NEWS
author ballarin
Thu Aug 03 14:57:26 2006 +0200 (2006-08-03)
changeset 20318 0e0ea63fe768
parent 20217 25b068a99d2b
child 20327 69a9d839dcc8
permissions -rw-r--r--
Restructured algebra library, added ideals and quotient rings.
wenzelm@5363
     1
Isabelle NEWS -- history user-relevant changes
wenzelm@5363
     2
==============================================
wenzelm@2553
     3
wenzelm@17754
     4
New in this Isabelle release
wenzelm@17754
     5
----------------------------
wenzelm@17754
     6
wenzelm@17754
     7
*** General ***
wenzelm@17754
     8
wenzelm@17918
     9
* Theory syntax: the header format ``theory A = B + C:'' has been
wenzelm@17918
    10
discontinued in favour of ``theory A imports B C begin''.  Use isatool
wenzelm@17918
    11
fixheaders to convert existing theory files.  INCOMPATIBILITY.
wenzelm@17918
    12
wenzelm@17918
    13
* Theory syntax: the old non-Isar theory file format has been
wenzelm@17918
    14
discontinued altogether.  Note that ML proof scripts may still be used
wenzelm@17918
    15
with Isar theories; migration is usually quite simple with the ML
wenzelm@17918
    16
function use_legacy_bindings.  INCOMPATIBILITY.
wenzelm@17918
    17
wenzelm@19814
    18
* Theory syntax: some popular names (e.g. "class", "if") are now
wenzelm@19814
    19
keywords.  INCOMPATIBILITY, use double quotes.
wenzelm@19814
    20
wenzelm@17981
    21
* Legacy goal package: reduced interface to the bare minimum required
wenzelm@17981
    22
to keep existing proof scripts running.  Most other user-level
wenzelm@17981
    23
functions are now part of the OldGoals structure, which is *not* open
wenzelm@17981
    24
by default (consider isatool expandshort before open OldGoals).
wenzelm@17981
    25
Removed top_sg, prin, printyp, pprint_term/typ altogether, because
wenzelm@17981
    26
these tend to cause confusion about the actual goal (!) context being
wenzelm@17981
    27
used here, which is not necessarily the same as the_context().
wenzelm@17918
    28
wenzelm@17754
    29
* Command 'find_theorems': support "*" wildcard in "name:" criterion.
wenzelm@17754
    30
wenzelm@17754
    31
wenzelm@17865
    32
*** Document preparation ***
wenzelm@17865
    33
wenzelm@17869
    34
* Added antiquotations @{ML_type text} and @{ML_struct text} which
wenzelm@17869
    35
check the given source text as ML type/structure, printing verbatim.
wenzelm@17865
    36
wenzelm@17865
    37
wenzelm@17779
    38
*** Pure ***
wenzelm@17779
    39
haftmann@20188
    40
* class_package.ML offers a combination of axclasses and locales to achieve
haftmann@20188
    41
Haskell-like type classes in Isabelle. See HOL/ex/Classpackage.thy for examples.
haftmann@20188
    42
haftmann@20188
    43
* Yet another code generator framework allows to generate executable code
haftmann@20188
    44
for ML and Haskell (including "class"es). Most basic use cases:
haftmann@20188
    45
haftmann@20188
    46
    internal compilation:
haftmann@20188
    47
        code_serialize ml <list of constants (term syntax)> (-)
haftmann@20188
    48
    writing ML code to a file:
haftmann@20188
    49
        code_serialize ml <list of constants (term syntax)> (<filename>)
haftmann@20188
    50
    writing Haskell code to a bunch of files:
haftmann@20188
    51
        code_serialize haskell <list of constants (term syntax)> (<filename>)
haftmann@20188
    52
haftmann@20188
    53
See HOL/ex/Codegenerator.thy for examples.
haftmann@20188
    54
wenzelm@19254
    55
* Command 'no_translations' removes translation rules from theory
wenzelm@19254
    56
syntax.
wenzelm@19254
    57
wenzelm@19625
    58
* Overloaded definitions are now actually checked for acyclic
wenzelm@19714
    59
dependencies.  The overloading scheme is slightly more general than
wenzelm@19714
    60
that of Haskell98, although Isabelle does not demand an exact
wenzelm@19714
    61
correspondence to type class and instance declarations.
wenzelm@19714
    62
INCOMPATIBILITY, use ``defs (unchecked overloaded)'' to admit more
wenzelm@19714
    63
exotic versions of overloading -- at the discretion of the user!
wenzelm@19711
    64
wenzelm@19711
    65
Polymorphic constants are represented via type arguments, i.e. the
wenzelm@19711
    66
instantiation that matches an instance against the most general
wenzelm@19711
    67
declaration given in the signature.  For example, with the declaration
wenzelm@19711
    68
c :: 'a => 'a => 'a, an instance c :: nat => nat => nat is represented
wenzelm@19711
    69
as c(nat).  Overloading is essentially simultaneous structural
wenzelm@19711
    70
recursion over such type arguments.  Incomplete specification patterns
wenzelm@19714
    71
impose global constraints on all occurrences, e.g. c('a * 'a) on the
wenzelm@19715
    72
LHS means that more general c('a * 'b) will be disallowed on any RHS.
wenzelm@19714
    73
Command 'print_theory' outputs the normalized system of recursive
wenzelm@19714
    74
equations, see section "definitions".
wenzelm@19625
    75
wenzelm@17865
    76
* Isar: improper proof element 'guess' is like 'obtain', but derives
wenzelm@17865
    77
the obtained context from the course of reasoning!  For example:
wenzelm@17865
    78
wenzelm@17865
    79
  assume "EX x y. A x & B y"   -- "any previous fact"
wenzelm@17865
    80
  then guess x and y by clarify
wenzelm@17865
    81
wenzelm@17865
    82
This technique is potentially adventurous, depending on the facts and
wenzelm@17865
    83
proof tools being involved here.
wenzelm@17865
    84
wenzelm@18020
    85
* Isar: known facts from the proof context may be specified as literal
wenzelm@18020
    86
propositions, using ASCII back-quote syntax.  This works wherever
wenzelm@18020
    87
named facts used to be allowed so far, in proof commands, proof
wenzelm@18020
    88
methods, attributes etc.  Literal facts are retrieved from the context
wenzelm@18020
    89
according to unification of type and term parameters.  For example,
wenzelm@18020
    90
provided that "A" and "A ==> B" and "!!x. P x ==> Q x" are known
wenzelm@18020
    91
theorems in the current context, then these are valid literal facts:
wenzelm@18020
    92
`A` and `A ==> B` and `!!x. P x ==> Q x" as well as `P a ==> Q a` etc.
wenzelm@18020
    93
wenzelm@18020
    94
There is also a proof method "fact" which does the same composition
wenzelm@18044
    95
for explicit goal states, e.g. the following proof texts coincide with
wenzelm@18044
    96
certain special cases of literal facts:
wenzelm@18020
    97
wenzelm@18020
    98
  have "A" by fact                 ==  note `A`
wenzelm@18020
    99
  have "A ==> B" by fact           ==  note `A ==> B`
wenzelm@18020
   100
  have "!!x. P x ==> Q x" by fact  ==  note `!!x. P x ==> Q x`
wenzelm@18020
   101
  have "P a ==> Q a" by fact       ==  note `P a ==> Q a`
wenzelm@18020
   102
wenzelm@20118
   103
* Isar: ":" (colon) is no longer a symbolic identifier character in
wenzelm@20118
   104
outer syntax.  Thus symbolic identifiers may be used without
wenzelm@20118
   105
additional white space in declarations like this: ``assume *: A''.
wenzelm@20118
   106
wenzelm@20013
   107
* Isar: 'print_facts' prints all local facts of the current context,
wenzelm@20013
   108
both named and unnamed ones.
wenzelm@20013
   109
wenzelm@18308
   110
* Isar: 'def' now admits simultaneous definitions, e.g.:
wenzelm@18308
   111
wenzelm@18308
   112
  def x == "t" and y == "u"
wenzelm@18308
   113
wenzelm@18540
   114
* Isar: added command 'unfolding', which is structurally similar to
wenzelm@18540
   115
'using', but affects both the goal state and facts by unfolding given
wenzelm@18815
   116
rewrite rules.  Thus many occurrences of the 'unfold' method or
wenzelm@18540
   117
'unfolded' attribute may be replaced by first-class proof text.
wenzelm@18540
   118
wenzelm@18815
   119
* Isar: methods 'unfold' / 'fold', attributes 'unfolded' / 'folded',
wenzelm@18815
   120
and command 'unfolding' now all support object-level equalities
wenzelm@18815
   121
(potentially conditional).  The underlying notion of rewrite rule is
wenzelm@18815
   122
analogous to the 'rule_format' attribute, but *not* that of the
wenzelm@18815
   123
Simplifier (which is usually more generous).
wenzelm@18815
   124
wenzelm@19220
   125
* Isar: the goal restriction operator [N] (default N = 1) evaluates a
wenzelm@19220
   126
method expression within a sandbox consisting of the first N
wenzelm@19240
   127
sub-goals, which need to exist.  For example, ``simp_all [3]''
wenzelm@19240
   128
simplifies the first three sub-goals, while (rule foo, simp_all)[]
wenzelm@19240
   129
simplifies all new goals that emerge from applying rule foo to the
wenzelm@19240
   130
originally first one.
wenzelm@19220
   131
wenzelm@19814
   132
* Isar: schematic goals are no longer restricted to higher-order
wenzelm@19814
   133
patterns; e.g. ``lemma "?P(?x)" by (rule TrueI)'' now works as
wenzelm@19814
   134
expected.
wenzelm@19814
   135
wenzelm@18901
   136
* Isar: the conclusion of a long theorem statement is now either
wenzelm@18901
   137
'shows' (a simultaneous conjunction, as before), or 'obtains'
wenzelm@18901
   138
(essentially a disjunction of cases with local parameters and
wenzelm@18901
   139
assumptions).  The latter allows to express general elimination rules
wenzelm@18910
   140
adequately; in this notation common elimination rules look like this:
wenzelm@18901
   141
wenzelm@18901
   142
  lemma exE:    -- "EX x. P x ==> (!!x. P x ==> thesis) ==> thesis"
wenzelm@18901
   143
    assumes "EX x. P x"
wenzelm@18901
   144
    obtains x where "P x"
wenzelm@18901
   145
wenzelm@18901
   146
  lemma conjE:  -- "A & B ==> (A ==> B ==> thesis) ==> thesis"
wenzelm@18901
   147
    assumes "A & B"
wenzelm@18901
   148
    obtains A and B
wenzelm@18901
   149
wenzelm@18901
   150
  lemma disjE:  -- "A | B ==> (A ==> thesis) ==> (B ==> thesis) ==> thesis"
wenzelm@18901
   151
    assumes "A | B"
wenzelm@18901
   152
    obtains
wenzelm@18901
   153
      A
wenzelm@18901
   154
    | B
wenzelm@18901
   155
wenzelm@18910
   156
The subsequent classical rules even refer to the formal "thesis"
wenzelm@18901
   157
explicitly:
wenzelm@18901
   158
wenzelm@18901
   159
  lemma classical:     -- "(~ thesis ==> thesis) ==> thesis"
wenzelm@18901
   160
    obtains "~ thesis"
wenzelm@18901
   161
wenzelm@18910
   162
  lemma Peirce's_Law:  -- "((thesis ==> something) ==> thesis) ==> thesis"
wenzelm@18910
   163
    obtains "thesis ==> something"
wenzelm@18901
   164
wenzelm@18901
   165
The actual proof of an 'obtains' statement is analogous to that of the
wenzelm@18910
   166
Isar proof element 'obtain', only that there may be several cases.
wenzelm@18910
   167
Optional case names may be specified in parentheses; these will be
wenzelm@18910
   168
available both in the present proof and as annotations in the
wenzelm@18910
   169
resulting rule, for later use with the 'cases' method (cf. attribute
wenzelm@18910
   170
case_names).
wenzelm@18901
   171
wenzelm@19263
   172
* Isar: 'print_statement' prints theorems from the current theory or
wenzelm@19263
   173
proof context in long statement form, according to the syntax of a
wenzelm@19263
   174
top-level lemma.
wenzelm@19263
   175
wenzelm@18901
   176
* Isar: 'obtain' takes an optional case name for the local context
wenzelm@18901
   177
introduction rule (default "that").
wenzelm@18901
   178
wenzelm@19587
   179
* Isar: removed obsolete 'concl is' patterns.  INCOMPATIBILITY, use
wenzelm@19587
   180
explicit (is "_ ==> ?foo") in the rare cases where this still happens
wenzelm@19587
   181
to occur.
wenzelm@19587
   182
wenzelm@19682
   183
* Pure: syntax "CONST name" produces a fully internalized constant
wenzelm@19682
   184
according to the current context.  This is particularly useful for
wenzelm@19682
   185
syntax translations that should refer to internal constant
wenzelm@19682
   186
representations independently of name spaces.
wenzelm@19682
   187
wenzelm@19682
   188
* Isar/locales: 'const_syntax' provides a robust interface to the
wenzelm@19682
   189
'syntax' primitive that also works in a locale context.  Type
wenzelm@19682
   190
declaration and internal syntactic representation of given constants
wenzelm@19682
   191
retrieved from the context.
wenzelm@19682
   192
wenzelm@19665
   193
* Isar/locales: new derived specification elements 'axiomatization',
wenzelm@19665
   194
'definition', 'abbreviation', which support type-inference, admit
wenzelm@19083
   195
object-level specifications (equality, equivalence).  See also the
wenzelm@19083
   196
isar-ref manual.  Examples:
wenzelm@19081
   197
wenzelm@19665
   198
  axiomatization
wenzelm@19665
   199
    eq  (infix "===" 50)
wenzelm@19665
   200
    where eq_refl: "x === x" and eq_subst: "x === y ==> P x ==> P y"
wenzelm@19665
   201
wenzelm@19081
   202
  definition
wenzelm@19081
   203
    "f x y = x + y + 1"
wenzelm@19081
   204
    "g x = f x x"
wenzelm@19081
   205
wenzelm@19363
   206
  abbreviation
wenzelm@19081
   207
    neq  (infix "=!=" 50)
wenzelm@19363
   208
    "x =!= y == ~ (x === y)"
wenzelm@19081
   209
wenzelm@19083
   210
These specifications may be also used in a locale context.  Then the
wenzelm@19083
   211
constants being introduced depend on certain fixed parameters, and the
wenzelm@19083
   212
constant name is qualified by the locale base name.  An internal
wenzelm@19083
   213
abbreviation takes care for convenient input and output, making the
wenzelm@19088
   214
parameters implicit and using the original short name.  See also
wenzelm@19083
   215
HOL/ex/Abstract_NAT.thy for an example of deriving polymorphic
wenzelm@19083
   216
entities from a monomorphic theory.
wenzelm@19083
   217
wenzelm@19083
   218
Presently, abbreviations are only available 'in' a target locale, but
wenzelm@19363
   219
not inherited by general import expressions.  Also note that
wenzelm@19363
   220
'abbreviation' may be used as a type-safe replacement for 'syntax' +
wenzelm@19363
   221
'translations' in common applications.
wenzelm@19084
   222
wenzelm@19682
   223
Concrete syntax is attached to specified constants in internal form,
wenzelm@19682
   224
independently of name spaces.  The parse tree representation is
wenzelm@19682
   225
slightly different -- use 'const_syntax' instead of raw 'syntax', and
wenzelm@19682
   226
'translations' with explicit "CONST" markup to accommodate this.
wenzelm@19665
   227
ballarin@19783
   228
* Isar/locales: improved parameter handling:
ballarin@19783
   229
- use of locales "var" and "struct" no longer necessary;
ballarin@19783
   230
- parameter renamings are no longer required to be injective.
