poly/ML (approx. 20 MB instead
  2889 of 40 MB), cf. ML_OPTIONS;
  2890 
  2891 * system: Proof General keywords specification is now part of the
  2892 Isabelle distribution (see etc/isar-keywords.el);
  2893 
  2894 * system: support for persistent Proof General sessions (refrain from
  2895 outdating all loaded theories on startup); user may create writable
  2896 logic images like this: ``isabelle -q HOL Test'';
  2897 
  2898 * system: smart selection of Isabelle process versus Isabelle
  2899 interface, accommodates case-insensitive file systems (e.g. HFS+); may
  2900 run both "isabelle" and "Isabelle" even if file names are badly
  2901 damaged (executable inspects the case of the first letter of its own
  2902 name); added separate "isabelle-process" and "isabelle-interface";
  2903 
  2904 * system: refrain from any attempt at filtering input streams; no
  2905 longer support ``8bit'' encoding of old isabelle font, instead proper
  2906 iso-latin characters may now be used; the related isatools
  2907 "symbolinput" and "nonascii" have disappeared as well;
  2908 
  2909 * system: removed old "xterm" interface (the print modes "xterm" and
  2910 "xterm_color" are still available for direct use in a suitable
  2911 terminal);
  2912 
  2913 
  2914 
  2915 New in Isabelle99-2 (February 2001)
  2916 -----------------------------------
  2917 
  2918 *** Overview of INCOMPATIBILITIES ***
  2919 
  2920 * HOL: please note that theories in the Library and elsewhere often use the
  2921 new-style (Isar) format; to refer to their theorems in an ML script you must
  2922 bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
  2923 
  2924 * HOL: inductive package no longer splits induction rule aggressively,
  2925 but only as far as specified by the introductions given; the old
  2926 format may be recovered via ML function complete_split_rule or attribute
  2927 'split_rule (complete)';
  2928 
  2929 * HOL: induct renamed to lfp_induct, lfp_Tarski to lfp_unfold,
  2930 gfp_Tarski to gfp_unfold;
  2931 
  2932 * HOL: contrapos, contrapos2 renamed to contrapos_nn, contrapos_pp;
  2933 
  2934 * HOL: infix "dvd" now has priority 50 rather than 70 (because it is a
  2935 relation); infix "^^" has been renamed "``"; infix "``" has been
  2936 renamed "`"; "univalent" has been renamed "single_valued";
  2937 
  2938 * HOL/Real: "rinv" and "hrinv" replaced by overloaded "inverse"
  2939 operation;
  2940 
  2941 * HOLCF: infix "`" has been renamed "$"; the symbol syntax is \<cdot>;
  2942 
  2943 * Isar: 'obtain' no longer declares "that" fact as simp/intro;
  2944 
  2945 * Isar/HOL: method 'induct' now handles non-atomic goals; as a
  2946 consequence, it is no longer monotonic wrt. the local goal context
  2947 (which is now passed through the inductive cases);
  2948 
  2949 * Document preparation: renamed standard symbols \<ll> to \<lless> and
  2950 \<gg> to \<ggreater>;
  2951 
  2952 
  2953 *** Document preparation ***
  2954 
  2955 * \isabellestyle{NAME} selects version of Isabelle output (currently
  2956 available: are "it" for near math-mode best-style output, "sl" for
  2957 slanted text style, and "tt" for plain type-writer; if no
  2958 \isabellestyle command is given, output is according to slanted
  2959 type-writer);
  2960 
  2961 * support sub/super scripts (for single symbols only), input syntax is
  2962 like this: "A\<^sup>*" or "A\<^sup>\<star>";
  2963 
  2964 * some more standard symbols; see Appendix A of the system manual for
  2965 the complete list of symbols defined in isabellesym.sty;
  2966 
  2967 * improved isabelle style files; more abstract symbol implementation
  2968 (should now use \isamath{...} and \isatext{...} in custom symbol
  2969 definitions);
  2970 
  2971 * antiquotation @{goals} and @{subgoals} for output of *dynamic* goals
  2972 state; Note that presentation of goal states does not conform to
  2973 actual human-readable proof documents.  Please do not include goal
  2974 states into document output unless you really know what you are doing!
  2975 
  2976 * proper indentation of antiquoted output with proportional LaTeX
  2977 fonts;
  2978 
  2979 * no_document ML operator temporarily disables LaTeX document
  2980 generation;
  2981 
  2982 * isatool unsymbolize tunes sources for plain ASCII communication;
  2983 
  2984 
  2985 *** Isar ***
  2986 
  2987 * Pure: Isar now suffers initial goal statements to contain unbound
  2988 schematic variables (this does not conform to actual readable proof
  2989 documents, due to unpredictable outcome and non-compositional proof
  2990 checking); users who know what they are doing may use schematic goals
  2991 for Prolog-style synthesis of proven results;
  2992 
  2993 * Pure: assumption method (an implicit finishing) now handles actual
  2994 rules as well;
  2995 
  2996 * Pure: improved 'obtain' --- moved to Pure, insert "that" into
  2997 initial goal, declare "that" only as Pure intro (only for single
  2998 steps); the "that" rule assumption may now be involved in implicit
  2999 finishing, thus ".." becomes a feasible for trivial obtains;
  3000 
  3001 * Pure: default proof step now includes 'intro_classes'; thus trivial
  3002 instance proofs may be performed by "..";
  3003 
  3004 * Pure: ?thesis / ?this / "..." now work for pure meta-level
  3005 statements as well;
  3006 
  3007 * Pure: more robust selection of calculational rules;
  3008 
  3009 * Pure: the builtin notion of 'finished' goal now includes the ==-refl
  3010 rule (as well as the assumption rule);
  3011 
  3012 * Pure: 'thm_deps' command visualizes dependencies of theorems and
  3013 lemmas, using the graph browser tool;
  3014 
  3015 * Pure: predict failure of "show" in interactive mode;
  3016 
  3017 * Pure: 'thms_containing' now takes actual terms as arguments;
  3018 
  3019 * HOL: improved method 'induct' --- now handles non-atomic goals
  3020 (potential INCOMPATIBILITY); tuned error handling;
  3021 
  3022 * HOL: cases and induct rules now provide explicit hints about the
  3023 number of facts to be consumed (0 for "type" and 1 for "set" rules);
  3024 any remaining facts are inserted into the goal verbatim;
  3025 
  3026 * HOL: local contexts (aka cases) may now contain term bindings as
  3027 well; the 'cases' and 'induct' methods new provide a ?case binding for
  3028 the result to be shown in each case;
  3029 
  3030 * HOL: added 'recdef_tc' command;
  3031 
  3032 * isatool convert assists in eliminating legacy ML scripts;
  3033 
  3034 
  3035 *** HOL ***
  3036 
  3037 * HOL/Library: a collection of generic theories to be used together
  3038 with main HOL; the theory loader path already includes this directory
  3039 by default; the following existing theories have been moved here:
  3040 HOL/Induct/Multiset, HOL/Induct/Acc (as Accessible_Part), HOL/While
  3041 (as While_Combinator), HOL/Lex/Prefix (as List_Prefix);
  3042 
  3043 * HOL/Unix: "Some aspects of Unix file-system security", a typical
  3044 modelling and verification task performed in Isabelle/HOL +
  3045 Isabelle/Isar + Isabelle document preparation (by Markus Wenzel).
  3046 
  3047 * HOL/Algebra: special summation operator SUM no longer exists, it has
  3048 been replaced by setsum; infix 'assoc' now has priority 50 (like
  3049 'dvd'); axiom 'one_not_zero' has been moved from axclass 'ring' to
  3050 'domain', this makes the theory consistent with mathematical
  3051 literature;
  3052 
  3053 * HOL basics: added overloaded operations "inverse" and "divide"
  3054 (infix "/"), syntax for generic "abs" operation, generic summation
  3055 operator \<Sum>;
  3056 
  3057 * HOL/typedef: simplified package, provide more useful rules (see also
  3058 HOL/subset.thy);
  3059 
  3060 * HOL/datatype: induction rule for arbitrarily branching datatypes is
  3061 now expressed as a proper nested rule (old-style tactic scripts may
  3062 require atomize_strip_tac to cope with non-atomic premises);
  3063 
  3064 * HOL: renamed theory "Prod" to "Product_Type", renamed "split" rule
  3065 to "split_conv" (old name still available for compatibility);
  3066 
  3067 * HOL: improved concrete syntax for strings (e.g. allows translation
  3068 rules with string literals);
  3069 
  3070 * HOL-Real-Hyperreal: this extends HOL-Real with the hyperreals
  3071  and Fleuriot's mechanization of analysis, including the transcendental
  3072  functions for the reals;
  3073 
  3074 * HOL/Real, HOL/Hyperreal: improved arithmetic simplification;
  3075 
  3076 
  3077 *** CTT ***
  3078 
  3079 * CTT: x-symbol support for Pi, Sigma, -->, : (membership); note that
  3080 "lam" is displayed as TWO lambda-symbols
  3081 
  3082 * CTT: theory Main now available, containing everything (that is, Bool
  3083 and Arith);
  3084 
  3085 
  3086 *** General ***
  3087 
  3088 * Pure: the Simplifier has been implemented properly as a derived rule
  3089 outside of the actual kernel (at last!); the overall performance
  3090 penalty in practical applications is about 50%, while reliability of
  3091 the Isabelle inference kernel has been greatly improved;
  3092 
  3093 * print modes "brackets" and "no_brackets" control output of nested =>
  3094 (types) and ==> (props); the default behaviour is "brackets";
  3095 
  3096 * Provers: fast_tac (and friends) now handle actual object-logic rules
  3097 as assumptions as well;
  3098 
  3099 * system: support Poly/ML 4.0;
  3100 
  3101 * system: isatool install handles KDE version 1 or 2;
  3102 
  3103 
  3104 
  3105 New in Isabelle99-1 (October 2000)
  3106 ----------------------------------
  3107 
  3108 *** Overview of INCOMPATIBILITIES ***
  3109 
  3110 * HOL: simplification of natural numbers is much changed; to partly
  3111 recover the old behaviour (e.g. to prevent n+n rewriting to #2*n)
  3112 issue the following ML commands:
  3113 
  3114   Delsimprocs Nat_Numeral_Simprocs.cancel_numerals;
  3115   Delsimprocs [Nat_Numeral_Simprocs.combine_numerals];
  3116 
  3117 * HOL: simplification no longer dives into case-expressions; this is
  3118 controlled by "t.weak_case_cong" for each datatype t;
  3119 
  3120 * HOL: nat_less_induct renamed to less_induct;
  3121 
  3122 * HOL: systematic renaming of the SOME (Eps) rules, may use isatool
  3123 fixsome to patch .thy and .ML sources automatically;
  3124 
  3125   select_equality  -> some_equality
  3126   select_eq_Ex     -> some_eq_ex
  3127   selectI2EX       -> someI2_ex
  3128   selectI2         -> someI2
  3129   selectI          -> someI
  3130   select1_equality -> some1_equality
  3131   Eps_sym_eq       -> some_sym_eq_trivial
  3132   Eps_eq           -> some_eq_trivial
  3133 
  3134 * HOL: exhaust_tac on datatypes superceded by new generic case_tac;
  3135 
  3136 * HOL: removed obsolete theorem binding expand_if (refer to split_if
  3137 instead);
  3138 
  3139 * HOL: the recursion equations generated by 'recdef' are now called
  3140 f.simps instead of f.rules;
  3141 
  3142 * HOL: qed_spec_mp now also handles bounded ALL as well;
  3143 
  3144 * HOL: 0 is now overloaded, so the type constraint ":: nat" may
  3145 sometimes be needed;
  3146 
  3147 * HOL: the constant for "f``x" is now "image" rather than "op ``";
  3148 
  3149 * HOL: the constant for "f-``x" is now "vimage" rather than "op -``";
  3150 
  3151 * HOL: the disjoint sum is now "<+>" instead of "Plus"; the cartesian
  3152 product is now "<*>" instead of "Times"; the lexicographic product is
  3153 now "<*lex*>" instead of "**";
  3154 
  3155 * HOL: theory Sexp is now in HOL/Induct examples (it used to be part
  3156 of main HOL, but was unused); better use HOL's datatype package;
  3157 
  3158 * HOL: removed "symbols" syntax for constant "override" of theory Map;
  3159 the old syntax may be recovered as follows:
  3160 
  3161   syntax (symbols)
  3162     override  :: "('a ~=> 'b) => ('a ~=> 'b) => ('a ~=> 'b)"
  3163       (infixl "\\<oplus>" 100)
  3164 
  3165 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3166 
  3167 * HOL/ML: even fewer consts are declared as global (see theories Ord,
  3168 Lfp, Gfp, WF); this only affects ML packages that refer to const names
  3169 internally;
  3170 
  3171 * HOL and ZF: syntax for quotienting wrt an equivalence relation
  3172 changed from A/r to A//r;
  3173 
  3174 * ZF: new treatment of arithmetic (nat & int) may break some old
  3175 proofs;
  3176 
  3177 * Isar: renamed some attributes (RS -> THEN, simplify -> simplified,
  3178 rulify -> rule_format, elimify -> elim_format, ...);
  3179 
  3180 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3181 intro/intro!/intro!! flags to intro!/intro/intro? (in most cases, one
  3182 should have to change intro!! to intro? only); replaced "delrule" by
  3183 "rule del";
  3184 
  3185 * Isar/HOL: renamed "intrs" to "intros" in inductive definitions;
  3186 
  3187 * Provers: strengthened force_tac by using new first_best_tac;
  3188 
  3189 * LaTeX document preparation: several changes of isabelle.sty (see
  3190 lib/texinputs);
  3191 
  3192 
  3193 *** Document preparation ***
  3194 
  3195 * formal comments (text blocks etc.) in new-style theories may now
  3196 contain antiquotations of thm/prop/term/typ/text to be presented
  3197 according to latex print mode; concrete syntax is like this:
  3198 @{term[show_types] "f(x) = a + x"};
  3199 
  3200 * isatool mkdir provides easy setup of Isabelle session directories,
  3201 including proper document sources;
  3202 
  3203 * generated LaTeX sources are now deleted after successful run
  3204 (isatool document -c); may retain a copy somewhere else via -D option
  3205 of isatool usedir;
  3206 
  3207 * isatool usedir -D now lets isatool latex -o sty update the Isabelle
  3208 style files, achieving self-contained LaTeX sources and simplifying
  3209 LaTeX debugging;
  3210 
  3211 * old-style theories now produce (crude) LaTeX output as well;
  3212 
  3213 * browser info session directories are now self-contained (may be put
  3214 on WWW server seperately); improved graphs of nested sessions; removed
  3215 graph for 'all sessions';
  3216 
  3217 * several improvements in isabelle style files; \isabellestyle{it}
  3218 produces fake math mode output; \isamarkupheader is now \section by
  3219 default; see lib/texinputs/isabelle.sty etc.;
  3220 
  3221 
  3222 *** Isar ***
  3223 
  3224 * Isar/Pure: local results and corresponding term bindings are now
  3225 subject to Hindley-Milner polymorphism (similar to ML); this
  3226 accommodates incremental type-inference very nicely;
  3227 
  3228 * Isar/Pure: new derived language element 'obtain' supports
  3229 generalized existence reasoning;
  3230 
  3231 * Isar/Pure: new calculational elements 'moreover' and 'ultimately'
  3232 support accumulation of results, without applying any rules yet;
  3233 useful to collect intermediate results without explicit name
  3234 references, and for use with transitivity rules with more than 2
  3235 premises;
  3236 
  3237 * Isar/Pure: scalable support for case-analysis type proofs: new
  3238 'case' language element refers to local contexts symbolically, as
  3239 produced by certain proof methods; internally, case names are attached
  3240 to theorems as "tags";
  3241 
  3242 * Isar/Pure: theory command 'hide' removes declarations from
  3243 class/type/const name spaces;
  3244 
  3245 * Isar/Pure: theory command 'defs' supports option "(overloaded)" to
  3246 indicate potential overloading;
  3247 
  3248 * Isar/Pure: changed syntax of local blocks from {{ }} to { };
  3249 
  3250 * Isar/Pure: syntax of sorts made 'inner', i.e. have to write
  3251 "{a,b,c}" instead of {a,b,c};
  3252 
  3253 * Isar/Pure now provides its own version of intro/elim/dest
  3254 attributes; useful for building new logics, but beware of confusion
  3255 with the version in Provers/classical;
  3256 
  3257 * Isar/Pure: the local context of (non-atomic) goals is provided via
  3258 case name 'antecedent';
  3259 
  3260 * Isar/Pure: removed obsolete 'transfer' attribute (transfer of thms
  3261 to the current context is now done automatically);
  3262 
  3263 * Isar/Pure: theory command 'method_setup' provides a simple interface
  3264 for definining proof methods in ML;
  3265 
  3266 * Isar/Provers: intro/elim/dest attributes changed; renamed
  3267 intro/intro!/intro!! flags to intro!/intro/intro? (INCOMPATIBILITY, in
  3268 most cases, one should have to change intro!! to intro? only);
  3269 replaced "delrule" by "rule del";
  3270 
  3271 * Isar/Provers: new 'hypsubst' method, plain 'subst' method and
  3272 'symmetric' attribute (the latter supercedes [RS sym]);
  3273 
  3274 * Isar/Provers: splitter support (via 'split' attribute and 'simp'
  3275 method modifier); 'simp' method: 'only:' modifier removes loopers as
  3276 well (including splits);
  3277 
  3278 * Isar/Provers: Simplifier and Classical methods now support all kind
  3279 of modifiers used in the past, including 'cong', 'iff', etc.
  3280 
  3281 * Isar/Provers: added 'fastsimp' and 'clarsimp' methods (combination
  3282 of Simplifier and Classical reasoner);
  3283 
  3284 * Isar/HOL: new proof method 'cases' and improved version of 'induct'
  3285 now support named cases; major packages (inductive, datatype, primrec,
  3286 recdef) support case names and properly name parameters;
  3287 
  3288 * Isar/HOL: new transitivity rules for substitution in inequalities --
  3289 monotonicity conditions are extracted to be proven at end of
  3290 calculations;
  3291 
  3292 * Isar/HOL: removed 'case_split' thm binding, should use 'cases' proof
  3293 method anyway;
  3294 
  3295 * Isar/HOL: removed old expand_if = split_if; theorems if_splits =
  3296 split_if split_if_asm; datatype package provides theorems foo.splits =
  3297 foo.split foo.split_asm for each datatype;
  3298 
  3299 * Isar/HOL: tuned inductive package, rename "intrs" to "intros"
  3300 (potential INCOMPATIBILITY), emulation of mk_cases feature for proof
  3301 scripts: new 'inductive_cases' command and 'ind_cases' method; (Note:
  3302 use "(cases (simplified))" method in proper proof texts);
  3303 
  3304 * Isar/HOL: added global 'arith_split' attribute for 'arith' method;
  3305 
  3306 * Isar: names of theorems etc. may be natural numbers as well;
  3307 
  3308 * Isar: 'pr' command: optional arguments for goals_limit and
  3309 ProofContext.prems_limit; no longer prints theory contexts, but only
  3310 proof states;
  3311 
  3312 * Isar: diagnostic commands 'pr', 'thm', 'prop', 'term', 'typ' admit
  3313 additional print modes to be specified; e.g. "pr(latex)" will print
  3314 proof state according to the Isabelle LaTeX style;
  3315 
  3316 * Isar: improved support for emulating tactic scripts, including proof
  3317 methods 'rule_tac' etc., 'cut_tac', 'thin_tac', 'subgoal_tac',
  3318 'rename_tac', 'rotate_tac', 'tactic', and 'case_tac' / 'induct_tac'
  3319 (for HOL datatypes);
  3320 
  3321 * Isar: simplified (more robust) goal selection of proof methods: 1st
  3322 goal, all goals, or explicit goal specifier (tactic emulation); thus
  3323 'proof method scripts' have to be in depth-first order;
  3324 
  3325 * Isar: tuned 'let' syntax: replaced 'as' keyword by 'and';
  3326 
  3327 * Isar: removed 'help' command, which hasn't been too helpful anyway;
  3328 should instead use individual commands for printing items
  3329 (print_commands, print_methods etc.);
  3330 
  3331 * Isar: added 'nothing' --- the empty list of theorems;
  3332 
  3333 
  3334 *** HOL ***
  3335 
  3336 * HOL/MicroJava: formalization of a fragment of Java, together with a
  3337 corresponding virtual machine and a specification of its bytecode
  3338 verifier and a lightweight bytecode verifier, including proofs of
  3339 type-safety; by Gerwin Klein, Tobias Nipkow, David von Oheimb, and
  3340 Cornelia Pusch (see also the homepage of project Bali at
  3341 http://isabelle.in.tum.de/Bali/);
  3342 
  3343 * HOL/Algebra: new theory of rings and univariate polynomials, by
  3344 Clemens Ballarin;
  3345 
  3346 * HOL/NumberTheory: fundamental Theorem of Arithmetic, Chinese
  3347 Remainder Theorem, Fermat/Euler Theorem, Wilson's Theorem, by Thomas M
  3348 Rasmussen;
  3349 
  3350 * HOL/Lattice: fundamental concepts of lattice theory and order
  3351 structures, including duals, properties of bounds versus algebraic
  3352 laws, lattice operations versus set-theoretic ones, the Knaster-Tarski
  3353 Theorem for complete lattices etc.; may also serve as a demonstration
  3354 for abstract algebraic reasoning using axiomatic type classes, and
  3355 mathematics-style proof in Isabelle/Isar; by Markus Wenzel;
  3356 
  3357 * HOL/Prolog: a (bare-bones) implementation of Lambda-Prolog, by David
  3358 von Oheimb;
  3359 
  3360 * HOL/IMPP: extension of IMP with local variables and mutually
  3361 recursive procedures, by David von Oheimb;
  3362 
  3363 * HOL/Lambda: converted into new-style theory and document;
  3364 
  3365 * HOL/ex/Multiquote: example of multiple nested quotations and
  3366 anti-quotations -- basically a generalized version of de-Bruijn
  3367 representation; very useful in avoiding lifting of operations;
  3368 
  3369 * HOL/record: added general record equality rule to simpset; fixed
  3370 select-update simplification procedure to handle extended records as
  3371 well; admit "r" as field name;
  3372 
  3373 * HOL: 0 is now overloaded over the new sort "zero", allowing its use with
  3374 other numeric types and also as the identity of groups, rings, etc.;
  3375 
  3376 * HOL: new axclass plus_ac0 for addition with the AC-laws and 0 as identity.