ballarin@19783
   231
  This enables, for example, to define a locale for endomorphisms thus:
ballarin@19783
   232
  locale endom = homom mult mult h.
ballarin@19783
   233
ballarin@19931
   234
* Isar/locales: changed the way locales with predicates are defined.
ballarin@19931
   235
Instead of accumulating the specification, the imported expression is
ballarin@19931
   236
now an interpretation.
ballarin@19931
   237
INCOMPATIBILITY: different normal form of locale expressions.
ballarin@19931
   238
In particular, in interpretations of locales with predicates,
ballarin@19931
   239
goals repesenting already interpreted fragments are not removed
ballarin@19984
   240
automatically.  Use methods `intro_locales' and `unfold_locales'; see below.
ballarin@19984
   241
ballarin@19984
   242
* Isar/locales: new methods `intro_locales' and `unfold_locales' provide
ballarin@19984
   243
backward reasoning on locales predicates.  The methods are aware of
ballarin@19984
   244
interpretations and discharge corresponding goals.  `intro_locales' is
ballarin@19984
   245
less aggressive then `unfold_locales' and does not unfold predicates to
ballarin@19984
   246
assumptions.
ballarin@19931
   247
ballarin@19931
   248
* Isar/locales: the order in which locale fragments are accumulated
ballarin@19931
   249
has changed.  This enables to override declarations from fragments
ballarin@19931
   250
due to interpretations -- for example, unwanted simp rules.
ballarin@19931
   251
wenzelm@18233
   252
* Provers/induct: improved internal context management to support
wenzelm@18233
   253
local fixes and defines on-the-fly.  Thus explicit meta-level
wenzelm@18233
   254
connectives !! and ==> are rarely required anymore in inductive goals
wenzelm@18233
   255
(using object-logic connectives for this purpose has been long
wenzelm@18233
   256
obsolete anyway).  The subsequent proof patterns illustrate advanced
wenzelm@18233
   257
techniques of natural induction; general datatypes and inductive sets
wenzelm@18267
   258
work analogously (see also src/HOL/Lambda for realistic examples).
wenzelm@18267
   259
wenzelm@18267
   260
(1) This is how to ``strengthen'' an inductive goal wrt. certain
wenzelm@18239
   261
parameters:
wenzelm@18233
   262
wenzelm@18233
   263
  lemma
wenzelm@18233
   264
    fixes n :: nat and x :: 'a
wenzelm@18233
   265
    assumes a: "A n x"
wenzelm@18233
   266
    shows "P n x"
wenzelm@18233
   267
    using a                     -- {* make induct insert fact a *}
wenzelm@18233
   268
  proof (induct n fixing: x)    -- {* generalize goal to "!!x. A n x ==> P n x" *}
wenzelm@18248
   269
    case 0
wenzelm@18233
   270
    show ?case sorry
wenzelm@18233
   271
  next
wenzelm@18248
   272
    case (Suc n)
wenzelm@18239
   273
    note `!!x. A n x ==> P n x` -- {* induction hypothesis, according to induction rule *}
wenzelm@18239
   274
    note `A (Suc n) x`          -- {* induction premise, stemming from fact a *}
wenzelm@18233
   275
    show ?case sorry
wenzelm@18233
   276
  qed
wenzelm@18233
   277
wenzelm@18267
   278
(2) This is how to perform induction over ``expressions of a certain
wenzelm@18233
   279
form'', using a locally defined inductive parameter n == "a x"
wenzelm@18239
   280
together with strengthening (the latter is usually required to get
wenzelm@18267
   281
sufficiently flexible induction hypotheses):
wenzelm@18233
   282
wenzelm@18233
   283
  lemma
wenzelm@18233
   284
    fixes a :: "'a => nat"
wenzelm@18233
   285
    assumes a: "A (a x)"
wenzelm@18233
   286
    shows "P (a x)"
wenzelm@18233
   287
    using a
wenzelm@18233
   288
  proof (induct n == "a x" fixing: x)
wenzelm@18233
   289
    ...
wenzelm@18233
   290
wenzelm@18267
   291
See also HOL/Isar_examples/Puzzle.thy for an application of the this
wenzelm@18267
   292
particular technique.
wenzelm@18267
   293
wenzelm@18901
   294
(3) This is how to perform existential reasoning ('obtains' or
wenzelm@18901
   295
'obtain') by induction, while avoiding explicit object-logic
wenzelm@18901
   296
encodings:
wenzelm@18901
   297
wenzelm@18901
   298
  lemma
wenzelm@18901
   299
    fixes n :: nat
wenzelm@18901
   300
    obtains x :: 'a where "P n x" and "Q n x"
wenzelm@18267
   301
  proof (induct n fixing: thesis)
wenzelm@18267
   302
    case 0
wenzelm@18267
   303
    obtain x where "P 0 x" and "Q 0 x" sorry
wenzelm@18399
   304
    then show thesis by (rule 0)
wenzelm@18267
   305
  next
wenzelm@18267
   306
    case (Suc n)
wenzelm@18267
   307
    obtain x where "P n x" and "Q n x" by (rule Suc.hyps)
wenzelm@18267
   308
    obtain x where "P (Suc n) x" and "Q (Suc n) x" sorry
wenzelm@18267
   309
    then show thesis by (rule Suc.prems)
wenzelm@18267
   310
  qed
wenzelm@18267
   311
wenzelm@18267
   312
Here the 'fixing: thesis' specification essentially modifies the scope
wenzelm@18267
   313
of the formal thesis parameter, in order to the get the whole
wenzelm@18267
   314
existence statement through the induction as expected.
wenzelm@18233
   315
wenzelm@18506
   316
* Provers/induct: mutual induction rules are now specified as a list
wenzelm@18506
   317
of rule sharing the same induction cases.  HOL packages usually
wenzelm@18506
   318
provide foo_bar.inducts for mutually defined items foo and bar
wenzelm@18506
   319
(e.g. inductive sets or datatypes).  INCOMPATIBILITY, users need to
wenzelm@18506
   320
specify mutual induction rules differently, i.e. like this:
wenzelm@18506
   321
wenzelm@18506
   322
  (induct rule: foo_bar.inducts)
wenzelm@18506
   323
  (induct set: foo bar)
wenzelm@18506
   324
  (induct type: foo bar)
wenzelm@18506
   325
wenzelm@18506
   326
The ML function ProjectRule.projections turns old-style rules into the
wenzelm@18506
   327
new format.
wenzelm@18506
   328
wenzelm@18506
   329
* Provers/induct: improved handling of simultaneous goals.  Instead of
wenzelm@18506
   330
introducing object-level conjunction, the statement is now split into
wenzelm@18506
   331
several conclusions, while the corresponding symbolic cases are
wenzelm@18601
   332
nested accordingly.  INCOMPATIBILITY, proofs need to be structured
wenzelm@18601
   333
explicitly.  For example:
wenzelm@18480
   334
wenzelm@18480
   335
  lemma
wenzelm@18480
   336
    fixes n :: nat
wenzelm@18480
   337
    shows "P n" and "Q n"
wenzelm@18480
   338
  proof (induct n)
wenzelm@18601
   339
    case 0 case 1
wenzelm@18480
   340
    show "P 0" sorry
wenzelm@18480
   341
  next
wenzelm@18601
   342
    case 0 case 2
wenzelm@18480
   343
    show "Q 0" sorry
wenzelm@18480
   344
  next
wenzelm@18601
   345
    case (Suc n) case 1
wenzelm@18480
   346
    note `P n` and `Q n`
wenzelm@18480
   347
    show "P (Suc n)" sorry
wenzelm@18480
   348
  next
wenzelm@18601
   349
    case (Suc n) case 2
wenzelm@18480
   350
    note `P n` and `Q n`
wenzelm@18480
   351
    show "Q (Suc n)" sorry
wenzelm@18480
   352
  qed
wenzelm@18480
   353
wenzelm@18601
   354
The split into subcases may be deferred as follows -- this is
wenzelm@18601
   355
particularly relevant for goal statements with local premises.
wenzelm@18601
   356
wenzelm@18601
   357
  lemma
wenzelm@18601
   358
    fixes n :: nat
wenzelm@18601
   359
    shows "A n ==> P n" and "B n ==> Q n"
wenzelm@18601
   360
  proof (induct n)
wenzelm@18601
   361
    case 0
wenzelm@18601
   362
    {
wenzelm@18601
   363
      case 1
wenzelm@18601
   364
      note `A 0`
wenzelm@18601
   365
      show "P 0" sorry
wenzelm@18601
   366
    next
wenzelm@18601
   367
      case 2
wenzelm@18601
   368
      note `B 0`
wenzelm@18601
   369
      show "Q 0" sorry
wenzelm@18601
   370
    }
wenzelm@18601
   371
  next
wenzelm@18601
   372
    case (Suc n)
wenzelm@18601
   373
    note `A n ==> P n` and `B n ==> Q n`
wenzelm@18601
   374
    {
wenzelm@18601
   375
      case 1
wenzelm@18601
   376
      note `A (Suc n)`
wenzelm@18601
   377
      show "P (Suc n)" sorry
wenzelm@18601
   378
    next
wenzelm@18601
   379
      case 2
wenzelm@18601
   380
      note `B (Suc n)`
wenzelm@18601
   381
      show "Q (Suc n)" sorry
wenzelm@18601
   382
    }
wenzelm@18601
   383
  qed
wenzelm@18601
   384
wenzelm@18506
   385
If simultaneous goals are to be used with mutual rules, the statement
wenzelm@18506
   386
needs to be structured carefully as a two-level conjunction, using
wenzelm@18506
   387
lists of propositions separated by 'and':
wenzelm@18506
   388
wenzelm@18507
   389
  lemma
wenzelm@18507
   390
    shows "a : A ==> P1 a"
wenzelm@18507
   391
          "a : A ==> P2 a"
wenzelm@18507
   392
      and "b : B ==> Q1 b"
wenzelm@18507
   393
          "b : B ==> Q2 b"
wenzelm@18507
   394
          "b : B ==> Q3 b"
wenzelm@18507
   395
  proof (induct set: A B)
wenzelm@18480
   396
wenzelm@18399
   397
* Provers/induct: support coinduction as well.  See
wenzelm@18399
   398
src/HOL/Library/Coinductive_List.thy for various examples.
wenzelm@18399
   399
nipkow@18674
   400
* Simplifier: by default the simplifier trace only shows top level rewrites
nipkow@18674
   401
now. That is, trace_simp_depth_limit is set to 1 by default. Thus there is
nipkow@18674
   402
less danger of being flooded by the trace. The trace indicates where parts
nipkow@18674
   403
have been suppressed.
nipkow@18674
   404
  
wenzelm@18536
   405
* Provers/classical: removed obsolete classical version of elim_format
wenzelm@18536
   406
attribute; classical elim/dest rules are now treated uniformly when
wenzelm@18536
   407
manipulating the claset.
wenzelm@18536
   408
wenzelm@18694
   409
* Provers/classical: stricter checks to ensure that supplied intro,
wenzelm@18694
   410
dest and elim rules are well-formed; dest and elim rules must have at
wenzelm@18694
   411
least one premise.
wenzelm@18694
   412
wenzelm@18694
   413
* Provers/classical: attributes dest/elim/intro take an optional
wenzelm@18695
   414
weight argument for the rule (just as the Pure versions).  Weights are
wenzelm@18696
   415
ignored by automated tools, but determine the search order of single
wenzelm@18694
   416
rule steps.
paulson@18557
   417
wenzelm@18536
   418
* Syntax: input syntax now supports dummy variable binding "%_. b",
wenzelm@18536
   419
where the body does not mention the bound variable.  Note that dummy
wenzelm@18536
   420
patterns implicitly depend on their context of bounds, which makes
wenzelm@18536
   421
"{_. _}" match any set comprehension as expected.  Potential
wenzelm@18536
   422
INCOMPATIBILITY -- parse translations need to cope with syntactic
wenzelm@18536
   423
constant "_idtdummy" in the binding position.
wenzelm@18536
   424
wenzelm@18536
   425
* Syntax: removed obsolete syntactic constant "_K" and its associated
wenzelm@18536
   426
parse translation.  INCOMPATIBILITY -- use dummy abstraction instead,
wenzelm@18536
   427
for example "A -> B" => "Pi A (%_. B)".
wenzelm@17779
   428
wenzelm@17865
   429
nipkow@17806
   430
*** HOL ***
nipkow@17806
   431
nipkow@19895
   432
* New top level command 'normal_form' computes the normal form of a term
nipkow@19895
   433
that may contain free variables. For example 'normal_form "rev[a,b,c]"'
nipkow@19895
   434
prints '[b,c,a]'. This command is suitable for heavy-duty computations
nipkow@19895
   435
because the functions are compiled to ML first.
nipkow@19895
   436
INCOMPATIBILITY: new keywords 'normal_form' must quoted when used as
nipkow@19895
   437
an identifier.
nipkow@19895
   438
wenzelm@17996
   439
* Alternative iff syntax "A <-> B" for equality on bool (with priority
wenzelm@17996
   440
25 like -->); output depends on the "iff" print_mode, the default is
wenzelm@17996
   441
"A = B" (with priority 50).
wenzelm@17996
   442
ballarin@19279
   443
* Renamed constants in HOL.thy and Orderings.thy:
haftmann@19233
   444
    op +   ~> HOL.plus
haftmann@19233
   445
    op -   ~> HOL.minus
haftmann@19233
   446
    uminus ~> HOL.uminus
haftmann@19233
   447
    op *   ~> HOL.times
haftmann@19277
   448
    op <   ~> Orderings.less
haftmann@19277
   449
    op <=  ~> Orderings.less_eq
haftmann@19233
   450
haftmann@19233
   451
Adaptions may be required in the following cases:
haftmann@19233
   452
nipkow@19377
   453
a) User-defined constants using any of the names "plus", "minus", "times",
nipkow@19377
   454
"less" or "less_eq". The standard syntax translations for "+", "-" and "*"
nipkow@19377
   455
may go wrong.
haftmann@19233
   456
INCOMPATIBILITY: use more specific names.
haftmann@19233
   457
haftmann@19277
   458
b) Variables named "plus", "minus", "times", "less", "less_eq"
haftmann@19233
   459
INCOMPATIBILITY: use more specific names.
haftmann@19233
   460
nipkow@19377
   461
c) Permutative equations (e.g. "a + b = b + a")
nipkow@19377
   462
Since the change of names also changes the order of terms, permutative
nipkow@19377
   463
rewrite rules may get applied in a different order. Experience shows that
nipkow@19377
   464
this is rarely the case (only two adaptions in the whole Isabelle
nipkow@19377
   465
distribution).
nipkow@19377
   466
INCOMPATIBILITY: rewrite proofs
haftmann@19233
   467
haftmann@19233
   468
d) ML code directly refering to constant names
haftmann@19233
   469
This in general only affects hand-written proof tactics, simprocs and so on.
haftmann@19233
   470
INCOMPATIBILITY: grep your sourcecode and replace names.
haftmann@19233
   471
nipkow@18674
   472
* "LEAST x:A. P" expands to "LEAST x. x:A & P" (input only).
nipkow@18674
   473
nipkow@19377
   474
* The old set interval syntax "{m..n(}" (and relatives) has been removed.
nipkow@19377
   475
  Use "{m..<n}" (and relatives) instead.
nipkow@19377
   476
wenzelm@17865
   477
* In the context of the assumption "~(s = t)" the Simplifier rewrites
wenzelm@17865
   478
"t = s" to False (by simproc "neq_simproc").  For backward
wenzelm@17865
   479
compatibility this can be disabled by ML "reset use_neq_simproc".
wenzelm@17779
   480
nipkow@18979
   481
* "m dvd n" where m and n are numbers is evaluated to True/False by simp.
nipkow@18979
   482
nipkow@19211
   483
* Theorem Cons_eq_map_conv no longer has attribute `simp'.