  3377 Types nat and int belong to this axclass;
  3378 
  3379 * HOL: greatly improved simplification involving numerals of type nat, int, real:
  3380    (i + #8 + j) = Suc k simplifies to  #7 + (i + j) = k
  3381    i*j + k + j*#3*i     simplifies to  #4*(i*j) + k
  3382   two terms #m*u and #n*u are replaced by #(m+n)*u
  3383     (where #m, #n and u can implicitly be 1; this is simproc combine_numerals)
  3384   and the term/formula #m*u+x ~~ #n*u+y simplifies simplifies to #(m-n)+x ~~ y
  3385     or x ~~ #(n-m)+y, where ~~ is one of = < <= or - (simproc cancel_numerals);
  3386 
  3387 * HOL: meson_tac is available (previously in ex/meson.ML); it is a
  3388 powerful prover for predicate logic but knows nothing of clasets; see
  3389 ex/mesontest.ML and ex/mesontest2.ML for example applications;
  3390 
  3391 * HOL: new version of "case_tac" subsumes both boolean case split and
  3392 "exhaust_tac" on datatypes; INCOMPATIBILITY: exhaust_tac no longer
  3393 exists, may define val exhaust_tac = case_tac for ad-hoc portability;
  3394 
  3395 * HOL: simplification no longer dives into case-expressions: only the
  3396 selector expression is simplified, but not the remaining arms; to
  3397 enable full simplification of case-expressions for datatype t, you may
  3398 remove t.weak_case_cong from the simpset, either globally (Delcongs
  3399 [thm"t.weak_case_cong"];) or locally (delcongs [...]).
  3400 
  3401 * HOL/recdef: the recursion equations generated by 'recdef' for
  3402 function 'f' are now called f.simps instead of f.rules; if all
  3403 termination conditions are proved automatically, these simplification
  3404 rules are added to the simpset, as in primrec; rules may be named
  3405 individually as well, resulting in a separate list of theorems for
  3406 each equation;
  3407 
  3408 * HOL/While is a new theory that provides a while-combinator. It
  3409 permits the definition of tail-recursive functions without the
  3410 provision of a termination measure. The latter is necessary once the
  3411 invariant proof rule for while is applied.
  3412 
  3413 * HOL: new (overloaded) notation for the set of elements below/above
  3414 some element: {..u}, {..u(}, {l..}, {)l..}. See theory SetInterval.
  3415 
  3416 * HOL: theorems impI, allI, ballI bound as "strip";
  3417 
  3418 * HOL: new tactic induct_thm_tac: thm -> string -> int -> tactic
  3419 induct_tac th "x1 ... xn" expects th to have a conclusion of the form
  3420 P v1 ... vn and abbreviates res_inst_tac [("v1","x1"),...,("vn","xn")] th;
  3421 
  3422 * HOL/Real: "rabs" replaced by overloaded "abs" function;
  3423 
  3424 * HOL: theory Sexp now in HOL/Induct examples (it used to be part of
  3425 main HOL, but was unused);
  3426 
  3427 * HOL: fewer consts declared as global (e.g. have to refer to
  3428 "Lfp.lfp" instead of "lfp" internally; affects ML packages only);
  3429 
  3430 * HOL: tuned AST representation of nested pairs, avoiding bogus output
  3431 in case of overlap with user translations (e.g. judgements over
  3432 tuples); (note that the underlying logical represenation is still
  3433 bogus);
  3434 
  3435 
  3436 *** ZF ***
  3437 
  3438 * ZF: simplification automatically cancels common terms in arithmetic
  3439 expressions over nat and int;
  3440 
  3441 * ZF: new treatment of nat to minimize type-checking: all operators
  3442 coerce their operands to a natural number using the function natify,
  3443 making the algebraic laws unconditional;
  3444 
  3445 * ZF: as above, for int: operators coerce their operands to an integer
  3446 using the function intify;
  3447 
  3448 * ZF: the integer library now contains many of the usual laws for the
  3449 orderings, including $<=, and monotonicity laws for $+ and $*;
  3450 
  3451 * ZF: new example ZF/ex/NatSum to demonstrate integer arithmetic
  3452 simplification;
  3453 
  3454 * FOL and ZF: AddIffs now available, giving theorems of the form P<->Q
  3455 to the simplifier and classical reasoner simultaneously;
  3456 
  3457 
  3458 *** General ***
  3459 
  3460 * Provers: blast_tac now handles actual object-logic rules as
  3461 assumptions; note that auto_tac uses blast_tac internally as well;
  3462 
  3463 * Provers: new functions rulify/rulify_no_asm: thm -> thm for turning
  3464 outer -->/All/Ball into ==>/!!; qed_spec_mp now uses rulify_no_asm;
  3465 
  3466 * Provers: delrules now handles destruct rules as well (no longer need
  3467 explicit make_elim);
  3468 
  3469 * Provers: Blast_tac now warns of and ignores "weak elimination rules" e.g.
  3470   [| inj ?f;          ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3471 use instead the strong form,
  3472   [| inj ?f; ~ ?W ==> ?f ?x = ?f ?y; ?x = ?y ==> ?W |] ==> ?W
  3473 in HOL, FOL and ZF the function cla_make_elim will create such rules
  3474 from destruct-rules;
  3475 
  3476 * Provers: Simplifier.easy_setup provides a fast path to basic
  3477 Simplifier setup for new object-logics;
  3478 
  3479 * Pure: AST translation rules no longer require constant head on LHS;
  3480 
  3481 * Pure: improved name spaces: ambiguous output is qualified; support
  3482 for hiding of names;
  3483 
  3484 * system: smart setup of canonical ML_HOME, ISABELLE_INTERFACE, and
  3485 XSYMBOL_HOME; no longer need to do manual configuration in most
  3486 situations;
  3487 
  3488 * system: compression of ML heaps images may now be controlled via -c
  3489 option of isabelle and isatool usedir (currently only observed by
  3490 Poly/ML);
  3491 
  3492 * system: isatool installfonts may handle X-Symbol fonts as well (very
  3493 useful for remote X11);
  3494 
  3495 * system: provide TAGS file for Isabelle sources;
  3496 
  3497 * ML: infix 'OF' is a version of 'MRS' with more appropriate argument
  3498 order;
  3499 
  3500 * ML: renamed flags Syntax.trace_norm_ast to Syntax.trace_ast; global
  3501 timing flag supersedes proof_timing and Toplevel.trace;
  3502 
  3503 * ML: new combinators |>> and |>>> for incremental transformations
  3504 with secondary results (e.g. certain theory extensions):
  3505 
  3506 * ML: PureThy.add_defs gets additional argument to indicate potential
  3507 overloading (usually false);
  3508 
  3509 * ML: PureThy.add_thms/add_axioms/add_defs now return theorems as
  3510 results;
  3511 
  3512 
  3513 
  3514 New in Isabelle99 (October 1999)
  3515 --------------------------------
  3516 
  3517 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3518 
  3519 * HOL: The THEN and ELSE parts of conditional expressions (if P then x else y)
  3520 are no longer simplified.  (This allows the simplifier to unfold recursive
  3521 functional programs.)  To restore the old behaviour, declare
  3522 
  3523     Delcongs [if_weak_cong];
  3524 
  3525 * HOL: Removed the obsolete syntax "Compl A"; use -A for set
  3526 complement;
  3527 
  3528 * HOL: the predicate "inj" is now defined by translation to "inj_on";
  3529 
  3530 * HOL/datatype: mutual_induct_tac no longer exists --
  3531   use induct_tac "x_1 ... x_n" instead of mutual_induct_tac ["x_1", ..., "x_n"]
  3532 
  3533 * HOL/typedef: fixed type inference for representing set; type
  3534 arguments now have to occur explicitly on the rhs as type constraints;
  3535 
  3536 * ZF: The con_defs part of an inductive definition may no longer refer
  3537 to constants declared in the same theory;
  3538 
  3539 * HOL, ZF: the function mk_cases, generated by the inductive
  3540 definition package, has lost an argument.  To simplify its result, it
  3541 uses the default simpset instead of a supplied list of theorems.
  3542 
  3543 * HOL/List: the constructors of type list are now Nil and Cons;
  3544 
  3545 * Simplifier: the type of the infix ML functions
  3546         setSSolver addSSolver setSolver addSolver
  3547 is now  simpset * solver -> simpset  where `solver' is a new abstract type
  3548 for packaging solvers. A solver is created via
  3549         mk_solver: string -> (thm list -> int -> tactic) -> solver
  3550 where the string argument is only a comment.
  3551 
  3552 
  3553 *** Proof tools ***
  3554 
  3555 * Provers/Arith/fast_lin_arith.ML contains a functor for creating a
  3556 decision procedure for linear arithmetic. Currently it is used for
  3557 types `nat', `int', and `real' in HOL (see below); it can, should and
  3558 will be instantiated for other types and logics as well.
  3559 
  3560 * The simplifier now accepts rewrite rules with flexible heads, eg
  3561      hom ?f ==> ?f(?x+?y) = ?f ?x + ?f ?y
  3562   They are applied like any rule with a non-pattern lhs, i.e. by first-order
  3563   matching.
  3564 
  3565 
  3566 *** General ***
  3567 
  3568 * New Isabelle/Isar subsystem provides an alternative to traditional
  3569 tactical theorem proving; together with the ProofGeneral/isar user
  3570 interface it offers an interactive environment for developing human
  3571 readable proof documents (Isar == Intelligible semi-automated
  3572 reasoning); for further information see isatool doc isar-ref,
  3573 src/HOL/Isar_examples and http://isabelle.in.tum.de/Isar/
  3574 
  3575 * improved and simplified presentation of theories: better HTML markup
  3576 (including colors), graph views in several sizes; isatool usedir now
  3577 provides a proper interface for user theories (via -P option); actual
  3578 document preparation based on (PDF)LaTeX is available as well (for
  3579 new-style theories only); see isatool doc system for more information;
  3580 
  3581 * native support for Proof General, both for classic Isabelle and
  3582 Isabelle/Isar;
  3583 
  3584 * ML function thm_deps visualizes dependencies of theorems and lemmas,
  3585 using the graph browser tool;
  3586 
  3587 * Isabelle manuals now also available as PDF;
  3588 
  3589 * theory loader rewritten from scratch (may not be fully
  3590 bug-compatible); old loadpath variable has been replaced by show_path,
  3591 add_path, del_path, reset_path functions; new operations such as
  3592 update_thy, touch_thy, remove_thy, use/update_thy_only (see also
  3593 isatool doc ref);
  3594 
  3595 * improved isatool install: option -k creates KDE application icon,
  3596 option -p DIR installs standalone binaries;
  3597 
  3598 * added ML_PLATFORM setting (useful for cross-platform installations);
  3599 more robust handling of platform specific ML images for SML/NJ;
  3600 
  3601 * the settings environment is now statically scoped, i.e. it is never
  3602 created again in sub-processes invoked from isabelle, isatool, or
  3603 Isabelle;
  3604 
  3605 * path element specification '~~' refers to '$ISABELLE_HOME';
  3606 
  3607 * in locales, the "assumes" and "defines" parts may be omitted if
  3608 empty;
  3609 
  3610 * new print_mode "xsymbols" for extended symbol support (e.g. genuine
  3611 long arrows);
  3612 
  3613 * new print_mode "HTML";
  3614 
  3615 * new flag show_tags controls display of tags of theorems (which are
  3616 basically just comments that may be attached by some tools);
  3617 
  3618 * Isamode 2.6 requires patch to accomodate change of Isabelle font
  3619 mode and goal output format:
  3620 
  3621 diff -r Isamode-2.6/elisp/isa-load.el Isamode/elisp/isa-load.el
  3622 244c244
  3623 <       (list (isa-getenv "ISABELLE") "-msymbols" logic-name)
  3624 ---
  3625 >       (list (isa-getenv "ISABELLE") "-misabelle_font" "-msymbols" logic-name)
  3626 diff -r Isabelle-2.6/elisp/isa-proofstate.el Isamode/elisp/isa-proofstate.el
  3627 181c181
  3628 < (defconst proofstate-proofstart-regexp "^Level [0-9]+$"
  3629 ---
  3630 > (defconst proofstate-proofstart-regexp "^Level [0-9]+"
  3631 
  3632 * function bind_thms stores lists of theorems (cf. bind_thm);
  3633 
  3634 * new shorthand tactics ftac, eatac, datac, fatac;
  3635 
  3636 * qed (and friends) now accept "" as result name; in that case the
  3637 theorem is not stored, but proper checks and presentation of the
  3638 result still apply;
  3639 
  3640 * theorem database now also indexes constants "Trueprop", "all",
  3641 "==>", "=="; thus thms_containing, findI etc. may retrieve more rules;
  3642 
  3643 
  3644 *** HOL ***
  3645 
  3646 ** HOL arithmetic **
  3647 
  3648 * There are now decision procedures for linear arithmetic over nat and
  3649 int:
  3650 
  3651 1. arith_tac copes with arbitrary formulae involving `=', `<', `<=',
  3652 `+', `-', `Suc', `min', `max' and numerical constants; other subterms
  3653 are treated as atomic; subformulae not involving type `nat' or `int'
  3654 are ignored; quantified subformulae are ignored unless they are
  3655 positive universal or negative existential. The tactic has to be
  3656 invoked by hand and can be a little bit slow. In particular, the
  3657 running time is exponential in the number of occurrences of `min' and
  3658 `max', and `-' on `nat'.
  3659 
  3660 2. fast_arith_tac is a cut-down version of arith_tac: it only takes
  3661 (negated) (in)equalities among the premises and the conclusion into
  3662 account (i.e. no compound formulae) and does not know about `min' and
  3663 `max', and `-' on `nat'. It is fast and is used automatically by the
  3664 simplifier.
  3665 
  3666 NB: At the moment, these decision procedures do not cope with mixed
  3667 nat/int formulae where the two parts interact, such as `m < n ==>
  3668 int(m) < int(n)'.
  3669 
  3670 * HOL/Numeral provides a generic theory of numerals (encoded
  3671 efficiently as bit strings); setup for types nat/int/real is in place;
  3672 INCOMPATIBILITY: since numeral syntax is now polymorphic, rather than
  3673 int, existing theories and proof scripts may require a few additional
  3674 type constraints;
  3675 
  3676 * integer division and remainder can now be performed on constant
  3677 arguments;
  3678 
  3679 * many properties of integer multiplication, division and remainder
  3680 are now available;
  3681 
  3682 * An interface to the Stanford Validity Checker (SVC) is available through the
  3683 tactic svc_tac.  Propositional tautologies and theorems of linear arithmetic
  3684 are proved automatically.  SVC must be installed separately, and its results
  3685 must be TAKEN ON TRUST (Isabelle does not check the proofs, but tags any
  3686 invocation of the underlying oracle).  For SVC see
  3687   http://verify.stanford.edu/SVC
  3688 
  3689 * IsaMakefile: the HOL-Real target now builds an actual image;
  3690 
  3691 
  3692 ** HOL misc **
  3693 
  3694 * HOL/Real/HahnBanach: the Hahn-Banach theorem for real vector spaces
  3695 (in Isabelle/Isar) -- by Gertrud Bauer;
  3696 
  3697 * HOL/BCV: generic model of bytecode verification, i.e. data-flow
  3698 analysis for assembly languages with subtypes;
  3699 
  3700 * HOL/TLA (Lamport's Temporal Logic of Actions): major reorganization
  3701 -- avoids syntactic ambiguities and treats state, transition, and
  3702 temporal levels more uniformly; introduces INCOMPATIBILITIES due to
  3703 changed syntax and (many) tactics;
  3704 
  3705 * HOL/inductive: Now also handles more general introduction rules such
  3706   as "ALL y. (y, x) : r --> y : acc r ==> x : acc r"; monotonicity
  3707   theorems are now maintained within the theory (maintained via the
  3708   "mono" attribute);
  3709 
  3710 * HOL/datatype: Now also handles arbitrarily branching datatypes
  3711   (using function types) such as
  3712 
  3713   datatype 'a tree = Atom 'a | Branch "nat => 'a tree"
  3714 
  3715 * HOL/record: record_simproc (part of the default simpset) takes care
  3716 of selectors applied to updated records; record_split_tac is no longer
  3717 part of the default claset; update_defs may now be removed from the
  3718 simpset in many cases; COMPATIBILITY: old behavior achieved by
  3719 
  3720   claset_ref () := claset() addSWrapper record_split_wrapper;
  3721   Delsimprocs [record_simproc]
  3722 
  3723 * HOL/typedef: fixed type inference for representing set; type
  3724 arguments now have to occur explicitly on the rhs as type constraints;
  3725 
  3726 * HOL/recdef (TFL): 'congs' syntax now expects comma separated list of theorem
  3727 names rather than an ML expression;
  3728 
  3729 * HOL/defer_recdef (TFL): like recdef but the well-founded relation can be
  3730 supplied later.  Program schemes can be defined, such as
  3731     "While B C s = (if B s then While B C (C s) else s)"
  3732 where the well-founded relation can be chosen after B and C have been given.