nipkow@19211
   484
ballarin@19279
   485
* Theorem setsum_mult renamed to setsum_right_distrib.
ballarin@19279
   486
nipkow@19211
   487
* Prefer ex1I over ex_ex1I in single-step reasoning, e.g. by the
nipkow@19211
   488
'rule' method.
nipkow@19211
   489
webertj@17809
   490
* Tactics 'sat' and 'satx' reimplemented, several improvements: goals
webertj@17809
   491
no longer need to be stated as "<prems> ==> False", equivalences (i.e.
wenzelm@17865
   492
"=" on type bool) are handled, variable names of the form "lit_<n>"
wenzelm@17865
   493
are no longer reserved, significant speedup.
wenzelm@17865
   494
webertj@20136
   495
* Tactics 'sat' and 'satx' can now replay MiniSat proof traces.  zChaff is
webertj@20136
   496
still supported as well.
webertj@20136
   497
wenzelm@18480
   498
* inductive and datatype: provide projections of mutual rules, bundled
wenzelm@18480
   499
as foo_bar.inducts;
wenzelm@18480
   500
wenzelm@19572
   501
* Library: theory Accessible_Part has been move to main HOL.
wenzelm@19572
   502
wenzelm@18446
   503
* Library: added theory Coinductive_List of potentially infinite lists
wenzelm@18446
   504
as greatest fixed-point.
wenzelm@18399
   505
wenzelm@19254
   506
* Library: added theory AssocList which implements (finite) maps as
schirmer@19252
   507
association lists.
webertj@17809
   508
berghofe@19855
   509
* New proof method "evaluation" for efficiently solving a goal
berghofe@19855
   510
  (i.e. a boolean expression) by compiling it to ML. The goal is
berghofe@19855
   511
  "proved" (via the oracle "Evaluation") if it evaluates to True.
berghofe@19855
   512
webertj@20217
   513
* Linear arithmetic now splits certain operators (e.g. min, max, abs) also
webertj@20217
   514
when invoked by the simplifier.  This results in the simplifier being more
webertj@20217
   515
powerful on arithmetic goals.
webertj@20217
   516
INCOMPATIBILTY: rewrite proofs.  Set fast_arith_split_limit to 0 to obtain
webertj@20217
   517
the old behavior.
webertj@20217
   518
kleing@20067
   519
* Support for hex (0x20) and binary (0b1001) numerals. 
wenzelm@19254
   520
ballarin@20169
   521
*** HOL-Algebra ***
ballarin@20169
   522
ballarin@20169
   523
* Method algebra is now set up via an attribute.  For examples see CRing.thy.
ballarin@20318
   524
  INCOMPATIBILITY: the method is now weaker on combinations of algebraic
ballarin@20318
   525
  structures.
ballarin@20318
   526
ballarin@20318
   527
* Formalisation of ideals and the quotient construction over rings, contributed
ballarin@20318
   528
  by Stephan Hohe.
ballarin@20318
   529
ballarin@20318
   530
* Renamed `CRing.thy' to `Ring.thy'.  INCOMPATIBILITY.
ballarin@20169
   531
wenzelm@19653
   532
*** HOL-Complex ***
wenzelm@19653
   533
wenzelm@19653
   534
* Theory Real: new method ferrack implements quantifier elimination
wenzelm@19653
   535
for linear arithmetic over the reals. The quantifier elimination
wenzelm@19653
   536
feature is used only for decision, for compatibility with arith. This
wenzelm@19653
   537
means a goal is either solved or left unchanged, no simplification.
wenzelm@19653
   538
wenzelm@19653
   539
wenzelm@17878
   540
*** ML ***
wenzelm@17878
   541
haftmann@18450
   542
* Pure/library:
haftmann@18450
   543
haftmann@18450
   544
  val burrow: ('a list -> 'b list) -> 'a list list -> 'b list list
haftmann@18549
   545
  val fold_burrow: ('a list -> 'c -> 'b list * 'd) -> 'a list list -> 'c -> 'b list list * 'd
haftmann@18450
   546
wenzelm@18540
   547
The semantics of "burrow" is: "take a function with *simulatanously*
wenzelm@18540
   548
transforms a list of value, and apply it *simulatanously* to a list of
wenzelm@18540
   549
list of values of the appropriate type". Confer this with "map" which
wenzelm@18540
   550
would *not* apply its argument function simulatanously but in
haftmann@18549
   551
sequence. "fold_burrow" has an additional context.
wenzelm@18540
   552
wenzelm@18540
   553
Both actually avoid the usage of "unflat" since they hide away
wenzelm@18540
   554
"unflat" from the user.
haftmann@18450
   555
wenzelm@18446
   556
* Pure/library: functions map2 and fold2 with curried syntax for
wenzelm@18446
   557
simultanous mapping and folding:
wenzelm@18446
   558
haftmann@18422
   559
    val map2: ('a -> 'b -> 'c) -> 'a list -> 'b list -> 'c list
haftmann@18422
   560
    val fold2: ('a -> 'b -> 'c -> 'c) -> 'a list -> 'b list -> 'c -> 'c
haftmann@18422
   561
wenzelm@18446
   562
* Pure/library: indexed lists - some functions in the Isabelle library
wenzelm@18446
   563
treating lists over 'a as finite mappings from [0...n] to 'a have been
wenzelm@18446
   564
given more convenient names and signatures reminiscent of similar
wenzelm@18446
   565
functions for alists, tables, etc:
haftmann@18051
   566
haftmann@18051
   567
  val nth: 'a list -> int -> 'a 
haftmann@18051
   568
  val nth_update: int * 'a -> 'a list -> 'a list
haftmann@18051
   569
  val nth_map: int -> ('a -> 'a) -> 'a list -> 'a list
haftmann@18051
   570
  val fold_index: (int * 'a -> 'b -> 'b) -> 'a list -> 'b -> 'b
haftmann@18051
   571
wenzelm@18446
   572
Note that fold_index starts counting at index 0, not 1 like foldln
wenzelm@18446
   573
used to.
wenzelm@18446
   574
wenzelm@19653
   575
* Pure/library: general ``divide_and_conquer'' combinator on lists.
wenzelm@19653
   576
wenzelm@19032
   577
* Pure/General/name_mangler.ML provides a functor for generic name
wenzelm@18446
   578
mangling (bijective mapping from any expression values to strings).
wenzelm@18446
   579
wenzelm@19032
   580
* Pure/General/rat.ML implements rational numbers.
wenzelm@19032
   581
wenzelm@19032
   582
* Pure/General/table.ML: the join operations now works via exceptions
wenzelm@19081
   583
DUP/SAME instead of type option.  This is simpler in simple cases, and
wenzelm@19081
   584
admits slightly more efficient complex applications.
wenzelm@18446
   585
wenzelm@18642
   586
* Pure: datatype Context.generic joins theory/Proof.context and
wenzelm@18644
   587
provides some facilities for code that works in either kind of
wenzelm@18642
   588
context, notably GenericDataFun for uniform theory and proof data.
wenzelm@18642
   589
wenzelm@18862
   590
* Pure: 'advanced' translation functions (parse_translation etc.) now
wenzelm@18862
   591
use Context.generic instead of just theory.
wenzelm@18862
   592
wenzelm@18737
   593
* Pure: simplified internal attribute type, which is now always
wenzelm@18737
   594
Context.generic * thm -> Context.generic * thm.  Global (theory)
wenzelm@18737
   595
vs. local (Proof.context) attributes have been discontinued, while
wenzelm@18738
   596
minimizing code duplication.  Thm.rule_attribute and
wenzelm@18738
   597
Thm.declaration_attribute build canonical attributes; see also
wenzelm@19006
   598
structure Context for further operations on Context.generic, notably
wenzelm@19006
   599
GenericDataFun.  INCOMPATIBILITY, need to adapt attribute type
wenzelm@19006
   600
declarations and definitions.
wenzelm@19006
   601
wenzelm@19508
   602
* Pure/kernel: consts certification ignores sort constraints given in
wenzelm@19508
   603
signature declarations.  (This information is not relevant to the
wenzelm@19508
   604
logic, but only for type inference.)  IMPORTANT INTERNAL CHANGE.
wenzelm@19508
   605
wenzelm@19508
   606
* Pure: axiomatic type classes are now purely definitional, with
wenzelm@19508
   607
explicit proofs of class axioms and super class relations performed
wenzelm@19508
   608
internally.  See Pure/axclass.ML for the main internal interfaces --
wenzelm@19508
   609
notably AxClass.define_class supercedes AxClass.add_axclass, and
wenzelm@19508
   610
AxClass.axiomatize_class/classrel/arity supercede
wenzelm@19508
   611
Sign.add_classes/classrel/arities.
wenzelm@19508
   612
wenzelm@19006
   613
* Pure/Isar: Args/Attrib parsers operate on Context.generic --
wenzelm@19006
   614
global/local versions on theory vs. Proof.context have been
wenzelm@19006
   615
discontinued; Attrib.syntax and Method.syntax have been adapted
wenzelm@19006
   616
accordingly.  INCOMPATIBILITY, need to adapt parser expressions for
wenzelm@19006
   617
attributes, methods, etc.
wenzelm@18642
   618
wenzelm@18446
   619
* Pure: several functions of signature "... -> theory -> theory * ..."
wenzelm@18446
   620
have been reoriented to "... -> theory -> ... * theory" in order to
wenzelm@18446
   621
allow natural usage in combination with the ||>, ||>>, |-> and
wenzelm@18446
   622
fold_map combinators.
haftmann@18051
   623
wenzelm@18020
   624
* Pure: primitive rule lift_rule now takes goal cterm instead of an
wenzelm@18145
   625
actual goal state (thm).  Use Thm.lift_rule (Thm.cprem_of st i) to
wenzelm@18020
   626
achieve the old behaviour.
wenzelm@18020
   627
wenzelm@18020
   628
* Pure: the "Goal" constant is now called "prop", supporting a
wenzelm@18020
   629
slightly more general idea of ``protecting'' meta-level rule
wenzelm@18020
   630
statements.
wenzelm@18020
   631
wenzelm@20040
   632
* Pure: Logic.(un)varify only works in a global context, which is now
wenzelm@20040
   633
enforced instead of silently assumed.  INCOMPATIBILITY, may use
wenzelm@20040
   634
Logic.legacy_(un)varify as temporary workaround.
wenzelm@20040
   635
wenzelm@20090
   636
* Pure: structure Name provides scalable operations for generating
wenzelm@20090
   637
internal variable names, notably Name.variants etc.  This replaces
wenzelm@20090
   638
some popular functions from term.ML:
wenzelm@20090
   639
wenzelm@20090
   640
  Term.variant		->  Name.variant
wenzelm@20090
   641
  Term.variantlist	->  Name.variant_list  (*canonical argument order*)
wenzelm@20090
   642
  Term.invent_names	->  Name.invent_list
wenzelm@20090
   643
wenzelm@20090
   644
Note that low-level renaming rarely occurs in new code -- operations
wenzelm@20090
   645
from structure Variable are used instead (see below).
wenzelm@20090
   646
wenzelm@20040
   647
* Pure: structure Variable provides fundamental operations for proper
wenzelm@20040
   648
treatment of fixed/schematic variables in a context.  For example,
wenzelm@20040
   649
Variable.import introduces fixes for schematics of given facts and
wenzelm@20040
   650
Variable.export reverses the effect (up to renaming) -- this replaces
wenzelm@20040
   651
various freeze_thaw operations.
wenzelm@20040
   652
wenzelm@18567
   653
* Pure: structure Goal provides simple interfaces for
wenzelm@17981
   654
init/conclude/finish and tactical prove operations (replacing former
wenzelm@20040
   655
Tactic.prove).  Goal.prove is the canonical way to prove results
wenzelm@20040
   656
within a given context; Goal.prove_global is a degraded version for
wenzelm@20040
   657
theory level goals, including a global Drule.standard.  Note that
wenzelm@20040
   658
OldGoals.prove_goalw_cterm has long been obsolete, since it is
wenzelm@20040
   659
ill-behaved in a local proof context (e.g. with local fixes/assumes or
wenzelm@20040
   660
in a locale context).
wenzelm@17981
   661
wenzelm@18815
   662
* Isar: simplified treatment of user-level errors, using exception
wenzelm@18687
   663
ERROR of string uniformly.  Function error now merely raises ERROR,
wenzelm@18686
   664
without any side effect on output channels.  The Isar toplevel takes
wenzelm@18686
   665
care of proper display of ERROR exceptions.  ML code may use plain
wenzelm@18686
   666
handle/can/try; cat_error may be used to concatenate errors like this:
wenzelm@18686
   667
wenzelm@18686
   668
  ... handle ERROR msg => cat_error msg "..."
wenzelm@18686
   669
wenzelm@18686
   670
Toplevel ML code (run directly or through the Isar toplevel) may be
wenzelm@18687
   671
embedded into the Isar toplevel with exception display/debug like
wenzelm@18687
   672
this:
wenzelm@18686
   673
wenzelm@18686
   674
  Isar.toplevel (fn () => ...)
wenzelm@18686
   675
wenzelm@18686
   676
INCOMPATIBILITY, removed special transform_error facilities, removed
wenzelm@18686
   677
obsolete variants of user-level exceptions (ERROR_MESSAGE,
wenzelm@18686
   678
Context.PROOF, ProofContext.CONTEXT, Proof.STATE, ProofHistory.FAIL)
wenzelm@18686
   679
-- use plain ERROR instead.
wenzelm@18686
   680
wenzelm@18815
   681
* Isar: theory setup now has type (theory -> theory), instead of a
wenzelm@18722
   682
list.  INCOMPATIBILITY, may use #> to compose setup functions.
wenzelm@18722
   683
wenzelm@18815
   684
* Isar: installed ML toplevel pretty printer for type Proof.context,
wenzelm@18815
   685
subject to ProofContext.debug/verbose flags.
wenzelm@18815
   686
wenzelm@18815
   687
* Isar: Toplevel.theory_to_proof admits transactions that modify the
wenzelm@18815
   688
theory before entering a proof state.  Transactions now always see a
wenzelm@18815
   689
quasi-functional intermediate checkpoint, both in interactive and
wenzelm@18590
   690
batch mode.
wenzelm@18567
   691
wenzelm@17878
   692
* Simplifier: the simpset of a running simplification process now
wenzelm@17878
   693
contains a proof context (cf. Simplifier.the_context), which is the
wenzelm@17878
   694
very context that the initial simpset has been retrieved from (by
wenzelm@17890
   695
simpset_of/local_simpset_of).  Consequently, all plug-in components
wenzelm@17878
   696
(solver, looper etc.) may depend on arbitrary proof data.
wenzelm@17878
   697
wenzelm@17878
   698
* Simplifier.inherit_context inherits the proof context (plus the
wenzelm@17878
   699
local bounds) of the current simplification process; any simproc
wenzelm@17878
   700
etc. that calls the Simplifier recursively should do this!  Removed
wenzelm@17878
   701
former Simplifier.inherit_bounds, which is already included here --
wenzelm@17890
   702
INCOMPATIBILITY.  Tools based on low-level rewriting may even have to
wenzelm@17890
   703
specify an explicit context using Simplifier.context/theory_context.
wenzelm@17878
   704
wenzelm@17878
   705
* Simplifier/Classical Reasoner: more abstract interfaces
wenzelm@17878
   706
change_simpset/claset for modifying the simpset/claset reference of a
wenzelm@17878
   707
theory; raw versions simpset/claset_ref etc. have been discontinued --
wenzelm@17878
   708
INCOMPATIBILITY.
wenzelm@17878
   709
wenzelm@18540
   710
* Provers: more generic wrt. syntax of object-logics, avoid hardwired
wenzelm@18540
   711
"Trueprop" etc.