  3733 
  3734 * HOL/List: the constructors of type list are now Nil and Cons;
  3735 INCOMPATIBILITY: while [] and infix # syntax is still there, of
  3736 course, ML tools referring to List.list.op # etc. have to be adapted;
  3737 
  3738 * HOL_quantifiers flag superseded by "HOL" print mode, which is
  3739 disabled by default; run isabelle with option -m HOL to get back to
  3740 the original Gordon/HOL-style output;
  3741 
  3742 * HOL/Ord.thy: new bounded quantifier syntax (input only): ALL x<y. P,
  3743 ALL x<=y. P, EX x<y. P, EX x<=y. P;
  3744 
  3745 * HOL basic syntax simplified (more orthogonal): all variants of
  3746 All/Ex now support plain / symbolic / HOL notation; plain syntax for
  3747 Eps operator is provided as well: "SOME x. P[x]";
  3748 
  3749 * HOL/Sum.thy: sum_case has been moved to HOL/Datatype;
  3750 
  3751 * HOL/Univ.thy: infix syntax <*>, <+>, <**>, <+> eliminated and made
  3752 thus available for user theories;
  3753 
  3754 * HOLCF/IOA/Sequents: renamed 'Cons' to 'Consq' to avoid clash with
  3755 HOL/List; hardly an INCOMPATIBILITY since '>>' syntax is used all the
  3756 time;
  3757 
  3758 * HOL: new tactic smp_tac: int -> int -> tactic, which applies spec
  3759 several times and then mp;
  3760 
  3761 
  3762 *** LK ***
  3763 
  3764 * the notation <<...>> is now available as a notation for sequences of
  3765 formulas;
  3766 
  3767 * the simplifier is now installed
  3768 
  3769 * the axiom system has been generalized (thanks to Soren Heilmann)
  3770 
  3771 * the classical reasoner now has a default rule database
  3772 
  3773 
  3774 *** ZF ***
  3775 
  3776 * new primrec section allows primitive recursive functions to be given
  3777 directly (as in HOL) over datatypes and the natural numbers;
  3778 
  3779 * new tactics induct_tac and exhaust_tac for induction (or case
  3780 analysis) over datatypes and the natural numbers;
  3781 
  3782 * the datatype declaration of type T now defines the recursor T_rec;
  3783 
  3784 * simplification automatically does freeness reasoning for datatype
  3785 constructors;
  3786 
  3787 * automatic type-inference, with AddTCs command to insert new
  3788 type-checking rules;
  3789 
  3790 * datatype introduction rules are now added as Safe Introduction rules
  3791 to the claset;
  3792 
  3793 * the syntax "if P then x else y" is now available in addition to
  3794 if(P,x,y);
  3795 
  3796 
  3797 *** Internal programming interfaces ***
  3798 
  3799 * tuned simplifier trace output; new flag debug_simp;
  3800 
  3801 * structures Vartab / Termtab (instances of TableFun) offer efficient
  3802 tables indexed by indexname_ord / term_ord (compatible with aconv);
  3803 
  3804 * AxClass.axclass_tac lost the theory argument;
  3805 
  3806 * tuned current_goals_markers semantics: begin / end goal avoids
  3807 printing empty lines;
  3808 
  3809 * removed prs and prs_fn hook, which was broken because it did not
  3810 include \n in its semantics, forcing writeln to add one
  3811 uncoditionally; replaced prs_fn by writeln_fn; consider std_output:
  3812 string -> unit if you really want to output text without newline;
  3813 
  3814 * Symbol.output subject to print mode; INCOMPATIBILITY: defaults to
  3815 plain output, interface builders may have to enable 'isabelle_font'
  3816 mode to get Isabelle font glyphs as before;
  3817 
  3818 * refined token_translation interface; INCOMPATIBILITY: output length
  3819 now of type real instead of int;
  3820 
  3821 * theory loader actions may be traced via new ThyInfo.add_hook
  3822 interface (see src/Pure/Thy/thy_info.ML); example application: keep
  3823 your own database of information attached to *whole* theories -- as
  3824 opposed to intra-theory data slots offered via TheoryDataFun;
  3825 
  3826 * proper handling of dangling sort hypotheses (at last!);
  3827 Thm.strip_shyps and Drule.strip_shyps_warning take care of removing
  3828 extra sort hypotheses that can be witnessed from the type signature;
  3829 the force_strip_shyps flag is gone, any remaining shyps are simply
  3830 left in the theorem (with a warning issued by strip_shyps_warning);
  3831 
  3832 
  3833 
  3834 New in Isabelle98-1 (October 1998)
  3835 ----------------------------------
  3836 
  3837 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  3838 
  3839 * several changes of automated proof tools;
  3840 
  3841 * HOL: major changes to the inductive and datatype packages, including
  3842 some minor incompatibilities of theory syntax;
  3843 
  3844 * HOL: renamed r^-1 to 'converse' from 'inverse'; 'inj_onto' is now
  3845 called `inj_on';
  3846 
  3847 * HOL: removed duplicate thms in Arith:
  3848   less_imp_add_less  should be replaced by  trans_less_add1
  3849   le_imp_add_le      should be replaced by  trans_le_add1
  3850 
  3851 * HOL: unary minus is now overloaded (new type constraints may be
  3852 required);
  3853 
  3854 * HOL and ZF: unary minus for integers is now #- instead of #~.  In
  3855 ZF, expressions such as n#-1 must be changed to n#- 1, since #-1 is
  3856 now taken as an integer constant.
  3857 
  3858 * Pure: ML function 'theory_of' renamed to 'theory';
  3859 
  3860 
  3861 *** Proof tools ***
  3862 
  3863 * Simplifier:
  3864   1. Asm_full_simp_tac is now more aggressive.
  3865      1. It will sometimes reorient premises if that increases their power to
  3866         simplify.
  3867      2. It does no longer proceed strictly from left to right but may also
  3868         rotate premises to achieve further simplification.
  3869      For compatibility reasons there is now Asm_lr_simp_tac which is like the
  3870      old Asm_full_simp_tac in that it does not rotate premises.
  3871   2. The simplifier now knows a little bit about nat-arithmetic.
  3872 
  3873 * Classical reasoner: wrapper mechanism for the classical reasoner now
  3874 allows for selected deletion of wrappers, by introduction of names for
  3875 wrapper functionals.  This implies that addbefore, addSbefore,
  3876 addaltern, and addSaltern now take a pair (name, tactic) as argument,
  3877 and that adding two tactics with the same name overwrites the first
  3878 one (emitting a warning).
  3879   type wrapper = (int -> tactic) -> (int -> tactic)
  3880   setWrapper, setSWrapper, compWrapper and compSWrapper are replaced by
  3881   addWrapper, addSWrapper: claset * (string * wrapper) -> claset
  3882   delWrapper, delSWrapper: claset *  string            -> claset
  3883   getWrapper is renamed to appWrappers, getSWrapper to appSWrappers;
  3884 
  3885 * Classical reasoner: addbefore/addSbefore now have APPEND/ORELSE
  3886 semantics; addbefore now affects only the unsafe part of step_tac
  3887 etc.; this affects addss/auto_tac/force_tac, so EXISTING PROOFS MAY
  3888 FAIL, but proofs should be fixable easily, e.g. by replacing Auto_tac
  3889 by Force_tac;
  3890 
  3891 * Classical reasoner: setwrapper to setWrapper and compwrapper to
  3892 compWrapper; added safe wrapper (and access functions for it);
  3893 
  3894 * HOL/split_all_tac is now much faster and fails if there is nothing
  3895 to split.  Some EXISTING PROOFS MAY REQUIRE ADAPTION because the order
  3896 and the names of the automatically generated variables have changed.
  3897 split_all_tac has moved within claset() from unsafe wrappers to safe
  3898 wrappers, which means that !!-bound variables are split much more
  3899 aggressively, and safe_tac and clarify_tac now split such variables.
  3900 If this splitting is not appropriate, use delSWrapper "split_all_tac".
  3901 Note: the same holds for record_split_tac, which does the job of
  3902 split_all_tac for record fields.
  3903 
  3904 * HOL/Simplifier: Rewrite rules for case distinctions can now be added
  3905 permanently to the default simpset using Addsplits just like
  3906 Addsimps. They can be removed via Delsplits just like
  3907 Delsimps. Lower-case versions are also available.
  3908 
  3909 * HOL/Simplifier: The rule split_if is now part of the default
  3910 simpset. This means that the simplifier will eliminate all occurrences
  3911 of if-then-else in the conclusion of a goal. To prevent this, you can
  3912 either remove split_if completely from the default simpset by
  3913 `Delsplits [split_if]' or remove it in a specific call of the
  3914 simplifier using `... delsplits [split_if]'.  You can also add/delete
  3915 other case splitting rules to/from the default simpset: every datatype
  3916 generates suitable rules `split_t_case' and `split_t_case_asm' (where
  3917 t is the name of the datatype).
  3918 
  3919 * Classical reasoner / Simplifier combination: new force_tac (and
  3920 derivatives Force_tac, force) combines rewriting and classical
  3921 reasoning (and whatever other tools) similarly to auto_tac, but is
  3922 aimed to solve the given subgoal completely.
  3923 
  3924 
  3925 *** General ***
  3926 
  3927 * new top-level commands `Goal' and `Goalw' that improve upon `goal'
  3928 and `goalw': the theory is no longer needed as an explicit argument -
  3929 the current theory context is used; assumptions are no longer returned
  3930 at the ML-level unless one of them starts with ==> or !!; it is
  3931 recommended to convert to these new commands using isatool fixgoal
  3932 (backup your sources first!);
  3933 
  3934 * new top-level commands 'thm' and 'thms' for retrieving theorems from
  3935 the current theory context, and 'theory' to lookup stored theories;
  3936 
  3937 * new theory section 'locale' for declaring constants, assumptions and
  3938 definitions that have local scope;
  3939 
  3940 * new theory section 'nonterminals' for purely syntactic types;
  3941 
  3942 * new theory section 'setup' for generic ML setup functions
  3943 (e.g. package initialization);
  3944 
  3945 * the distribution now includes Isabelle icons: see
  3946 lib/logo/isabelle-{small,tiny}.xpm;
  3947 
  3948 * isatool install - install binaries with absolute references to
  3949 ISABELLE_HOME/bin;
  3950 
  3951 * isatool logo -- create instances of the Isabelle logo (as EPS);
  3952 
  3953 * print mode 'emacs' reserved for Isamode;
  3954 
  3955 * support multiple print (ast) translations per constant name;
  3956 
  3957 * theorems involving oracles are now printed with a suffixed [!];
  3958 
  3959 
  3960 *** HOL ***
  3961 
  3962 * there is now a tutorial on Isabelle/HOL (do 'isatool doc tutorial');
  3963 
  3964 * HOL/inductive package reorganized and improved: now supports mutual
  3965 definitions such as
  3966 
  3967   inductive EVEN ODD
  3968     intrs
  3969       null "0 : EVEN"
  3970       oddI "n : EVEN ==> Suc n : ODD"
  3971       evenI "n : ODD ==> Suc n : EVEN"
  3972 
  3973 new theorem list "elims" contains an elimination rule for each of the
  3974 recursive sets; inductive definitions now handle disjunctive premises
  3975 correctly (also ZF);
  3976 
  3977 INCOMPATIBILITIES: requires Inductive as an ancestor; component
  3978 "mutual_induct" no longer exists - the induction rule is always
  3979 contained in "induct";
  3980 
  3981 
  3982 * HOL/datatype package re-implemented and greatly improved: now
  3983 supports mutually recursive datatypes such as
  3984 
  3985   datatype
  3986     'a aexp = IF_THEN_ELSE ('a bexp) ('a aexp) ('a aexp)
  3987             | SUM ('a aexp) ('a aexp)
  3988             | DIFF ('a aexp) ('a aexp)
  3989             | NUM 'a
  3990   and
  3991     'a bexp = LESS ('a aexp) ('a aexp)
  3992             | AND ('a bexp) ('a bexp)
  3993             | OR ('a bexp) ('a bexp)
  3994 
  3995 as well as indirectly recursive datatypes such as
  3996 
  3997   datatype
  3998     ('a, 'b) term = Var 'a
  3999                   | App 'b ((('a, 'b) term) list)
  4000 
  4001 The new tactic  mutual_induct_tac [<var_1>, ..., <var_n>] i  performs
  4002 induction on mutually / indirectly recursive datatypes.
  4003 
  4004 Primrec equations are now stored in theory and can be accessed via
  4005 <function_name>.simps.
  4006 
  4007 INCOMPATIBILITIES:
  4008 
  4009   - Theories using datatypes must now have theory Datatype as an
  4010     ancestor.
  4011   - The specific <typename>.induct_tac no longer exists - use the
  4012     generic induct_tac instead.
  4013   - natE has been renamed to nat.exhaust - use exhaust_tac
  4014     instead of res_inst_tac ... natE. Note that the variable
  4015     names in nat.exhaust differ from the names in natE, this
  4016     may cause some "fragile" proofs to fail.
  4017   - The theorems split_<typename>_case and split_<typename>_case_asm
  4018     have been renamed to <typename>.split and <typename>.split_asm.
  4019   - Since default sorts of type variables are now handled correctly,
  4020     some datatype definitions may have to be annotated with explicit
  4021     sort constraints.
  4022   - Primrec definitions no longer require function name and type
  4023     of recursive argument.
  4024 
  4025 Consider using isatool fixdatatype to adapt your theories and proof
  4026 scripts to the new package (backup your sources first!).
  4027 
  4028 
  4029 * HOL/record package: considerably improved implementation; now
  4030 includes concrete syntax for record types, terms, updates; theorems
  4031 for surjective pairing and splitting !!-bound record variables; proof
  4032 support is as follows:
  4033 
  4034   1) standard conversions (selectors or updates applied to record
  4035 constructor terms) are part of the standard simpset;
  4036 
  4037   2) inject equations of the form ((x, y) = (x', y')) == x=x' & y=y' are
  4038 made part of standard simpset and claset via addIffs;
  4039 
  4040   3) a tactic for record field splitting (record_split_tac) is part of
  4041 the standard claset (addSWrapper);
  4042 
  4043 To get a better idea about these rules you may retrieve them via
  4044 something like 'thms "foo.simps"' or 'thms "foo.iffs"', where "foo" is
  4045 the name of your record type.
  4046 
  4047 The split tactic 3) conceptually simplifies by the following rule:
  4048 
  4049   "(!!x. PROP ?P x) == (!!a b. PROP ?P (a, b))"
  4050 
  4051 Thus any record variable that is bound by meta-all will automatically
  4052 blow up into some record constructor term, consequently the
  4053 simplifications of 1), 2) apply.  Thus force_tac, auto_tac etc. shall
  4054 solve record problems automatically.
  4055 
  4056 
  4057 * reorganized the main HOL image: HOL/Integ and String loaded by
  4058 default; theory Main includes everything;
  4059 
  4060 * automatic simplification of integer sums and comparisons, using cancellation;
  4061 
  4062 * added option_map_eq_Some and not_Some_eq to the default simpset and claset;
  4063 
  4064 * added disj_not1 = "(~P | Q) = (P --> Q)" to the default simpset;
  4065 
  4066 * many new identities for unions, intersections, set difference, etc.;
  4067 
  4068 * expand_if, expand_split, expand_sum_case and expand_nat_case are now
  4069 called split_if, split_split, split_sum_case and split_nat_case (to go
  4070 with add/delsplits);
  4071 
  4072 * HOL/Prod introduces simplification procedure unit_eq_proc rewriting
  4073 (?x::unit) = (); this is made part of the default simpset, which COULD
  4074 MAKE EXISTING PROOFS FAIL under rare circumstances (consider
  4075 'Delsimprocs [unit_eq_proc];' as last resort); also note that
  4076 unit_abs_eta_conv is added in order to counter the effect of
  4077 unit_eq_proc on (%u::unit. f u), replacing it by f rather than by
  4078 %u.f();
  4079 
  4080 * HOL/Fun INCOMPATIBILITY: `inj_onto' is now called `inj_on' (which
  4081 makes more sense);
  4082 
  4083 * HOL/Set INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4084   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4085   disjointness reasoning but breaking a few old proofs.
  4086 
  4087 * HOL/Relation INCOMPATIBILITY: renamed the relational operator r^-1
  4088 to 'converse' from 'inverse' (for compatibility with ZF and some
  4089 literature);
  4090 
  4091 * HOL/recdef can now declare non-recursive functions, with {} supplied as
  4092 the well-founded relation;
  4093 
  4094 * HOL/Set INCOMPATIBILITY: the complement of set A is now written -A instead of
  4095     Compl A.  The "Compl" syntax remains available as input syntax for this
  4096     release ONLY.
  4097 
  4098 * HOL/Update: new theory of function updates:
  4099     f(a:=b) == %x. if x=a then b else f x
  4100 may also be iterated as in f(a:=b,c:=d,...);
  4101 
  4102 * HOL/Vimage: new theory for inverse image of a function, syntax f-``B;
  4103 
  4104 * HOL/List:
  4105   - new function list_update written xs[i:=v] that updates the i-th
  4106     list position. May also be iterated as in xs[i:=a,j:=b,...].
  4107   - new function `upt' written [i..j(] which generates the list
  4108     [i,i+1,...,j-1], i.e. the upper bound is excluded. To include the upper
  4109     bound write [i..j], which is a shorthand for [i..j+1(].
  4110   - new lexicographic orderings and corresponding wellfoundedness theorems.
  4111 
  4112 * HOL/Arith:
  4113   - removed 'pred' (predecessor) function;
  4114   - generalized some theorems about n-1;
  4115   - many new laws about "div" and "mod";
  4116   - new laws about greatest common divisors (see theory ex/Primes);
  4117 
  4118 * HOL/Relation: renamed the relational operator r^-1 "converse"
  4119 instead of "inverse";
  4120 
  4121 * HOL/Induct/Multiset: a theory of multisets, including the wellfoundedness
  4122   of the multiset ordering;
  4123 
  4124 * directory HOL/Real: a construction of the reals using Dedekind cuts
  4125   (not included by default);
  4126 
  4127 * directory HOL/UNITY: Chandy and Misra's UNITY formalism;
  4128 
  4129 * directory HOL/Hoare: a new version of Hoare logic which permits many-sorted
  4130   programs, i.e. different program variables may have different types.
  4131 
  4132 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4133   [previously, this check worked only if the rewrite rule was unconditional]
  4134   Now rew can involve either definitions or equalities (either == or =).
  4135 
  4136 
  4137 *** ZF ***
  4138 
  4139 * theory Main includes everything; INCOMPATIBILITY: theory ZF.thy contains
  4140   only the theorems proved on ZF.ML;
  4141 
  4142 * ZF INCOMPATIBILITY: rule `equals0D' is now a well-formed destruct rule;
  4143   It and 'sym RS equals0D' are now in the default  claset, giving automatic
  4144   disjointness reasoning but breaking a few old proofs.
  4145 
  4146 * ZF/Update: new theory of function updates
  4147     with default rewrite rule  f(x:=y) ` z = if(z=x, y, f`z)
  4148   may also be iterated as in f(a:=b,c:=d,...);
  4149 
  4150 * in  let x=t in u(x), neither t nor u(x) has to be an FOL term.
  4151 
  4152 * calling (stac rew i) now fails if "rew" has no effect on the goal
  4153   [previously, this check worked only if the rewrite rule was unconditional]