wenzelm@18540
   712
wenzelm@17878
   713
wenzelm@17754
   714
wenzelm@17720
   715
New in Isabelle2005 (October 2005)
wenzelm@17720
   716
----------------------------------
wenzelm@14655
   717
wenzelm@14655
   718
*** General ***
wenzelm@14655
   719
nipkow@15130
   720
* Theory headers: the new header syntax for Isar theories is
nipkow@15130
   721
nipkow@15130
   722
  theory <name>
wenzelm@16234
   723
  imports <theory1> ... <theoryN>
wenzelm@16234
   724
  uses <file1> ... <fileM>
nipkow@15130
   725
  begin
nipkow@15130
   726
wenzelm@16234
   727
where the 'uses' part is optional.  The previous syntax
wenzelm@16234
   728
wenzelm@16234
   729
  theory <name> = <theory1> + ... + <theoryN>:
wenzelm@16234
   730
wenzelm@16717
   731
will disappear in the next release.  Use isatool fixheaders to convert
wenzelm@16717
   732
existing theory files.  Note that there is no change in ancient
wenzelm@17371
   733
non-Isar theories now, but these will disappear soon.
nipkow@15130
   734
berghofe@15475
   735
* Theory loader: parent theories can now also be referred to via
wenzelm@16234
   736
relative and absolute paths.
wenzelm@16234
   737
wenzelm@17408
   738
* Command 'find_theorems' searches for a list of criteria instead of a
wenzelm@17408
   739
list of constants. Known criteria are: intro, elim, dest, name:string,
wenzelm@17408
   740
simp:term, and any term. Criteria can be preceded by '-' to select
wenzelm@17408
   741
theorems that do not match. Intro, elim, dest select theorems that
wenzelm@17408
   742
match the current goal, name:s selects theorems whose fully qualified
wenzelm@17408
   743
name contain s, and simp:term selects all simplification rules whose
wenzelm@17408
   744
lhs match term.  Any other term is interpreted as pattern and selects
wenzelm@17408
   745
all theorems matching the pattern. Available in ProofGeneral under
wenzelm@17408
   746
'ProofGeneral -> Find Theorems' or C-c C-f.  Example:
wenzelm@16234
   747
wenzelm@17275
   748
  C-c C-f (100) "(_::nat) + _ + _" intro -name: "HOL."
wenzelm@16234
   749
wenzelm@16234
   750
prints the last 100 theorems matching the pattern "(_::nat) + _ + _",
wenzelm@16234
   751
matching the current goal as introduction rule and not having "HOL."
wenzelm@16234
   752
in their name (i.e. not being defined in theory HOL).
wenzelm@16013
   753
wenzelm@17408
   754
* Command 'thms_containing' has been discontinued in favour of
wenzelm@17408
   755
'find_theorems'; INCOMPATIBILITY.
wenzelm@17408
   756
wenzelm@17385
   757
* Communication with Proof General is now 8bit clean, which means that
wenzelm@17385
   758
Unicode text in UTF-8 encoding may be used within theory texts (both
wenzelm@17408
   759
formal and informal parts).  Cf. option -U of the Isabelle Proof
wenzelm@17538
   760
General interface.  Here are some simple examples (cf. src/HOL/ex):
wenzelm@17538
   761
wenzelm@17538
   762
  http://isabelle.in.tum.de/library/HOL/ex/Hebrew.html
wenzelm@17538
   763
  http://isabelle.in.tum.de/library/HOL/ex/Chinese.html
wenzelm@17385
   764
wenzelm@17425
   765
* Improved efficiency of the Simplifier and, to a lesser degree, the
wenzelm@17425
   766
Classical Reasoner.  Typical big applications run around 2 times
wenzelm@17425
   767
faster.
wenzelm@17425
   768
wenzelm@15703
   769
wenzelm@15703
   770
*** Document preparation ***
wenzelm@15703
   771
wenzelm@16234
   772
* Commands 'display_drafts' and 'print_drafts' perform simple output
wenzelm@16234
   773
of raw sources.  Only those symbols that do not require additional
wenzelm@16234
   774
LaTeX packages (depending on comments in isabellesym.sty) are
wenzelm@16234
   775
displayed properly, everything else is left verbatim.  isatool display
wenzelm@16234
   776
and isatool print are used as front ends (these are subject to the
wenzelm@16234
   777
DVI/PDF_VIEWER and PRINT_COMMAND settings, respectively).
wenzelm@16234
   778
wenzelm@17047
   779
* Command tags control specific markup of certain regions of text,
wenzelm@17047
   780
notably folding and hiding.  Predefined tags include "theory" (for
wenzelm@17047
   781
theory begin and end), "proof" for proof commands, and "ML" for
wenzelm@17047
   782
commands involving ML code; the additional tags "visible" and
wenzelm@17047
   783
"invisible" are unused by default.  Users may give explicit tag
wenzelm@17047
   784
specifications in the text, e.g. ''by %invisible (auto)''.  The
wenzelm@17047
   785
interpretation of tags is determined by the LaTeX job during document
wenzelm@17047
   786
preparation: see option -V of isatool usedir, or options -n and -t of
wenzelm@17047
   787
isatool document, or even the LaTeX macros \isakeeptag, \isafoldtag,
wenzelm@17047
   788
\isadroptag.
wenzelm@17047
   789
wenzelm@17047
   790
Several document versions may be produced at the same time via isatool
wenzelm@17047
   791
usedir (the generated index.html will link all of them).  Typical
wenzelm@17047
   792
specifications include ''-V document=theory,proof,ML'' to present
wenzelm@17047
   793
theory/proof/ML parts faithfully, ''-V outline=/proof,/ML'' to fold
wenzelm@17047
   794
proof and ML commands, and ''-V mutilated=-theory,-proof,-ML'' to omit
wenzelm@17047
   795
these parts without any formal replacement text.  The Isabelle site
wenzelm@17047
   796
default settings produce ''document'' and ''outline'' versions as
wenzelm@17047
   797
specified above.
wenzelm@16234
   798
haftmann@17402
   799
* Several new antiquotations:
wenzelm@15979
   800
wenzelm@15979
   801
  @{term_type term} prints a term with its type annotated;
wenzelm@15979
   802
wenzelm@15979
   803
  @{typeof term} prints the type of a term;
wenzelm@15979
   804
wenzelm@16234
   805
  @{const const} is the same as @{term const}, but checks that the
wenzelm@16234
   806
  argument is a known logical constant;
wenzelm@15979
   807
wenzelm@15979
   808
  @{term_style style term} and @{thm_style style thm} print a term or
wenzelm@16234
   809
  theorem applying a "style" to it
wenzelm@16234
   810
wenzelm@17117
   811
  @{ML text}
wenzelm@17117
   812
wenzelm@16234
   813
Predefined styles are 'lhs' and 'rhs' printing the lhs/rhs of
wenzelm@16234
   814
definitions, equations, inequations etc., 'concl' printing only the
schirmer@17393
   815
conclusion of a meta-logical statement theorem, and 'prem1' .. 'prem19'
wenzelm@16234
   816
to print the specified premise.  TermStyle.add_style provides an ML
wenzelm@16234
   817
interface for introducing further styles.  See also the "LaTeX Sugar"
wenzelm@17117
   818
document practical applications.  The ML antiquotation prints
wenzelm@17117
   819
type-checked ML expressions verbatim.
wenzelm@16234
   820
wenzelm@17259
   821
* Markup commands 'chapter', 'section', 'subsection', 'subsubsection',
wenzelm@17259
   822
and 'text' support optional locale specification '(in loc)', which
wenzelm@17269
   823
specifies the default context for interpreting antiquotations.  For
wenzelm@17269
   824
example: 'text (in lattice) {* @{thm inf_assoc}*}'.
wenzelm@17259
   825
wenzelm@17259
   826
* Option 'locale=NAME' of antiquotations specifies an alternative
wenzelm@17259
   827
context interpreting the subsequent argument.  For example: @{thm
wenzelm@17269
   828
[locale=lattice] inf_assoc}.
wenzelm@17259
   829
wenzelm@17097
   830
* Proper output of proof terms (@{prf ...} and @{full_prf ...}) within
wenzelm@17097
   831
a proof context.
wenzelm@17097
   832
wenzelm@17097
   833
* Proper output of antiquotations for theory commands involving a
wenzelm@17097
   834
proof context (such as 'locale' or 'theorem (in loc) ...').
wenzelm@17097
   835
wenzelm@17193
   836
* Delimiters of outer tokens (string etc.) now produce separate LaTeX
wenzelm@17193
   837
macros (\isachardoublequoteopen, isachardoublequoteclose etc.).
wenzelm@17193
   838
wenzelm@17193
   839
* isatool usedir: new option -C (default true) controls whether option
wenzelm@17193
   840
-D should include a copy of the original document directory; -C false
wenzelm@17193
   841
prevents unwanted effects such as copying of administrative CVS data.
wenzelm@17193
   842
wenzelm@16234
   843
wenzelm@16234
   844
*** Pure ***
wenzelm@16234
   845
wenzelm@16234
   846
* Considerably improved version of 'constdefs' command.  Now performs
wenzelm@16234
   847
automatic type-inference of declared constants; additional support for
wenzelm@16234
   848
local structure declarations (cf. locales and HOL records), see also
wenzelm@16234
   849
isar-ref manual.  Potential INCOMPATIBILITY: need to observe strictly
wenzelm@16234
   850
sequential dependencies of definitions within a single 'constdefs'
wenzelm@16234
   851
section; moreover, the declared name needs to be an identifier.  If
wenzelm@16234
   852
all fails, consider to fall back on 'consts' and 'defs' separately.
wenzelm@16234
   853
wenzelm@16234
   854
* Improved indexed syntax and implicit structures.  First of all,
wenzelm@16234
   855
indexed syntax provides a notational device for subscripted
wenzelm@16234
   856
application, using the new syntax \<^bsub>term\<^esub> for arbitrary
wenzelm@16234
   857
expressions.  Secondly, in a local context with structure
wenzelm@16234
   858
declarations, number indexes \<^sub>n or the empty index (default
wenzelm@16234
   859
number 1) refer to a certain fixed variable implicitly; option
wenzelm@16234
   860
show_structs controls printing of implicit structures.  Typical
wenzelm@16234
   861
applications of these concepts involve record types and locales.
wenzelm@16234
   862
wenzelm@16234
   863
* New command 'no_syntax' removes grammar declarations (and
wenzelm@16234
   864
translations) resulting from the given syntax specification, which is
wenzelm@16234
   865
interpreted in the same manner as for the 'syntax' command.
wenzelm@16234
   866
wenzelm@16234
   867
* 'Advanced' translation functions (parse_translation etc.) may depend
wenzelm@16234
   868
on the signature of the theory context being presently used for
wenzelm@16234
   869
parsing/printing, see also isar-ref manual.
wenzelm@16234
   870
wenzelm@16856
   871
* Improved 'oracle' command provides a type-safe interface to turn an
wenzelm@16856
   872
ML expression of type theory -> T -> term into a primitive rule of
wenzelm@16856
   873
type theory -> T -> thm (i.e. the functionality of Thm.invoke_oracle
wenzelm@16856
   874
is already included here); see also FOL/ex/IffExample.thy;
wenzelm@16856
   875
INCOMPATIBILITY.
wenzelm@16856
   876
wenzelm@17275
   877
* axclass: name space prefix for class "c" is now "c_class" (was "c"
wenzelm@17275
   878
before); "cI" is no longer bound, use "c.intro" instead.
wenzelm@17275
   879
INCOMPATIBILITY.  This change avoids clashes of fact bindings for
wenzelm@17275
   880
axclasses vs. locales.
wenzelm@17275
   881
wenzelm@16234
   882
* Improved internal renaming of symbolic identifiers -- attach primes
wenzelm@16234
   883
instead of base 26 numbers.
wenzelm@16234
   884
wenzelm@16234
   885
* New flag show_question_marks controls printing of leading question
wenzelm@16234
   886
marks in schematic variable names.
wenzelm@16234
   887
wenzelm@16234
   888
* In schematic variable names, *any* symbol following \<^isub> or
wenzelm@16234
   889
\<^isup> is now treated as part of the base name.  For example, the
wenzelm@16234
   890
following works without printing of awkward ".0" indexes:
wenzelm@16234
   891
wenzelm@16234
   892
  lemma "x\<^isub>1 = x\<^isub>2 ==> x\<^isub>2 = x\<^isub>1"
wenzelm@16234
   893
    by simp
wenzelm@16234
   894
wenzelm@16234
   895
* Inner syntax includes (*(*nested*) comments*).
wenzelm@16234
   896
wenzelm@17548
   897
* Pretty printer now supports unbreakable blocks, specified in mixfix
wenzelm@16234
   898
annotations as "(00...)".
wenzelm@16234
   899
wenzelm@16234
   900
* Clear separation of logical types and nonterminals, where the latter
wenzelm@16234
   901
may only occur in 'syntax' specifications or type abbreviations.
wenzelm@16234
   902
Before that distinction was only partially implemented via type class
wenzelm@16234
   903
"logic" vs. "{}".  Potential INCOMPATIBILITY in rare cases of improper
wenzelm@16234
   904
use of 'types'/'consts' instead of 'nonterminals'/'syntax'.  Some very
wenzelm@16234
   905
exotic syntax specifications may require further adaption
wenzelm@17691
   906
(e.g. Cube/Cube.thy).
wenzelm@16234
   907
wenzelm@16234
   908
* Removed obsolete type class "logic", use the top sort {} instead.
wenzelm@16234
   909
Note that non-logical types should be declared as 'nonterminals'
wenzelm@16234
   910
rather than 'types'.  INCOMPATIBILITY for new object-logic
wenzelm@16234
   911
specifications.
wenzelm@16234
   912
ballarin@17095
   913
* Attributes 'induct' and 'cases': type or set names may now be
ballarin@17095
   914
locally fixed variables as well.
ballarin@17095
   915
wenzelm@16234
   916
* Simplifier: can now control the depth to which conditional rewriting
wenzelm@16234
   917
is traced via the PG menu Isabelle -> Settings -> Trace Simp Depth
wenzelm@16234
   918
Limit.
wenzelm@16234
   919
wenzelm@16234
   920
* Simplifier: simplification procedures may now take the current
wenzelm@16234
   921
simpset into account (cf. Simplifier.simproc(_i) / mk_simproc
wenzelm@16234
   922
interface), which is very useful for calling the Simplifier
wenzelm@16234
   923
recursively.  Minor INCOMPATIBILITY: the 'prems' argument of simprocs
wenzelm@16234
   924
is gone -- use prems_of_ss on the simpset instead.  Moreover, the
wenzelm@16234
   925
low-level mk_simproc no longer applies Logic.varify internally, to
wenzelm@16234
   926
allow for use in a context of fixed variables.
wenzelm@16234
   927
wenzelm@16234
   928
* thin_tac now works even if the assumption being deleted contains !!
wenzelm@16234
   929
or ==>.  More generally, erule now works even if the major premise of
wenzelm@16234
   930
the elimination rule contains !! or ==>.
wenzelm@16234
   931
wenzelm@17597
   932
* Method 'rules' has been renamed to 'iprover'. INCOMPATIBILITY.
nipkow@17590
   933
wenzelm@16234
   934
* Reorganized bootstrapping of the Pure theories; CPure is now derived
wenzelm@16234
   935
from Pure, which contains all common declarations already.  Both
wenzelm@16234
   936
theories are defined via plain Isabelle/Isar .thy files.
wenzelm@16234
   937
INCOMPATIBILITY: elements of CPure (such as the CPure.intro /
wenzelm@16234
   938
CPure.elim / CPure.dest attributes) now appear in the Pure name space;
wenzelm@16234
   939
use isatool fixcpure to adapt your theory and ML sources.
wenzelm@16234
   940
wenzelm@16234
   941
* New syntax 'name(i-j, i-, i, ...)' for referring to specific
wenzelm@16234
   942
selections of theorems in named facts via index ranges.