  4154   Now rew can involve either definitions or equalities (either == or =).
  4155 
  4156 * case_tac provided for compatibility with HOL
  4157     (like the old excluded_middle_tac, but with subgoals swapped)
  4158 
  4159 
  4160 *** Internal programming interfaces ***
  4161 
  4162 * Pure: several new basic modules made available for general use, see
  4163 also src/Pure/README;
  4164 
  4165 * improved the theory data mechanism to support encapsulation (data
  4166 kind name replaced by private Object.kind, acting as authorization
  4167 key); new type-safe user interface via functor TheoryDataFun; generic
  4168 print_data function becomes basically useless;
  4169 
  4170 * removed global_names compatibility flag -- all theory declarations
  4171 are qualified by default;
  4172 
  4173 * module Pure/Syntax now offers quote / antiquote translation
  4174 functions (useful for Hoare logic etc. with implicit dependencies);
  4175 see HOL/ex/Antiquote for an example use;
  4176 
  4177 * Simplifier now offers conversions (asm_)(full_)rewrite: simpset ->
  4178 cterm -> thm;
  4179 
  4180 * new tactical CHANGED_GOAL for checking that a tactic modifies a
  4181 subgoal;
  4182 
  4183 * Display.print_goals function moved to Locale.print_goals;
  4184 
  4185 * standard print function for goals supports current_goals_markers
  4186 variable for marking begin of proof, end of proof, start of goal; the
  4187 default is ("", "", ""); setting current_goals_markers := ("<proof>",
  4188 "</proof>", "<goal>") causes SGML like tagged proof state printing,
  4189 for example;
  4190 
  4191 
  4192 
  4193 New in Isabelle98 (January 1998)
  4194 --------------------------------
  4195 
  4196 *** Overview of INCOMPATIBILITIES (see below for more details) ***
  4197 
  4198 * changed lexical syntax of terms / types: dots made part of long
  4199 identifiers, e.g. "%x.x" no longer possible, should be "%x. x";
  4200 
  4201 * simpset (and claset) reference variable replaced by functions
  4202 simpset / simpset_ref;
  4203 
  4204 * no longer supports theory aliases (via merge) and non-trivial
  4205 implicit merge of thms' signatures;
  4206 
  4207 * most internal names of constants changed due to qualified names;
  4208 
  4209 * changed Pure/Sequence interface (see Pure/seq.ML);
  4210 
  4211 
  4212 *** General Changes ***
  4213 
  4214 * hierachically structured name spaces (for consts, types, axms, thms
  4215 etc.); new lexical class 'longid' (e.g. Foo.bar.x) may render much of
  4216 old input syntactically incorrect (e.g. "%x.x"); COMPATIBILITY:
  4217 isatool fixdots ensures space after dots (e.g. "%x. x"); set
  4218 long_names for fully qualified output names; NOTE: ML programs
  4219 (special tactics, packages etc.) referring to internal names may have
  4220 to be adapted to cope with fully qualified names; in case of severe
  4221 backward campatibility problems try setting 'global_names' at compile
  4222 time to have enrything declared within a flat name space; one may also
  4223 fine tune name declarations in theories via the 'global' and 'local'
  4224 section;
  4225 
  4226 * reimplemented the implicit simpset and claset using the new anytype
  4227 data filed in signatures; references simpset:simpset ref etc. are
  4228 replaced by functions simpset:unit->simpset and
  4229 simpset_ref:unit->simpset ref; COMPATIBILITY: use isatool fixclasimp
  4230 to patch your ML files accordingly;
  4231 
  4232 * HTML output now includes theory graph data for display with Java
  4233 applet or isatool browser; data generated automatically via isatool
  4234 usedir (see -i option, ISABELLE_USEDIR_OPTIONS);
  4235 
  4236 * defs may now be conditional; improved rewrite_goals_tac to handle
  4237 conditional equations;
  4238 
  4239 * defs now admits additional type arguments, using TYPE('a) syntax;
  4240 
  4241 * theory aliases via merge (e.g. M=A+B+C) no longer supported, always
  4242 creates a new theory node; implicit merge of thms' signatures is
  4243 restricted to 'trivial' ones; COMPATIBILITY: one may have to use
  4244 transfer:theory->thm->thm in (rare) cases;
  4245 
  4246 * improved handling of draft signatures / theories; draft thms (and
  4247 ctyps, cterms) are automatically promoted to real ones;
  4248 
  4249 * slightly changed interfaces for oracles: admit many per theory, named
  4250 (e.g. oracle foo = mlfun), additional name argument for invoke_oracle;
  4251 
  4252 * print_goals: optional output of const types (set show_consts and
  4253 show_types);
  4254 
  4255 * improved output of warnings (###) and errors (***);
  4256 
  4257 * subgoal_tac displays a warning if the new subgoal has type variables;
  4258 
  4259 * removed old README and Makefiles;
  4260 
  4261 * replaced print_goals_ref hook by print_current_goals_fn and result_error_fn;
  4262 
  4263 * removed obsolete init_pps and init_database;
  4264 
  4265 * deleted the obsolete tactical STATE, which was declared by
  4266     fun STATE tacfun st = tacfun st st;
  4267 
  4268 * cd and use now support path variables, e.g. $ISABELLE_HOME, or ~
  4269 (which abbreviates $HOME);
  4270 
  4271 * changed Pure/Sequence interface (see Pure/seq.ML); COMPATIBILITY:
  4272 use isatool fixseq to adapt your ML programs (this works for fully
  4273 qualified references to the Sequence structure only!);
  4274 
  4275 * use_thy no longer requires writable current directory; it always
  4276 reloads .ML *and* .thy file, if either one is out of date;
  4277 
  4278 
  4279 *** Classical Reasoner ***
  4280 
  4281 * Clarify_tac, clarify_tac, clarify_step_tac, Clarify_step_tac: new
  4282 tactics that use classical reasoning to simplify a subgoal without
  4283 splitting it into several subgoals;
  4284 
  4285 * Safe_tac: like safe_tac but uses the default claset;
  4286 
  4287 
  4288 *** Simplifier ***
  4289 
  4290 * added simplification meta rules:
  4291     (asm_)(full_)simplify: simpset -> thm -> thm;
  4292 
  4293 * simplifier.ML no longer part of Pure -- has to be loaded by object
  4294 logics (again);
  4295 
  4296 * added prems argument to simplification procedures;
  4297 
  4298 * HOL, FOL, ZF: added infix function `addsplits':
  4299   instead of `<simpset> setloop (split_tac <thms>)'
  4300   you can simply write `<simpset> addsplits <thms>'
  4301 
  4302 
  4303 *** Syntax ***
  4304 
  4305 * TYPE('a) syntax for type reflection terms;
  4306 
  4307 * no longer handles consts with name "" -- declare as 'syntax' instead;
  4308 
  4309 * pretty printer: changed order of mixfix annotation preference (again!);
  4310 
  4311 * Pure: fixed idt/idts vs. pttrn/pttrns syntactic categories;
  4312 
  4313 
  4314 *** HOL ***
  4315 
  4316 * HOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4317   with `addloop' of the simplifier to faciliate case splitting in premises.
  4318 
  4319 * HOL/TLA: Stephan Merz's formalization of Lamport's Temporal Logic of Actions;
  4320 
  4321 * HOL/Auth: new protocol proofs including some for the Internet
  4322   protocol TLS;
  4323 
  4324 * HOL/Map: new theory of `maps' a la VDM;
  4325 
  4326 * HOL/simplifier: simplification procedures nat_cancel_sums for
  4327 cancelling out common nat summands from =, <, <= (in)equalities, or
  4328 differences; simplification procedures nat_cancel_factor for
  4329 cancelling common factor from =, <, <= (in)equalities over natural
  4330 sums; nat_cancel contains both kinds of procedures, it is installed by
  4331 default in Arith.thy -- this COULD MAKE EXISTING PROOFS FAIL;
  4332 
  4333 * HOL/simplifier: terms of the form
  4334   `? x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x)'  (or t=x)
  4335   are rewritten to
  4336   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t)',
  4337   and those of the form
  4338   `! x. P1(x) & ... & Pn(x) & x=t & Q1(x) & ... Qn(x) --> R(x)'  (or t=x)
  4339   are rewritten to
  4340   `P1(t) & ... & Pn(t) & Q1(t) & ... Qn(t) --> R(t)',
  4341 
  4342 * HOL/datatype
  4343   Each datatype `t' now comes with a theorem `split_t_case' of the form
  4344 
  4345   P(t_case f1 ... fn x) =
  4346      ( (!y1 ... ym1. x = C1 y1 ... ym1 --> P(f1 y1 ... ym1)) &
  4347         ...
  4348        (!y1 ... ymn. x = Cn y1 ... ymn --> P(f1 y1 ... ymn))
  4349      )
  4350 
  4351   and a theorem `split_t_case_asm' of the form
  4352 
  4353   P(t_case f1 ... fn x) =
  4354     ~( (? y1 ... ym1. x = C1 y1 ... ym1 & ~P(f1 y1 ... ym1)) |
  4355         ...
  4356        (? y1 ... ymn. x = Cn y1 ... ymn & ~P(f1 y1 ... ymn))
  4357      )
  4358   which can be added to a simpset via `addsplits'. The existing theorems
  4359   expand_list_case and expand_option_case have been renamed to
  4360   split_list_case and split_option_case.
  4361 
  4362 * HOL/Arithmetic:
  4363   - `pred n' is automatically converted to `n-1'.
  4364     Users are strongly encouraged not to use `pred' any longer,
  4365     because it will disappear altogether at some point.
  4366   - Users are strongly encouraged to write "0 < n" rather than
  4367     "n ~= 0". Theorems and proof tools have been modified towards this
  4368     `standard'.
  4369 
  4370 * HOL/Lists:
  4371   the function "set_of_list" has been renamed "set" (and its theorems too);
  4372   the function "nth" now takes its arguments in the reverse order and
  4373   has acquired the infix notation "!" as in "xs!n".
  4374 
  4375 * HOL/Set: UNIV is now a constant and is no longer translated to Compl{};
  4376 
  4377 * HOL/Set: The operator (UN x.B x) now abbreviates (UN x:UNIV. B x) and its
  4378   specialist theorems (like UN1_I) are gone.  Similarly for (INT x.B x);
  4379 
  4380 * HOL/record: extensible records with schematic structural subtyping
  4381 (single inheritance); EXPERIMENTAL version demonstrating the encoding,
  4382 still lacks various theorems and concrete record syntax;
  4383 
  4384 
  4385 *** HOLCF ***
  4386 
  4387 * removed "axioms" and "generated by" sections;
  4388 
  4389 * replaced "ops" section by extended "consts" section, which is capable of
  4390   handling the continuous function space "->" directly;
  4391 
  4392 * domain package:
  4393   . proves theorems immediately and stores them in the theory,
  4394   . creates hierachical name space,
  4395   . now uses normal mixfix annotations (instead of cinfix...),
  4396   . minor changes to some names and values (for consistency),
  4397   . e.g. cases -> casedist, dists_eq -> dist_eqs, [take_lemma] -> take_lemmas,
  4398   . separator between mutual domain defs: changed "," to "and",
  4399   . improved handling of sort constraints;  now they have to
  4400     appear on the left-hand side of the equations only;
  4401 
  4402 * fixed LAM <x,y,zs>.b syntax;
  4403 
  4404 * added extended adm_tac to simplifier in HOLCF -- can now discharge
  4405 adm (%x. P (t x)), where P is chainfinite and t continuous;
  4406 
  4407 
  4408 *** FOL and ZF ***
  4409 
  4410 * FOL: there is a new splitter `split_asm_tac' that can be used e.g.
  4411   with `addloop' of the simplifier to faciliate case splitting in premises.
  4412 
  4413 * qed_spec_mp, qed_goal_spec_mp, qed_goalw_spec_mp are available, as
  4414 in HOL, they strip ALL and --> from proved theorems;
  4415 
  4416 
  4417 
  4418 New in Isabelle94-8 (May 1997)
  4419 ------------------------------
  4420 
  4421 *** General Changes ***
  4422 
  4423 * new utilities to build / run / maintain Isabelle etc. (in parts
  4424 still somewhat experimental); old Makefiles etc. still functional;
  4425 
  4426 * new 'Isabelle System Manual';
  4427 
  4428 * INSTALL text, together with ./configure and ./build scripts;
  4429 
  4430 * reimplemented type inference for greater efficiency, better error
  4431 messages and clean internal interface;
  4432 
  4433 * prlim command for dealing with lots of subgoals (an easier way of
  4434 setting goals_limit);
  4435 
  4436 
  4437 *** Syntax ***
  4438 
  4439 * supports alternative (named) syntax tables (parser and pretty
  4440 printer); internal interface is provided by add_modesyntax(_i);
  4441 
  4442 * Pure, FOL, ZF, HOL, HOLCF now support symbolic input and output; to
  4443 be used in conjunction with the Isabelle symbol font; uses the
  4444 "symbols" syntax table;
  4445 
  4446 * added token_translation interface (may translate name tokens in
  4447 arbitrary ways, dependent on their type (free, bound, tfree, ...) and
  4448 the current print_mode); IMPORTANT: user print translation functions
  4449 are responsible for marking newly introduced bounds
  4450 (Syntax.mark_boundT);
  4451 
  4452 * token translations for modes "xterm" and "xterm_color" that display
  4453 names in bold, underline etc. or colors (which requires a color
  4454 version of xterm);
  4455 
  4456 * infixes may now be declared with names independent of their syntax;
  4457 
  4458 * added typed_print_translation (like print_translation, but may
  4459 access type of constant);
  4460 
  4461 
  4462 *** Classical Reasoner ***
  4463 
  4464 Blast_tac: a new tactic!  It is often more powerful than fast_tac, but has
  4465 some limitations.  Blast_tac...