wenzelm@16234
   943
wenzelm@17097
   944
* 'print_theorems': in theory mode, really print the difference
wenzelm@17097
   945
wrt. the last state (works for interactive theory development only),
wenzelm@17097
   946
in proof mode print all local facts (cf. 'print_facts');
wenzelm@17097
   947
wenzelm@17397
   948
* 'hide': option '(open)' hides only base names.
wenzelm@17397
   949
wenzelm@17275
   950
* More efficient treatment of intermediate checkpoints in interactive
wenzelm@17275
   951
theory development.
wenzelm@17275
   952
berghofe@17663
   953
* Code generator is now invoked via code_module (incremental code
wenzelm@17664
   954
generation) and code_library (modular code generation, ML structures
wenzelm@17664
   955
for each theory).  INCOMPATIBILITY: new keywords 'file' and 'contains'
wenzelm@17664
   956
must be quoted when used as identifiers.
wenzelm@17664
   957
wenzelm@17664
   958
* New 'value' command for reading, evaluating and printing terms using
wenzelm@17664
   959
the code generator.  INCOMPATIBILITY: command keyword 'value' must be
wenzelm@17664
   960
quoted when used as identifier.
berghofe@17663
   961
wenzelm@16234
   962
wenzelm@16234
   963
*** Locales ***
ballarin@17095
   964
wenzelm@17385
   965
* New commands for the interpretation of locale expressions in
wenzelm@17385
   966
theories (1), locales (2) and proof contexts (3).  These generate
wenzelm@17385
   967
proof obligations from the expression specification.  After the
wenzelm@17385
   968
obligations have been discharged, theorems of the expression are added
wenzelm@17385
   969
to the theory, target locale or proof context.  The synopsis of the
wenzelm@17385
   970
commands is a follows:
wenzelm@17385
   971
ballarin@17095
   972
  (1) interpretation expr inst
ballarin@17095
   973
  (2) interpretation target < expr
ballarin@17095
   974
  (3) interpret expr inst
wenzelm@17385
   975
ballarin@17095
   976
Interpretation in theories and proof contexts require a parameter
ballarin@17095
   977
instantiation of terms from the current context.  This is applied to
wenzelm@17385
   978
specifications and theorems of the interpreted expression.
wenzelm@17385
   979
Interpretation in locales only permits parameter renaming through the
wenzelm@17385
   980
locale expression.  Interpretation is smart in that interpretations
wenzelm@17385
   981
that are active already do not occur in proof obligations, neither are
wenzelm@17385
   982
instantiated theorems stored in duplicate.  Use 'print_interps' to
wenzelm@17385
   983
inspect active interpretations of a particular locale.  For details,
ballarin@17436
   984
see the Isar Reference manual.  Examples can be found in
ballarin@17436
   985
HOL/Finite_Set.thy and HOL/Algebra/UnivPoly.thy.
wenzelm@16234
   986
wenzelm@16234
   987
INCOMPATIBILITY: former 'instantiate' has been withdrawn, use
wenzelm@16234
   988
'interpret' instead.
wenzelm@16234
   989
wenzelm@17385
   990
* New context element 'constrains' for adding type constraints to
wenzelm@17385
   991
parameters.
wenzelm@17385
   992
wenzelm@17385
   993
* Context expressions: renaming of parameters with syntax
wenzelm@17385
   994
redeclaration.
ballarin@17095
   995
ballarin@17095
   996
* Locale declaration: 'includes' disallowed.
ballarin@17095
   997
wenzelm@16234
   998
* Proper static binding of attribute syntax -- i.e. types / terms /
wenzelm@16234
   999
facts mentioned as arguments are always those of the locale definition
wenzelm@16234
  1000
context, independently of the context of later invocations.  Moreover,
wenzelm@16234
  1001
locale operations (renaming and type / term instantiation) are applied
wenzelm@16234
  1002
to attribute arguments as expected.
wenzelm@16234
  1003
wenzelm@16234
  1004
INCOMPATIBILITY of the ML interface: always pass Attrib.src instead of
wenzelm@16234
  1005
actual attributes; rare situations may require Attrib.attribute to
wenzelm@16234
  1006
embed those attributes into Attrib.src that lack concrete syntax.
wenzelm@16234
  1007
Attribute implementations need to cooperate properly with the static
wenzelm@16234
  1008
binding mechanism.  Basic parsers Args.XXX_typ/term/prop and
wenzelm@16234
  1009
Attrib.XXX_thm etc. already do the right thing without further
wenzelm@16234
  1010
intervention.  Only unusual applications -- such as "where" or "of"
wenzelm@16234
  1011
(cf. src/Pure/Isar/attrib.ML), which process arguments depending both
wenzelm@16234
  1012
on the context and the facts involved -- may have to assign parsed
wenzelm@16234
  1013
values to argument tokens explicitly.
wenzelm@16234
  1014
wenzelm@16234
  1015
* Changed parameter management in theorem generation for long goal
wenzelm@16234
  1016
statements with 'includes'.  INCOMPATIBILITY: produces a different
wenzelm@16234
  1017
theorem statement in rare situations.
wenzelm@16234
  1018
ballarin@17228
  1019
* Locale inspection command 'print_locale' omits notes elements.  Use
ballarin@17228
  1020
'print_locale!' to have them included in the output.
ballarin@17228
  1021
wenzelm@16234
  1022
wenzelm@16234
  1023
*** Provers ***
wenzelm@16234
  1024
wenzelm@16234
  1025
* Provers/hypsubst.ML: improved version of the subst method, for
wenzelm@16234
  1026
single-step rewriting: it now works in bound variable contexts. New is
wenzelm@16234
  1027
'subst (asm)', for rewriting an assumption.  INCOMPATIBILITY: may
wenzelm@16234
  1028
rewrite a different subterm than the original subst method, which is
wenzelm@16234
  1029
still available as 'simplesubst'.
wenzelm@16234
  1030
wenzelm@16234
  1031
* Provers/quasi.ML: new transitivity reasoners for transitivity only
wenzelm@16234
  1032
and quasi orders.
wenzelm@16234
  1033
wenzelm@16234
  1034
* Provers/trancl.ML: new transitivity reasoner for transitive and
wenzelm@16234
  1035
reflexive-transitive closure of relations.
wenzelm@16234
  1036
wenzelm@16234
  1037
* Provers/blast.ML: new reference depth_limit to make blast's depth
wenzelm@16234
  1038
limit (previously hard-coded with a value of 20) user-definable.
wenzelm@16234
  1039
wenzelm@16234
  1040
* Provers/simplifier.ML has been moved to Pure, where Simplifier.setup
wenzelm@16234
  1041
is peformed already.  Object-logics merely need to finish their
wenzelm@16234
  1042
initial simpset configuration as before.  INCOMPATIBILITY.
wenzelm@15703
  1043
berghofe@15475
  1044
schirmer@14700
  1045
*** HOL ***
schirmer@14700
  1046
wenzelm@16234
  1047
* Symbolic syntax of Hilbert Choice Operator is now as follows:
wenzelm@14878
  1048
wenzelm@14878
  1049
  syntax (epsilon)
wenzelm@14878
  1050
    "_Eps" :: "[pttrn, bool] => 'a"    ("(3\<some>_./ _)" [0, 10] 10)
wenzelm@14878
  1051
wenzelm@16234
  1052
The symbol \<some> is displayed as the alternative epsilon of LaTeX
wenzelm@16234
  1053
and x-symbol; use option '-m epsilon' to get it actually printed.
wenzelm@16234
  1054
Moreover, the mathematically important symbolic identifier \<epsilon>
wenzelm@16234
  1055
becomes available as variable, constant etc.  INCOMPATIBILITY,
wenzelm@16234
  1056
wenzelm@16234
  1057
* "x > y" abbreviates "y < x" and "x >= y" abbreviates "y <= x".
wenzelm@16234
  1058
Similarly for all quantifiers: "ALL x > y" etc.  The x-symbol for >=
wenzelm@17371
  1059
is \<ge>. New transitivity rules have been added to HOL/Orderings.thy to
avigad@17016
  1060
support corresponding Isar calculations.
wenzelm@16234
  1061
wenzelm@16234
  1062
* "{x:A. P}" abbreviates "{x. x:A & P}", and similarly for "\<in>"
wenzelm@16234
  1063
instead of ":".
wenzelm@16234
  1064
wenzelm@16234
  1065
* theory SetInterval: changed the syntax for open intervals:
wenzelm@16234
  1066
wenzelm@16234
  1067
  Old       New
wenzelm@16234
  1068
  {..n(}    {..<n}
wenzelm@16234
  1069
  {)n..}    {n<..}
wenzelm@16234
  1070
  {m..n(}   {m..<n}
wenzelm@16234
  1071
  {)m..n}   {m<..n}
wenzelm@16234
  1072
  {)m..n(}  {m<..<n}
wenzelm@16234
  1073
wenzelm@16234
  1074
The old syntax is still supported but will disappear in the next
wenzelm@16234
  1075
release.  For conversion use the following Emacs search and replace
wenzelm@16234
  1076
patterns (these are not perfect but work quite well):
nipkow@15046
  1077
nipkow@15046
  1078
  {)\([^\.]*\)\.\.  ->  {\1<\.\.}
nipkow@15046
  1079
  \.\.\([^(}]*\)(}  ->  \.\.<\1}
nipkow@15046
  1080
wenzelm@17533
  1081
* Theory Commutative_Ring (in Library): method comm_ring for proving
wenzelm@17533
  1082
equalities in commutative rings; method 'algebra' provides a generic
wenzelm@17533
  1083
interface.
wenzelm@17389
  1084
wenzelm@17389
  1085
* Theory Finite_Set: changed the syntax for 'setsum', summation over
wenzelm@16234
  1086
finite sets: "setsum (%x. e) A", which used to be "\<Sum>x:A. e", is
wenzelm@17371
  1087
now either "SUM x:A. e" or "\<Sum>x \<in> A. e". The bound variable can
paulson@17189
  1088
be a tuple pattern.
wenzelm@16234
  1089
wenzelm@16234
  1090
Some new syntax forms are available:
wenzelm@16234
  1091
wenzelm@16234
  1092
  "\<Sum>x | P. e"      for     "setsum (%x. e) {x. P}"
wenzelm@16234
  1093
  "\<Sum>x = a..b. e"   for     "setsum (%x. e) {a..b}"
wenzelm@16234
  1094
  "\<Sum>x = a..<b. e"  for     "setsum (%x. e) {a..<b}"
wenzelm@16234
  1095
  "\<Sum>x < k. e"      for     "setsum (%x. e) {..<k}"
wenzelm@16234
  1096
wenzelm@16234
  1097
The latter form "\<Sum>x < k. e" used to be based on a separate
wenzelm@16234
  1098
function "Summation", which has been discontinued.
wenzelm@16234
  1099
wenzelm@16234
  1100
* theory Finite_Set: in structured induction proofs, the insert case
wenzelm@16234
  1101
is now 'case (insert x F)' instead of the old counterintuitive 'case
wenzelm@16234
  1102
(insert F x)'.
wenzelm@16234
  1103
wenzelm@16234
  1104
* The 'refute' command has been extended to support a much larger
wenzelm@16234
  1105
fragment of HOL, including axiomatic type classes, constdefs and
wenzelm@16234
  1106
typedefs, inductive datatypes and recursion.
wenzelm@16234
  1107
webertj@17700
  1108
* New tactics 'sat' and 'satx' to prove propositional tautologies.
webertj@17700
  1109
Requires zChaff with proof generation to be installed.  See
webertj@17700
  1110
HOL/ex/SAT_Examples.thy for examples.
webertj@17619
  1111
wenzelm@16234
  1112
* Datatype induction via method 'induct' now preserves the name of the
wenzelm@16234
  1113
induction variable. For example, when proving P(xs::'a list) by
wenzelm@16234
  1114
induction on xs, the induction step is now P(xs) ==> P(a#xs) rather
wenzelm@16234
  1115
than P(list) ==> P(a#list) as previously.  Potential INCOMPATIBILITY
wenzelm@16234
  1116
in unstructured proof scripts.
wenzelm@16234
  1117
wenzelm@16234
  1118
* Reworked implementation of records.  Improved scalability for
wenzelm@16234
  1119
records with many fields, avoiding performance problems for type
wenzelm@16234
  1120
inference. Records are no longer composed of nested field types, but
wenzelm@16234
  1121
of nested extension types. Therefore the record type only grows linear
wenzelm@16234
  1122
in the number of extensions and not in the number of fields.  The
wenzelm@16234
  1123
top-level (users) view on records is preserved.  Potential
wenzelm@16234
  1124
INCOMPATIBILITY only in strange cases, where the theory depends on the
wenzelm@16234
  1125
old record representation. The type generated for a record is called
wenzelm@16234
  1126
<record_name>_ext_type.
wenzelm@16234
  1127
wenzelm@16234
  1128
Flag record_quick_and_dirty_sensitive can be enabled to skip the
wenzelm@16234
  1129
proofs triggered by a record definition or a simproc (if
wenzelm@16234
  1130
quick_and_dirty is enabled).  Definitions of large records can take
wenzelm@16234
  1131
quite long.
wenzelm@16234
  1132
wenzelm@16234
  1133
New simproc record_upd_simproc for simplification of multiple record
wenzelm@16234
  1134
updates enabled by default.  Moreover, trivial updates are also
wenzelm@16234
  1135
removed: r(|x := x r|) = r.  INCOMPATIBILITY: old proofs break
wenzelm@16234
  1136
occasionally, since simplification is more powerful by default.
wenzelm@16234
  1137
wenzelm@17275
  1138
* typedef: proper support for polymorphic sets, which contain extra
wenzelm@17275
  1139
type-variables in the term.
wenzelm@17275
  1140
wenzelm@16234
  1141
* Simplifier: automatically reasons about transitivity chains
wenzelm@16234
  1142
involving "trancl" (r^+) and "rtrancl" (r^*) by setting up tactics
wenzelm@16234
  1143
provided by Provers/trancl.ML as additional solvers.  INCOMPATIBILITY:
wenzelm@16234
  1144
old proofs break occasionally as simplification may now solve more
wenzelm@16234
  1145
goals than previously.
wenzelm@16234
  1146
wenzelm@16234
  1147
* Simplifier: converts x <= y into x = y if assumption y <= x is
wenzelm@16234
  1148
present.  Works for all partial orders (class "order"), in particular
wenzelm@16234
  1149
numbers and sets.  For linear orders (e.g. numbers) it treats ~ x < y
wenzelm@16234
  1150
just like y <= x.
wenzelm@16234
  1151
wenzelm@16234
  1152
* Simplifier: new simproc for "let x = a in f x".  If a is a free or
wenzelm@16234
  1153
bound variable or a constant then the let is unfolded.  Otherwise
wenzelm@16234
  1154
first a is simplified to b, and then f b is simplified to g. If
wenzelm@16234
  1155
possible we abstract b from g arriving at "let x = b in h x",
wenzelm@16234
  1156
otherwise we unfold the let and arrive at g.  The simproc can be
wenzelm@16234
  1157
enabled/disabled by the reference use_let_simproc.  Potential
wenzelm@16234
  1158
INCOMPATIBILITY since simplification is more powerful by default.
webertj@15776
  1159
paulson@16563
  1160
* Classical reasoning: the meson method now accepts theorems as arguments.
paulson@16563
  1161
paulson@17595
  1162
* Prover support: pre-release of the Isabelle-ATP linkup, which runs background
paulson@17595
  1163
jobs to provide advice on the provability of subgoals.
paulson@17595
  1164
wenzelm@16891
  1165
* Theory OrderedGroup and Ring_and_Field: various additions and
wenzelm@16891
  1166
improvements to faciliate calculations involving equalities and
wenzelm@16891
  1167
inequalities.