  4466   + ignores addss, addbefore, addafter; this restriction is intrinsic
  4467   + ignores elimination rules that don't have the correct format
  4468         (the conclusion MUST be a formula variable)
  4469   + ignores types, which can make HOL proofs fail
  4470   + rules must not require higher-order unification, e.g. apply_type in ZF
  4471     [message "Function Var's argument not a bound variable" relates to this]
  4472   + its proof strategy is more general but can actually be slower
  4473 
  4474 * substitution with equality assumptions no longer permutes other
  4475 assumptions;
  4476 
  4477 * minor changes in semantics of addafter (now called addaltern); renamed
  4478 setwrapper to setWrapper and compwrapper to compWrapper; added safe wrapper
  4479 (and access functions for it);
  4480 
  4481 * improved combination of classical reasoner and simplifier:
  4482   + functions for handling clasimpsets
  4483   + improvement of addss: now the simplifier is called _after_ the
  4484     safe steps.
  4485   + safe variant of addss called addSss: uses safe simplifications
  4486     _during_ the safe steps. It is more complete as it allows multiple
  4487     instantiations of unknowns (e.g. with slow_tac).
  4488 
  4489 *** Simplifier ***
  4490 
  4491 * added interface for simplification procedures (functions that
  4492 produce *proven* rewrite rules on the fly, depending on current
  4493 redex);
  4494 
  4495 * ordering on terms as parameter (used for ordered rewriting);
  4496 
  4497 * new functions delcongs, deleqcongs, and Delcongs. richer rep_ss;
  4498 
  4499 * the solver is now split into a safe and an unsafe part.
  4500 This should be invisible for the normal user, except that the
  4501 functions setsolver and addsolver have been renamed to setSolver and
  4502 addSolver; added safe_asm_full_simp_tac;
  4503 
  4504 
  4505 *** HOL ***
  4506 
  4507 * a generic induction tactic `induct_tac' which works for all datatypes and
  4508 also for type `nat';
  4509 
  4510 * a generic case distinction tactic `exhaust_tac' which works for all
  4511 datatypes and also for type `nat';
  4512 
  4513 * each datatype comes with a function `size';
  4514 
  4515 * patterns in case expressions allow tuple patterns as arguments to
  4516 constructors, for example `case x of [] => ... | (x,y,z)#ps => ...';
  4517 
  4518 * primrec now also works with type nat;
  4519 
  4520 * recdef: a new declaration form, allows general recursive functions to be
  4521 defined in theory files.  See HOL/ex/Fib, HOL/ex/Primes, HOL/Subst/Unify.
  4522 
  4523 * the constant for negation has been renamed from "not" to "Not" to
  4524 harmonize with FOL, ZF, LK, etc.;
  4525 
  4526 * HOL/ex/LFilter theory of a corecursive "filter" functional for
  4527 infinite lists;
  4528 
  4529 * HOL/Modelcheck demonstrates invocation of model checker oracle;
  4530 
  4531 * HOL/ex/Ring.thy declares cring_simp, which solves equational
  4532 problems in commutative rings, using axiomatic type classes for + and *;
  4533 
  4534 * more examples in HOL/MiniML and HOL/Auth;
  4535 
  4536 * more default rewrite rules for quantifiers, union/intersection;
  4537 
  4538 * a new constant `arbitrary == @x.False';
  4539 
  4540 * HOLCF/IOA replaces old HOL/IOA;
  4541 
  4542 * HOLCF changes: derived all rules and arities
  4543   + axiomatic type classes instead of classes
  4544   + typedef instead of faking type definitions
  4545   + eliminated the internal constants less_fun, less_cfun, UU_fun, UU_cfun etc.
  4546   + new axclasses cpo, chfin, flat with flat < chfin < pcpo < cpo < po
  4547   + eliminated the types void, one, tr
  4548   + use unit lift and bool lift (with translations) instead of one and tr
  4549   + eliminated blift from Lift3.thy (use Def instead of blift)
  4550   all eliminated rules are derived as theorems --> no visible changes ;
  4551 
  4552 
  4553 *** ZF ***
  4554 
  4555 * ZF now has Fast_tac, Simp_tac and Auto_tac.  Union_iff is a now a default
  4556 rewrite rule; this may affect some proofs.  eq_cs is gone but can be put back
  4557 as ZF_cs addSIs [equalityI];
  4558 
  4559 
  4560 
  4561 New in Isabelle94-7 (November 96)
  4562 ---------------------------------
  4563 
  4564 * allowing negative levels (as offsets) in prlev and choplev;
  4565 
  4566 * super-linear speedup for large simplifications;
  4567 
  4568 * FOL, ZF and HOL now use miniscoping: rewriting pushes
  4569 quantifications in as far as possible (COULD MAKE EXISTING PROOFS
  4570 FAIL); can suppress it using the command Delsimps (ex_simps @
  4571 all_simps); De Morgan laws are also now included, by default;
  4572 
  4573 * improved printing of ==>  :  ~:
  4574 
  4575 * new object-logic "Sequents" adds linear logic, while replacing LK
  4576 and Modal (thanks to Sara Kalvala);
  4577 
  4578 * HOL/Auth: correctness proofs for authentication protocols;
  4579 
  4580 * HOL: new auto_tac combines rewriting and classical reasoning (many
  4581 examples on HOL/Auth);
  4582 
  4583 * HOL: new command AddIffs for declaring theorems of the form P=Q to
  4584 the rewriter and classical reasoner simultaneously;
  4585 
  4586 * function uresult no longer returns theorems in "standard" format;
  4587 regain previous version by: val uresult = standard o uresult;
  4588 
  4589 
  4590 
  4591 New in Isabelle94-6
  4592 -------------------
  4593 
  4594 * oracles -- these establish an interface between Isabelle and trusted
  4595 external reasoners, which may deliver results as theorems;
  4596 
  4597 * proof objects (in particular record all uses of oracles);
  4598 
  4599 * Simp_tac, Fast_tac, etc. that refer to implicit simpset / claset;
  4600 
  4601 * "constdefs" section in theory files;
  4602 
  4603 * "primrec" section (HOL) no longer requires names;
  4604 
  4605 * internal type "tactic" now simply "thm -> thm Sequence.seq";
  4606 
  4607 
  4608 
  4609 New in Isabelle94-5
  4610 -------------------
  4611 
  4612 * reduced space requirements;
  4613 
  4614 * automatic HTML generation from theories;
  4615 
  4616 * theory files no longer require "..." (quotes) around most types;
  4617 
  4618 * new examples, including two proofs of the Church-Rosser theorem;
  4619 
  4620 * non-curried (1994) version of HOL is no longer distributed;
  4621 
  4622 
  4623 
  4624 New in Isabelle94-4
  4625 -------------------
  4626 
  4627 * greatly reduced space requirements;
  4628 
  4629 * theory files (.thy) no longer require \...\ escapes at line breaks;
  4630 
  4631 * searchable theorem database (see the section "Retrieving theorems" on
  4632 page 8 of the Reference Manual);
  4633 
  4634 * new examples, including Grabczewski's monumental case study of the
  4635 Axiom of Choice;
  4636 
  4637 * The previous version of HOL renamed to Old_HOL;
  4638 
  4639 * The new version of HOL (previously called CHOL) uses a curried syntax
  4640 for functions.  Application looks like f a b instead of f(a,b);
  4641 
  4642 * Mutually recursive inductive definitions finally work in HOL;
  4643 
  4644 * In ZF, pattern-matching on tuples is now available in all abstractions and
  4645 translates to the operator "split";
  4646 
  4647 
  4648 
  4649 New in Isabelle94-3
  4650 -------------------
  4651 
  4652 * new infix operator, addss, allowing the classical reasoner to
  4653 perform simplification at each step of its search.  Example:
  4654         fast_tac (cs addss ss)
  4655 
  4656 * a new logic, CHOL, the same as HOL, but with a curried syntax
  4657 for functions.  Application looks like f a b instead of f(a,b).  Also pairs
  4658 look like (a,b) instead of <a,b>;
  4659 
  4660 * PLEASE NOTE: CHOL will eventually replace HOL!
  4661 
  4662 * In CHOL, pattern-matching on tuples is now available in all abstractions.
  4663 It translates to the operator "split".  A new theory of integers is available;
  4664 
  4665 * In ZF, integer numerals now denote two's-complement binary integers.
  4666 Arithmetic operations can be performed by rewriting.  See ZF/ex/Bin.ML;
  4667 
  4668 * Many new examples: I/O automata, Church-Rosser theorem, equivalents
  4669 of the Axiom of Choice;
  4670 
  4671 
  4672 
  4673 New in Isabelle94-2
  4674 -------------------
  4675 
  4676 * Significantly faster resolution;
  4677 
  4678 * the different sections in a .thy file can now be mixed and repeated
  4679 freely;
  4680 
  4681 * Database of theorems for FOL, HOL and ZF.  New
  4682 commands including qed, qed_goal and bind_thm store theorems in the database.
  4683 
  4684 * Simple database queries: return a named theorem (get_thm) or all theorems of
  4685 a given theory (thms_of), or find out what theory a theorem was proved in
  4686 (theory_of_thm);
  4687 
  4688 * Bugs fixed in the inductive definition and datatype packages;
  4689 
  4690 * The classical reasoner provides deepen_tac and depth_tac, making FOL_dup_cs
  4691 and HOL_dup_cs obsolete;
  4692 
  4693 * Syntactic ambiguities caused by the new treatment of syntax in Isabelle94-1
  4694 have been removed;
  4695 
  4696 * Simpler definition of function space in ZF;
  4697 
  4698 * new results about cardinal and ordinal arithmetic in ZF;
  4699 
  4700 * 'subtype' facility in HOL for introducing new types as subsets of existing
  4701 types;
  4702 
  4703 
  4704 $Id$