wenzelm@16891
  1168
wenzelm@16891
  1169
The following theorems have been eliminated or modified
wenzelm@16891
  1170
(INCOMPATIBILITY):
avigad@16888
  1171
avigad@16888
  1172
  abs_eq             now named abs_of_nonneg
wenzelm@17371
  1173
  abs_of_ge_0        now named abs_of_nonneg
wenzelm@17371
  1174
  abs_minus_eq       now named abs_of_nonpos
avigad@16888
  1175
  imp_abs_id         now named abs_of_nonneg
avigad@16888
  1176
  imp_abs_neg_id     now named abs_of_nonpos
avigad@16888
  1177
  mult_pos           now named mult_pos_pos
avigad@16888
  1178
  mult_pos_le        now named mult_nonneg_nonneg
avigad@16888
  1179
  mult_pos_neg_le    now named mult_nonneg_nonpos
avigad@16888
  1180
  mult_pos_neg2_le   now named mult_nonneg_nonpos2
avigad@16888
  1181
  mult_neg           now named mult_neg_neg
avigad@16888
  1182
  mult_neg_le        now named mult_nonpos_nonpos
avigad@16888
  1183
wenzelm@16891
  1184
* Theory Parity: added rules for simplifying exponents.
wenzelm@16891
  1185
nipkow@17092
  1186
* Theory List:
nipkow@17092
  1187
nipkow@17092
  1188
The following theorems have been eliminated or modified
nipkow@17092
  1189
(INCOMPATIBILITY):
nipkow@17092
  1190
nipkow@17092
  1191
  list_all_Nil       now named list_all.simps(1)
nipkow@17092
  1192
  list_all_Cons      now named list_all.simps(2)
nipkow@17092
  1193
  list_all_conv      now named list_all_iff
nipkow@17092
  1194
  set_mem_eq         now named mem_iff
nipkow@17092
  1195
wenzelm@16929
  1196
* Theories SetsAndFunctions and BigO (see HOL/Library) support
wenzelm@16929
  1197
asymptotic "big O" calculations.  See the notes in BigO.thy.
wenzelm@16929
  1198
avigad@16888
  1199
avigad@16888
  1200
*** HOL-Complex ***
avigad@16888
  1201
wenzelm@16891
  1202
* Theory RealDef: better support for embedding natural numbers and
wenzelm@16891
  1203
integers in the reals.
wenzelm@16891
  1204
wenzelm@16891
  1205
The following theorems have been eliminated or modified
wenzelm@16891
  1206
(INCOMPATIBILITY):
wenzelm@16891
  1207
avigad@17016
  1208
  exp_ge_add_one_self  now requires no hypotheses
avigad@17016
  1209
  real_of_int_add      reversed direction of equality (use [symmetric])
avigad@17016
  1210
  real_of_int_minus    reversed direction of equality (use [symmetric])
avigad@17016
  1211
  real_of_int_diff     reversed direction of equality (use [symmetric])
avigad@17016
  1212
  real_of_int_mult     reversed direction of equality (use [symmetric])
wenzelm@16891
  1213
wenzelm@16891
  1214
* Theory RComplete: expanded support for floor and ceiling functions.
avigad@16888
  1215
avigad@16962
  1216
* Theory Ln is new, with properties of the natural logarithm
avigad@16962
  1217
wenzelm@17423
  1218
* Hyperreal: There is a new type constructor "star" for making
wenzelm@17423
  1219
nonstandard types.  The old type names are now type synonyms:
wenzelm@17423
  1220
wenzelm@17423
  1221
  hypreal = real star
wenzelm@17423
  1222
  hypnat = nat star
wenzelm@17423
  1223
  hcomplex = complex star
wenzelm@17423
  1224
wenzelm@17423
  1225
* Hyperreal: Many groups of similarly-defined constants have been
huffman@17442
  1226
replaced by polymorphic versions (INCOMPATIBILITY):
wenzelm@17423
  1227
wenzelm@17423
  1228
  star_of <-- hypreal_of_real, hypnat_of_nat, hcomplex_of_complex
wenzelm@17423
  1229
wenzelm@17423
  1230
  starset      <-- starsetNat, starsetC
wenzelm@17423
  1231
  *s*          <-- *sNat*, *sc*
wenzelm@17423
  1232
  starset_n    <-- starsetNat_n, starsetC_n
wenzelm@17423
  1233
  *sn*         <-- *sNatn*, *scn*
wenzelm@17423
  1234
  InternalSets <-- InternalNatSets, InternalCSets
wenzelm@17423
  1235
huffman@17442
  1236
  starfun      <-- starfun{Nat,Nat2,C,RC,CR}
wenzelm@17423
  1237
  *f*          <-- *fNat*, *fNat2*, *fc*, *fRc*, *fcR*
huffman@17442
  1238
  starfun_n    <-- starfun{Nat,Nat2,C,RC,CR}_n
wenzelm@17423
  1239
  *fn*         <-- *fNatn*, *fNat2n*, *fcn*, *fRcn*, *fcRn*
huffman@17442
  1240
  InternalFuns <-- InternalNatFuns, InternalNatFuns2, Internal{C,RC,CR}Funs
wenzelm@17423
  1241
wenzelm@17423
  1242
* Hyperreal: Many type-specific theorems have been removed in favor of
huffman@17442
  1243
theorems specific to various axiomatic type classes (INCOMPATIBILITY):
huffman@17442
  1244
huffman@17442
  1245
  add_commute <-- {hypreal,hypnat,hcomplex}_add_commute
huffman@17442
  1246
  add_assoc   <-- {hypreal,hypnat,hcomplex}_add_assocs
huffman@17442
  1247
  OrderedGroup.add_0 <-- {hypreal,hypnat,hcomplex}_add_zero_left
huffman@17442
  1248
  OrderedGroup.add_0_right <-- {hypreal,hcomplex}_add_zero_right
wenzelm@17423
  1249
  right_minus <-- hypreal_add_minus
huffman@17442
  1250
  left_minus <-- {hypreal,hcomplex}_add_minus_left
huffman@17442
  1251
  mult_commute <-- {hypreal,hypnat,hcomplex}_mult_commute
huffman@17442
  1252
  mult_assoc <-- {hypreal,hypnat,hcomplex}_mult_assoc
huffman@17442
  1253
  mult_1_left <-- {hypreal,hypnat}_mult_1, hcomplex_mult_one_left
wenzelm@17423
  1254
  mult_1_right <-- hcomplex_mult_one_right
wenzelm@17423
  1255
  mult_zero_left <-- hcomplex_mult_zero_left
huffman@17442
  1256
  left_distrib <-- {hypreal,hypnat,hcomplex}_add_mult_distrib
wenzelm@17423
  1257
  right_distrib <-- hypnat_add_mult_distrib2
huffman@17442
  1258
  zero_neq_one <-- {hypreal,hypnat,hcomplex}_zero_not_eq_one
wenzelm@17423
  1259
  right_inverse <-- hypreal_mult_inverse
wenzelm@17423
  1260
  left_inverse <-- hypreal_mult_inverse_left, hcomplex_mult_inv_left
huffman@17442
  1261
  order_refl <-- {hypreal,hypnat}_le_refl
huffman@17442
  1262
  order_trans <-- {hypreal,hypnat}_le_trans
huffman@17442
  1263
  order_antisym <-- {hypreal,hypnat}_le_anti_sym
huffman@17442
  1264
  order_less_le <-- {hypreal,hypnat}_less_le
huffman@17442
  1265
  linorder_linear <-- {hypreal,hypnat}_le_linear
huffman@17442
  1266
  add_left_mono <-- {hypreal,hypnat}_add_left_mono
huffman@17442
  1267
  mult_strict_left_mono <-- {hypreal,hypnat}_mult_less_mono2
wenzelm@17423
  1268
  add_nonneg_nonneg <-- hypreal_le_add_order
wenzelm@17423
  1269
wenzelm@17423
  1270
* Hyperreal: Separate theorems having to do with type-specific
wenzelm@17423
  1271
versions of constants have been merged into theorems that apply to the
huffman@17442
  1272
new polymorphic constants (INCOMPATIBILITY):
huffman@17442
  1273
huffman@17442
  1274
  STAR_UNIV_set <-- {STAR_real,NatStar_real,STARC_complex}_set
huffman@17442
  1275
  STAR_empty_set <-- {STAR,NatStar,STARC}_empty_set
huffman@17442
  1276
  STAR_Un <-- {STAR,NatStar,STARC}_Un
huffman@17442
  1277
  STAR_Int <-- {STAR,NatStar,STARC}_Int
huffman@17442
  1278
  STAR_Compl <-- {STAR,NatStar,STARC}_Compl
huffman@17442
  1279
  STAR_subset <-- {STAR,NatStar,STARC}_subset
huffman@17442
  1280
  STAR_mem <-- {STAR,NatStar,STARC}_mem
huffman@17442
  1281
  STAR_mem_Compl <-- {STAR,STARC}_mem_Compl
huffman@17442
  1282
  STAR_diff <-- {STAR,STARC}_diff
huffman@17442
  1283
  STAR_star_of_image_subset <-- {STAR_hypreal_of_real, NatStar_hypreal_of_real,
huffman@17442
  1284
    STARC_hcomplex_of_complex}_image_subset
huffman@17442
  1285
  starset_n_Un <-- starset{Nat,C}_n_Un
huffman@17442
  1286
  starset_n_Int <-- starset{Nat,C}_n_Int
huffman@17442
  1287
  starset_n_Compl <-- starset{Nat,C}_n_Compl
huffman@17442
  1288
  starset_n_diff <-- starset{Nat,C}_n_diff
huffman@17442
  1289
  InternalSets_Un <-- Internal{Nat,C}Sets_Un
huffman@17442
  1290
  InternalSets_Int <-- Internal{Nat,C}Sets_Int
huffman@17442
  1291
  InternalSets_Compl <-- Internal{Nat,C}Sets_Compl
huffman@17442
  1292
  InternalSets_diff <-- Internal{Nat,C}Sets_diff
huffman@17442
  1293
  InternalSets_UNIV_diff <-- Internal{Nat,C}Sets_UNIV_diff
huffman@17442
  1294
  InternalSets_starset_n <-- Internal{Nat,C}Sets_starset{Nat,C}_n
huffman@17442
  1295
  starset_starset_n_eq <-- starset{Nat,C}_starset{Nat,C}_n_eq
huffman@17442
  1296
  starset_n_starset <-- starset{Nat,C}_n_starset{Nat,C}
huffman@17442
  1297
  starfun_n_starfun <-- starfun{Nat,Nat2,C,RC,CR}_n_starfun{Nat,Nat2,C,RC,CR}
huffman@17442
  1298
  starfun <-- starfun{Nat,Nat2,C,RC,CR}
huffman@17442
  1299
  starfun_mult <-- starfun{Nat,Nat2,C,RC,CR}_mult
huffman@17442
  1300
  starfun_add <-- starfun{Nat,Nat2,C,RC,CR}_add
huffman@17442
  1301
  starfun_minus <-- starfun{Nat,Nat2,C,RC,CR}_minus
huffman@17442
  1302
  starfun_diff <-- starfun{C,RC,CR}_diff
huffman@17442
  1303
  starfun_o <-- starfun{NatNat2,Nat2,_stafunNat,C,C_starfunRC,_starfunCR}_o
huffman@17442
  1304
  starfun_o2 <-- starfun{NatNat2,_stafunNat,C,C_starfunRC,_starfunCR}_o2
huffman@17442
  1305
  starfun_const_fun <-- starfun{Nat,Nat2,C,RC,CR}_const_fun
huffman@17442
  1306
  starfun_inverse <-- starfun{Nat,C,RC,CR}_inverse
huffman@17442
  1307
  starfun_eq <-- starfun{Nat,Nat2,C,RC,CR}_eq
huffman@17442
  1308
  starfun_eq_iff <-- starfun{C,RC,CR}_eq_iff
wenzelm@17423
  1309
  starfun_Id <-- starfunC_Id
huffman@17442
  1310
  starfun_approx <-- starfun{Nat,CR}_approx
huffman@17442
  1311
  starfun_capprox <-- starfun{C,RC}_capprox
wenzelm@17423
  1312
  starfun_abs <-- starfunNat_rabs
huffman@17442
  1313
  starfun_lambda_cancel <-- starfun{C,CR,RC}_lambda_cancel
huffman@17442
  1314
  starfun_lambda_cancel2 <-- starfun{C,CR,RC}_lambda_cancel2
wenzelm@17423
  1315
  starfun_mult_HFinite_approx <-- starfunCR_mult_HFinite_capprox
huffman@17442
  1316
  starfun_mult_CFinite_capprox <-- starfun{C,RC}_mult_CFinite_capprox
huffman@17442
  1317
  starfun_add_capprox <-- starfun{C,RC}_add_capprox
wenzelm@17423
  1318
  starfun_add_approx <-- starfunCR_add_approx
wenzelm@17423
  1319
  starfun_inverse_inverse <-- starfunC_inverse_inverse
huffman@17442
  1320
  starfun_divide <-- starfun{C,CR,RC}_divide
huffman@17442
  1321
  starfun_n <-- starfun{Nat,C}_n
huffman@17442
  1322
  starfun_n_mult <-- starfun{Nat,C}_n_mult
huffman@17442
  1323
  starfun_n_add <-- starfun{Nat,C}_n_add
wenzelm@17423
  1324
  starfun_n_add_minus <-- starfunNat_n_add_minus
huffman@17442
  1325
  starfun_n_const_fun <-- starfun{Nat,C}_n_const_fun
huffman@17442
  1326
  starfun_n_minus <-- starfun{Nat,C}_n_minus
huffman@17442
  1327
  starfun_n_eq <-- starfun{Nat,C}_n_eq
huffman@17442
  1328
huffman@17442
  1329
  star_n_add <-- {hypreal,hypnat,hcomplex}_add
huffman@17442
  1330
  star_n_minus <-- {hypreal,hcomplex}_minus
huffman@17442
  1331
  star_n_diff <-- {hypreal,hcomplex}_diff
huffman@17442
  1332
  star_n_mult <-- {hypreal,hcomplex}_mult
huffman@17442
  1333
  star_n_inverse <-- {hypreal,hcomplex}_inverse
huffman@17442
  1334
  star_n_le <-- {hypreal,hypnat}_le
huffman@17442
  1335
  star_n_less <-- {hypreal,hypnat}_less
huffman@17442
  1336
  star_n_zero_num <-- {hypreal,hypnat,hcomplex}_zero_num
huffman@17442
  1337
  star_n_one_num <-- {hypreal,hypnat,hcomplex}_one_num
wenzelm@17423
  1338
  star_n_abs <-- hypreal_hrabs
wenzelm@17423
  1339
  star_n_divide <-- hcomplex_divide
wenzelm@17423
  1340
huffman@17442
  1341
  star_of_add <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_add
huffman@17442
  1342
  star_of_minus <-- {hypreal_of_real,hcomplex_of_complex}_minus
wenzelm@17423
  1343
  star_of_diff <-- hypreal_of_real_diff
huffman@17442
  1344
  star_of_mult <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_mult
huffman@17442
  1345
  star_of_one <-- {hypreal_of_real,hcomplex_of_complex}_one
huffman@17442
  1346
  star_of_zero <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_zero
huffman@17442
  1347
  star_of_le <-- {hypreal_of_real,hypnat_of_nat}_le_iff
huffman@17442
  1348
  star_of_less <-- {hypreal_of_real,hypnat_of_nat}_less_iff
huffman@17442
  1349
  star_of_eq <-- {hypreal_of_real,hypnat_of_nat,hcomplex_of_complex}_eq_iff
huffman@17442
  1350
  star_of_inverse <-- {hypreal_of_real,hcomplex_of_complex}_inverse
huffman@17442
  1351
  star_of_divide <-- {hypreal_of_real,hcomplex_of_complex}_divide
huffman@17442
  1352
  star_of_of_nat <-- {hypreal_of_real,hcomplex_of_complex}_of_nat
huffman@17442
  1353
  star_of_of_int <-- {hypreal_of_real,hcomplex_of_complex}_of_int
huffman@17442
  1354
  star_of_number_of <-- {hypreal,hcomplex}_number_of
wenzelm@17423
  1355
  star_of_number_less <-- number_of_less_hypreal_of_real_iff
wenzelm@17423
  1356
  star_of_number_le <-- number_of_le_hypreal_of_real_iff
wenzelm@17423
  1357
  star_of_eq_number <-- hypreal_of_real_eq_number_of_iff
wenzelm@17423
  1358
  star_of_less_number <-- hypreal_of_real_less_number_of_iff
wenzelm@17423
  1359
  star_of_le_number <-- hypreal_of_real_le_number_of_iff
wenzelm@17423
  1360
  star_of_power <-- hypreal_of_real_power
wenzelm@17423
  1361
  star_of_eq_0 <-- hcomplex_of_complex_zero_iff
wenzelm@17423
  1362
huffman@17442
  1363
* Hyperreal: new method "transfer" that implements the transfer
huffman@17442
  1364
principle of nonstandard analysis. With a subgoal that mentions
huffman@17442
  1365
nonstandard types like "'a star", the command "apply transfer"
huffman@17442
  1366
replaces it with an equivalent one that mentions only standard types.
huffman@17442
  1367
To be successful, all free variables must have standard types; non-
huffman@17442
  1368
standard variables must have explicit universal quantifiers.
huffman@17442
  1369
wenzelm@17641
  1370
* Hyperreal: A theory of Taylor series.
wenzelm@17641
  1371
wenzelm@14655
  1372
wenzelm@14682
  1373
*** HOLCF ***
wenzelm@14682
  1374
wenzelm@17533
  1375
* Discontinued special version of 'constdefs' (which used to support
wenzelm@17533
  1376
continuous functions) in favor of the general Pure one with full
wenzelm@17533
  1377
type-inference.
wenzelm@17533
  1378
wenzelm@17533
  1379
* New simplification procedure for solving continuity conditions; it
wenzelm@17533
  1380
is much faster on terms with many nested lambda abstractions (cubic
huffman@17442
  1381
instead of exponential time).
huffman@17442
  1382
wenzelm@17533
  1383
* New syntax for domain package: selector names are now optional.
huffman@17442
  1384
Parentheses should be omitted unless argument is lazy, for example:
huffman@17442
  1385
huffman@17442
  1386
  domain 'a stream = cons "'a" (lazy "'a stream")
huffman@17442
  1387
wenzelm@17533
  1388
* New command 'fixrec' for defining recursive functions with pattern
wenzelm@17533
  1389
matching; defining multiple functions with mutual recursion is also
wenzelm@17533
  1390
supported.  Patterns may include the constants cpair, spair, up, sinl,
wenzelm@17533
  1391
sinr, or any data constructor defined by the domain package. The given
wenzelm@17533
  1392
equations are proven as rewrite rules. See HOLCF/ex/Fixrec_ex.thy for
wenzelm@17533
  1393
syntax and examples.
wenzelm@17533
  1394
wenzelm@17533
  1395
* New commands 'cpodef' and 'pcpodef' for defining predicate subtypes
wenzelm@17533
  1396
of cpo and pcpo types. Syntax is exactly like the 'typedef' command,
wenzelm@17533
  1397
but the proof obligation additionally includes an admissibility
wenzelm@17533
  1398
requirement. The packages generate instances of class cpo or pcpo,
wenzelm@17533
  1399
with continuity and strictness theorems for Rep and Abs.
huffman@17442
  1400
huffman@17584
  1401
* HOLCF: Many theorems have been renamed according to a more standard naming
huffman@17584
  1402
scheme (INCOMPATIBILITY):
huffman@17584
  1403
huffman@17584
  1404
  foo_inject:  "foo$x = foo$y ==> x = y"
huffman@17584
  1405
  foo_eq:      "(foo$x = foo$y) = (x = y)"
huffman@17584
  1406
  foo_less:    "(foo$x << foo$y) = (x << y)"
huffman@17584
  1407
  foo_strict:  "foo$UU = UU"
huffman@17584
  1408
  foo_defined: "... ==> foo$x ~= UU"
huffman@17584
  1409
  foo_defined_iff: "(foo$x = UU) = (x = UU)"
huffman@17584
  1410
wenzelm@14682
  1411
paulson@14885
  1412
*** ZF ***
paulson@14885
  1413
wenzelm@16234
  1414
* ZF/ex: theories Group and Ring provide examples in abstract algebra,
wenzelm@16234
  1415
including the First Isomorphism Theorem (on quotienting by the kernel
wenzelm@16234
  1416
of a homomorphism).
wenzelm@15089
  1417
wenzelm@15089
  1418
* ZF/Simplifier: install second copy of type solver that actually
wenzelm@16234
  1419
makes use of TC rules declared to Isar proof contexts (or locales);
wenzelm@16234
  1420
the old version is still required for ML proof scripts.
wenzelm@15703
  1421
wenzelm@15703
  1422
wenzelm@17445
  1423
*** Cube ***
wenzelm@17445
  1424
wenzelm@17445
  1425
* Converted to Isar theory format; use locales instead of axiomatic
wenzelm@17445
  1426
theories.
wenzelm@17445
  1427
wenzelm@17445
  1428
wenzelm@15703
  1429
*** ML ***
wenzelm@15703
  1430
wenzelm@15973
  1431
* Pure/library.ML no longer defines its own option datatype, but uses
wenzelm@16234
  1432
that of the SML basis, which has constructors NONE and SOME instead of
wenzelm@16234
  1433
None and Some, as well as exception Option.Option instead of OPTION.
wenzelm@16234
  1434
The functions the, if_none, is_some, is_none have been adapted
wenzelm@16234
  1435
accordingly, while Option.map replaces apsome.
wenzelm@15973
  1436
wenzelm@16860
  1437
* Pure/library.ML: the exception LIST has been given up in favour of
wenzelm@16860
  1438
the standard exceptions Empty and Subscript, as well as
wenzelm@16860
  1439
Library.UnequalLengths.  Function like Library.hd and Library.tl are
wenzelm@16860
  1440
superceded by the standard hd and tl functions etc.
wenzelm@16860
  1441
wenzelm@16860
  1442
A number of basic list functions are no longer exported to the ML
wenzelm@16860
  1443
toplevel, as they are variants of predefined functions.  The following
wenzelm@16234
  1444
suggests how one can translate existing code:
wenzelm@15973
  1445
wenzelm@15973
  1446
    rev_append xs ys = List.revAppend (xs, ys)
wenzelm@15973
  1447
    nth_elem (i, xs) = List.nth (xs, i)
wenzelm@15973
  1448
    last_elem xs = List.last xs
wenzelm@15973
  1449
    flat xss = List.concat xss
wenzelm@16234
  1450
    seq fs = List.app fs
wenzelm@15973
  1451
    partition P xs = List.partition P xs
wenzelm@15973
  1452
    mapfilter f xs = List.mapPartial f xs
wenzelm@15973
  1453
wenzelm@16860
  1454
* Pure/library.ML: several combinators for linear functional
wenzelm@16860
  1455
transformations, notably reverse application and composition:
wenzelm@16860
  1456
wenzelm@17371
  1457
  x |> f                f #> g
wenzelm@17371
  1458
  (x, y) |-> f          f #-> g
wenzelm@16860
  1459
haftmann@17495
  1460
* Pure/library.ML: introduced/changed precedence of infix operators:
haftmann@17495
  1461
haftmann@17495
  1462
  infix 1 |> |-> ||> ||>> |>> |>>> #> #->;
haftmann@17495
  1463
  infix 2 ?;
haftmann@17495
  1464
  infix 3 o oo ooo oooo;
haftmann@17495
  1465
  infix 4 ~~ upto downto;
haftmann@17495
  1466
haftmann@17495
  1467
Maybe INCOMPATIBILITY when any of those is used in conjunction with other
haftmann@17495
  1468
infix operators.
haftmann@17495
  1469
wenzelm@17408
  1470
* Pure/library.ML: natural list combinators fold, fold_rev, and
haftmann@16869
  1471
fold_map support linear functional transformations and nesting.  For
wenzelm@16860
  1472
example:
wenzelm@16860
  1473
wenzelm@16860
  1474
  fold f [x1, ..., xN] y =
wenzelm@16860
  1475
    y |> f x1 |> ... |> f xN
wenzelm@16860
  1476
wenzelm@16860
  1477
  (fold o fold) f [xs1, ..., xsN] y =
wenzelm@16860
  1478
    y |> fold f xs1 |> ... |> fold f xsN
wenzelm@16860
  1479
wenzelm@16860
  1480
  fold f [x1, ..., xN] =
wenzelm@16860
  1481
    f x1 #> ... #> f xN
wenzelm@16860
  1482
wenzelm@16860
  1483
  (fold o fold) f [xs1, ..., xsN] =
wenzelm@16860
  1484
    fold f xs1 #> ... #> fold f xsN
wenzelm@16860
  1485
wenzelm@17408
  1486
* Pure/library.ML: the following selectors on type 'a option are
wenzelm@17408
  1487
available:
wenzelm@17408
  1488
wenzelm@17408
  1489
  the:               'a option -> 'a  (*partial*)
wenzelm@17408
  1490
  these:             'a option -> 'a  where 'a = 'b list
haftmann@17402
  1491
  the_default: 'a -> 'a option -> 'a
haftmann@17402
  1492
  the_list:          'a option -> 'a list
haftmann@17402
  1493
wenzelm@17408
  1494
* Pure/General: structure AList (cf. Pure/General/alist.ML) provides
wenzelm@17408
  1495
basic operations for association lists, following natural argument
haftmann@17564
  1496
order; moreover the explicit equality predicate passed here avoids
haftmann@17495
  1497
potentially expensive polymorphic runtime equality checks.
haftmann@17495
  1498
The old functions may be expressed as follows:
wenzelm@17408
  1499
wenzelm@17408
  1500
  assoc = uncurry (AList.lookup (op =))
wenzelm@17408
  1501
  assocs = these oo AList.lookup (op =)
wenzelm@17408
  1502
  overwrite = uncurry (AList.update (op =)) o swap
wenzelm@17408
  1503
haftmann@17564
  1504
* Pure/General: structure AList (cf. Pure/General/alist.ML) provides
haftmann@17564
  1505
haftmann@17564
  1506
  val make: ('a -> 'b) -> 'a list -> ('a * 'b) list
haftmann@17564
  1507
  val find: ('a * 'b -> bool) -> ('c * 'b) list -> 'a -> 'c list
haftmann@17564
  1508
haftmann@17564
  1509
replacing make_keylist and keyfilter (occassionally used)
haftmann@17564
  1510
Naive rewrites:
haftmann@17564
  1511
haftmann@17564
  1512
  make_keylist = AList.make
haftmann@17564
  1513
  keyfilter = AList.find (op =)
haftmann@17564
  1514
haftmann@17564
  1515
* eq_fst and eq_snd now take explicit equality parameter, thus
haftmann@17564
  1516
  avoiding eqtypes. Naive rewrites:
haftmann@17564
  1517
haftmann@17564
  1518
    eq_fst = eq_fst (op =)
haftmann@17564
  1519
    eq_snd = eq_snd (op =)
haftmann@17564
  1520
haftmann@17564
  1521
* Removed deprecated apl and apr (rarely used).
haftmann@17564
  1522
  Naive rewrites:
haftmann@17564
  1523
haftmann@17564
  1524
    apl (n, op) =>>= curry op n
haftmann@17564
  1525
    apr (op, m) =>>= fn n => op (n, m)
haftmann@17564
  1526
wenzelm@17408
  1527
* Pure/General: structure OrdList (cf. Pure/General/ord_list.ML)
wenzelm@17408
  1528
provides a reasonably efficient light-weight implementation of sets as
wenzelm@17408
  1529
lists.
wenzelm@17408
  1530
wenzelm@17408
  1531
* Pure/General: generic tables (cf. Pure/General/table.ML) provide a
wenzelm@17408
  1532
few new operations; existing lookup and update are now curried to
wenzelm@17408
  1533
follow natural argument order (for use with fold etc.);
wenzelm@17408
  1534
INCOMPATIBILITY, use (uncurry Symtab.lookup) etc. as last resort.
wenzelm@17408
  1535
wenzelm@17408
  1536
* Pure/General: output via the Isabelle channels of
wenzelm@17408
  1537
writeln/warning/error etc. is now passed through Output.output, with a
wenzelm@17408
  1538
hook for arbitrary transformations depending on the print_mode
wenzelm@17408
  1539
(cf. Output.add_mode -- the first active mode that provides a output
wenzelm@17408
  1540
function wins).  Already formatted output may be embedded into further
wenzelm@17408
  1541
text via Output.raw; the result of Pretty.string_of/str_of and derived
wenzelm@17408
  1542
functions (string_of_term/cterm/thm etc.) is already marked raw to
wenzelm@17408
  1543
accommodate easy composition of diagnostic messages etc.  Programmers
wenzelm@17408
  1544
rarely need to care about Output.output or Output.raw at all, with
wenzelm@17408
  1545
some notable exceptions: Output.output is required when bypassing the
wenzelm@17408
  1546
standard channels (writeln etc.), or in token translations to produce
wenzelm@17408
  1547
properly formatted results; Output.raw is required when capturing
wenzelm@17408
  1548
already output material that will eventually be presented to the user
wenzelm@17408
  1549
a second time.  For the default print mode, both Output.output and
wenzelm@17408
  1550
Output.raw have no effect.
wenzelm@17408
  1551
wenzelm@17408
  1552
* Pure/General: Output.time_accumulator NAME creates an operator ('a
wenzelm@17408
  1553
-> 'b) -> 'a -> 'b to measure runtime and count invocations; the
wenzelm@17408
  1554
cumulative results are displayed at the end of a batch session.
wenzelm@17408
  1555
wenzelm@17408
  1556
* Pure/General: File.sysify_path and File.quote_sysify path have been
wenzelm@17408
  1557
replaced by File.platform_path and File.shell_path (with appropriate
wenzelm@17408
  1558
hooks).  This provides a clean interface for unusual systems where the
wenzelm@17408
  1559
internal and external process view of file names are different.
wenzelm@17408
  1560
wenzelm@16689
  1561
* Pure: more efficient orders for basic syntactic entities: added
wenzelm@16689
  1562
fast_string_ord, fast_indexname_ord, fast_term_ord; changed sort_ord
wenzelm@16689
  1563
and typ_ord to use fast_string_ord and fast_indexname_ord (term_ord is
wenzelm@16689
  1564
NOT affected); structures Symtab, Vartab, Typtab, Termtab use the fast
wenzelm@16689
  1565
orders now -- potential INCOMPATIBILITY for code that depends on a
wenzelm@16689
  1566
particular order for Symtab.keys, Symtab.dest, etc. (consider using
wenzelm@16689
  1567
Library.sort_strings on result).
wenzelm@16689
  1568
wenzelm@17408
  1569
* Pure/term.ML: combinators fold_atyps, fold_aterms, fold_term_types,
wenzelm@17408
  1570
fold_types traverse types/terms from left to right, observing natural
wenzelm@17408
  1571
argument order.  Supercedes previous foldl_XXX versions, add_frees,
wenzelm@17408
  1572
add_vars etc. have been adapted as well: INCOMPATIBILITY.
wenzelm@17408
  1573
wenzelm@16151
  1574
* Pure: name spaces have been refined, with significant changes of the
wenzelm@16234
  1575
internal interfaces -- INCOMPATIBILITY.  Renamed cond_extern(_table)
wenzelm@16234
  1576
to extern(_table).  The plain name entry path is superceded by a
wenzelm@16234
  1577
general 'naming' context, which also includes the 'policy' to produce
wenzelm@16234
  1578
a fully qualified name and external accesses of a fully qualified
wenzelm@16234
  1579
name; NameSpace.extend is superceded by context dependent
wenzelm@16234
  1580
Sign.declare_name.  Several theory and proof context operations modify
wenzelm@16234
  1581
the naming context.  Especially note Theory.restore_naming and
wenzelm@16234
  1582
ProofContext.restore_naming to get back to a sane state; note that
wenzelm@16234
  1583
Theory.add_path is no longer sufficient to recover from
wenzelm@16234
  1584
Theory.absolute_path in particular.
wenzelm@16234
  1585
wenzelm@16234
  1586
* Pure: new flags short_names (default false) and unique_names
wenzelm@16234
  1587
(default true) for controlling output of qualified names.  If
wenzelm@16234
  1588
short_names is set, names are printed unqualified.  If unique_names is
wenzelm@16234
  1589
reset, the name prefix is reduced to the minimum required to achieve
wenzelm@16234
  1590
the original result when interning again, even if there is an overlap
wenzelm@16234
  1591
with earlier declarations.
wenzelm@16151
  1592
wenzelm@16456
  1593
* Pure/TheoryDataFun: change of the argument structure; 'prep_ext' is
wenzelm@16456
  1594
now 'extend', and 'merge' gets an additional Pretty.pp argument
wenzelm@16456
  1595
(useful for printing error messages).  INCOMPATIBILITY.
wenzelm@16456
  1596
wenzelm@16456
  1597
* Pure: major reorganization of the theory context.  Type Sign.sg and
wenzelm@16456
  1598
Theory.theory are now identified, referring to the universal
wenzelm@16456
  1599
Context.theory (see Pure/context.ML).  Actual signature and theory
wenzelm@16456
  1600
content is managed as theory data.  The old code and interfaces were
wenzelm@16456
  1601
spread over many files and structures; the new arrangement introduces
wenzelm@16456
  1602
considerable INCOMPATIBILITY to gain more clarity:
wenzelm@16456
  1603
wenzelm@16456
  1604
  Context -- theory management operations (name, identity, inclusion,
wenzelm@16456
  1605
    parents, ancestors, merge, etc.), plus generic theory data;
wenzelm@16456
  1606
wenzelm@16456
  1607
  Sign -- logical signature and syntax operations (declaring consts,
wenzelm@16456
  1608
    types, etc.), plus certify/read for common entities;
wenzelm@16456
  1609
wenzelm@16456
  1610
  Theory -- logical theory operations (stating axioms, definitions,
wenzelm@16456
  1611
    oracles), plus a copy of logical signature operations (consts,
wenzelm@16456
  1612
    types, etc.); also a few basic management operations (Theory.copy,
wenzelm@16456
  1613
    Theory.merge, etc.)
wenzelm@16456
  1614
wenzelm@16456
  1615
The most basic sign_of operations (Theory.sign_of, Thm.sign_of_thm
wenzelm@16456
  1616
etc.) as well as the sign field in Thm.rep_thm etc. have been retained
wenzelm@16456
  1617
for convenience -- they merely return the theory.
wenzelm@16456
  1618
wenzelm@17193
  1619
* Pure: type Type.tsig is superceded by theory in most interfaces.
wenzelm@17193
  1620
wenzelm@16547
  1621
* Pure: the Isar proof context type is already defined early in Pure
wenzelm@16547
  1622
as Context.proof (note that ProofContext.context and Proof.context are
wenzelm@16547
  1623
aliases, where the latter is the preferred name).  This enables other
wenzelm@16547
  1624
Isabelle components to refer to that type even before Isar is present.
wenzelm@16547
  1625
wenzelm@16373
  1626
* Pure/sign/theory: discontinued named name spaces (i.e. classK,
wenzelm@16373
  1627
typeK, constK, axiomK, oracleK), but provide explicit operations for
wenzelm@16373
  1628
any of these kinds.  For example, Sign.intern typeK is now
wenzelm@16373
  1629
Sign.intern_type, Theory.hide_space Sign.typeK is now
wenzelm@16373
  1630
Theory.hide_types.  Also note that former
wenzelm@16373
  1631
Theory.hide_classes/types/consts are now
wenzelm@16373
  1632
Theory.hide_classes_i/types_i/consts_i, while the non '_i' versions
wenzelm@16373
  1633
internalize their arguments!  INCOMPATIBILITY.
wenzelm@16373
  1634
wenzelm@16506
  1635
* Pure: get_thm interface (of PureThy and ProofContext) expects
wenzelm@16506
  1636
datatype thmref (with constructors Name and NameSelection) instead of
wenzelm@16506
  1637
plain string -- INCOMPATIBILITY;
wenzelm@16506
  1638
wenzelm@16151
  1639
* Pure: cases produced by proof methods specify options, where NONE
wenzelm@16234
  1640
means to remove case bindings -- INCOMPATIBILITY in
wenzelm@16234
  1641
(RAW_)METHOD_CASES.
wenzelm@16151
  1642
wenzelm@16373
  1643
* Pure: the following operations retrieve axioms or theorems from a
wenzelm@16373
  1644
theory node or theory hierarchy, respectively:
wenzelm@16373
  1645
wenzelm@16373
  1646
  Theory.axioms_of: theory -> (string * term) list
wenzelm@16373
  1647
  Theory.all_axioms_of: theory -> (string * term) list
wenzelm@16373
  1648
  PureThy.thms_of: theory -> (string * thm) list
wenzelm@16373
  1649
  PureThy.all_thms_of: theory -> (string * thm) list
wenzelm@16373
  1650
wenzelm@16718
  1651
* Pure: print_tac now outputs the goal through the trace channel.
wenzelm@16718
  1652
wenzelm@17408
  1653
* Isar toplevel: improved diagnostics, mostly for Poly/ML only.
wenzelm@17408
  1654
Reference Toplevel.debug (default false) controls detailed printing
wenzelm@17408
  1655
and tracing of low-level exceptions; Toplevel.profiling (default 0)
wenzelm@17408
  1656
controls execution profiling -- set to 1 for time and 2 for space
wenzelm@17408
  1657
(both increase the runtime).
wenzelm@17408
  1658
wenzelm@17408
  1659
* Isar session: The initial use of ROOT.ML is now always timed,
wenzelm@17408
  1660
i.e. the log will show the actual process times, in contrast to the
wenzelm@17408
  1661
elapsed wall-clock time that the outer shell wrapper produces.
wenzelm@17408
  1662
wenzelm@17408
  1663
* Simplifier: improved handling of bound variables (nameless
wenzelm@16997
  1664
representation, avoid allocating new strings).  Simprocs that invoke
wenzelm@16997
  1665
the Simplifier recursively should use Simplifier.inherit_bounds to
wenzelm@17720
  1666
avoid local name clashes.  Failure to do so produces warnings
wenzelm@17720
  1667
"Simplifier: renamed bound variable ..."; set Simplifier.debug_bounds
wenzelm@17720
  1668
for further details.
wenzelm@16234
  1669
wenzelm@17166
  1670
* ML functions legacy_bindings and use_legacy_bindings produce ML fact
wenzelm@17166
  1671
bindings for all theorems stored within a given theory; this may help
wenzelm@17166
  1672
in porting non-Isar theories to Isar ones, while keeping ML proof
wenzelm@17166
  1673
scripts for the time being.
wenzelm@17166
  1674
wenzelm@17457
  1675
* ML operator HTML.with_charset specifies the charset begin used for
wenzelm@17457
  1676
generated HTML files.  For example:
wenzelm@17457
  1677
wenzelm@17457
  1678
  HTML.with_charset "utf-8" use_thy "Hebrew";
wenzelm@17538
  1679
  HTML.with_charset "utf-8" use_thy "Chinese";
wenzelm@17457
  1680
wenzelm@16234
  1681
wenzelm@16234
  1682
*** System ***
wenzelm@16234
  1683
wenzelm@16234
  1684
* Allow symlinks to all proper Isabelle executables (Isabelle,
wenzelm@16234
  1685
isabelle, isatool etc.).
wenzelm@16234
  1686
wenzelm@16234
  1687
* ISABELLE_DOC_FORMAT setting specifies preferred document format (for
wenzelm@16234
  1688
isatool doc, isatool mkdir, display_drafts etc.).
wenzelm@16234
  1689
wenzelm@16234
  1690
* isatool usedir: option -f allows specification of the ML file to be
wenzelm@16234
  1691
used by Isabelle; default is ROOT.ML.
wenzelm@16234
  1692
wenzelm@16251
  1693
* New isatool version outputs the version identifier of the Isabelle
wenzelm@16251
  1694
distribution being used.
wenzelm@16251
  1695
wenzelm@16251
  1696
* HOL: new isatool dimacs2hol converts files in DIMACS CNF format
wenzelm@16234
  1697
(containing Boolean satisfiability problems) into Isabelle/HOL
wenzelm@16234
  1698
theories.
wenzelm@15703
  1699
wenzelm@15703
  1700
wenzelm@14655
  1701
wenzelm@14606
  1702
New in Isabelle2004 (April 2004)
wenzelm@14606
  1703
--------------------------------
wenzelm@13280
  1704
skalberg@14171
  1705
*** General ***
skalberg@14171
  1706
ballarin@14398
  1707
* Provers/order.ML:  new efficient reasoner for partial and linear orders.
ballarin@14398
  1708
  Replaces linorder.ML.
ballarin@14398
  1709
wenzelm@14606
  1710
* Pure: Greek letters (except small lambda, \<lambda>), as well as Gothic
wenzelm@14606
  1711
  (\<aa>...\<zz>\<AA>...\<ZZ>), calligraphic (\<A>...\<Z>), and Euler
skalberg@14173
  1712
  (\<a>...\<z>), are now considered normal letters, and can therefore
skalberg@14173
  1713
  be used anywhere where an ASCII letter (a...zA...Z) has until
skalberg@14173
  1714
  now. COMPATIBILITY: This obviously changes the parsing of some
skalberg@14173
  1715
  terms, especially where a symbol has been used as a binder, say
skalberg@14173
  1716
  '\<Pi>x. ...', which is now a type error since \<Pi>x will be parsed
skalberg@14173
  1717
  as an identifier.  Fix it by inserting a space around former
skalberg@14173
  1718
  symbols.  Call 'isatool fixgreek' to try to fix parsing errors in
skalberg@14173
  1719
  existing theory and ML files.
skalberg@14171
  1720
paulson@14237
  1721
* Pure: Macintosh and Windows line-breaks are now allowed in theory files.
paulson@14237
  1722
wenzelm@14731
  1723
* Pure: single letter sub/superscripts (\<^isub> and \<^isup>) are now
wenzelm@14731
  1724
  allowed in identifiers. Similar to Greek letters \<^isub> is now considered
wenzelm@14731
  1725
  a normal (but invisible) letter. For multiple letter subscripts repeat
wenzelm@14731
  1726
  \<^isub> like this: x\<^isub>1\<^isub>2.
kleing@14233
  1727
kleing@14333
  1728
* Pure: There are now sub-/superscripts that can span more than one
kleing@14333
  1729
  character. Text between \<^bsub> and \<^esub> is set in subscript in
wenzelm@14606
  1730
  ProofGeneral and LaTeX, text between \<^bsup> and \<^esup> in
wenzelm@14606
  1731
  superscript. The new control characters are not identifier parts.
kleing@14333
  1732
schirmer@14561
  1733
* Pure: Control-symbols of the form \<^raw:...> will literally print the
wenzelm@14606
  1734
  content of "..." to the latex file instead of \isacntrl... . The "..."
wenzelm@14606
  1735
  may consist of any printable characters excluding the end bracket >.
schirmer@14361
  1736
paulson@14237
  1737
* Pure: Using new Isar command "finalconsts" (or the ML functions
paulson@14237
  1738
  Theory.add_finals or Theory.add_finals_i) it is now possible to
paulson@14237
  1739
  declare constants "final", which prevents their being given a definition
paulson@14237
  1740
  later.  It is useful for constants whose behaviour is fixed axiomatically
skalberg@14224
  1741
  rather than definitionally, such as the meta-logic connectives.
skalberg@14224
  1742
wenzelm@14606
  1743
* Pure: 'instance' now handles general arities with general sorts
wenzelm@14606
  1744
  (i.e. intersections of classes),
skalberg@14503
  1745
kleing@14547
  1746
* Presentation: generated HTML now uses a CSS style sheet to make layout
wenzelm@14731
  1747
  (somewhat) independent of content. It is copied from lib/html/isabelle.css.
kleing@14547
  1748
  It can be changed to alter the colors/layout of generated pages.
kleing@14547
  1749
wenzelm@14556
  1750
ballarin@14175
  1751
*** Isar ***
ballarin@14175
  1752
ballarin@14508
  1753
* Tactic emulation methods rule_tac, erule_tac, drule_tac, frule_tac,
ballarin@14508
  1754
  cut_tac, subgoal_tac and thin_tac:
ballarin@14175
  1755
  - Now understand static (Isar) contexts.  As a consequence, users of Isar
ballarin@14175
  1756
    locales are no longer forced to write Isar proof scripts.
ballarin@14175
  1757
    For details see Isar Reference Manual, paragraph 4.3.2: Further tactic
ballarin@14175
  1758
    emulations.
ballarin@14175
  1759
  - INCOMPATIBILITY: names of variables to be instantiated may no
ballarin@14211
  1760
    longer be enclosed in quotes.  Instead, precede variable name with `?'.
ballarin@14211
  1761
    This is consistent with the instantiation attribute "where".
ballarin@14211
  1762
ballarin@14257
  1763
* Attributes "where" and "of":
ballarin@14285
  1764
  - Now take type variables of instantiated theorem into account when reading
ballarin@14285
  1765
    the instantiation string.  This fixes a bug that caused instantiated
ballarin@14285
  1766
    theorems to have too special types in some circumstances.
ballarin@14285
  1767
  - "where" permits explicit instantiations of type variables.
ballarin@14257
  1768
wenzelm@14556
  1769
* Calculation commands "moreover" and "also" no longer interfere with
wenzelm@14556
  1770
  current facts ("this"), admitting arbitrary combinations with "then"
wenzelm@14556
  1771
  and derived forms.
kleing@14283
  1772
ballarin@14211
  1773
* Locales:
ballarin@14211
  1774
  - Goal statements involving the context element "includes" no longer
ballarin@14211
  1775
    generate theorems with internal delta predicates (those ending on
ballarin@14211
  1776
    "_axioms") in the premise.
ballarin@14211
  1777
    Resolve particular premise with <locale>.intro to obtain old form.
ballarin@14211
  1778
  - Fixed bug in type inference ("unify_frozen") that prevented mix of target
ballarin@14211
  1779
    specification and "includes" elements in goal statement.
ballarin@14254
  1780
  - Rule sets <locale>.intro and <locale>.axioms no longer declared as
ballarin@14254
  1781
    [intro?] and [elim?] (respectively) by default.
ballarin@14508
  1782
  - Experimental command for instantiation of locales in proof contexts:
ballarin@14551
  1783
        instantiate <label>[<attrs>]: <loc>
ballarin@14508
  1784
    Instantiates locale <loc> and adds all its theorems to the current context
ballarin@14551
  1785
    taking into account their attributes.  Label and attrs are optional
ballarin@14551
  1786
    modifiers, like in theorem declarations.  If present, names of
ballarin@14551
  1787
    instantiated theorems are qualified with <label>, and the attributes
ballarin@14551
  1788
    <attrs> are applied after any attributes these theorems might have already.
ballarin@14551
  1789
      If the locale has assumptions, a chained fact of the form
ballarin@14508
  1790
    "<loc> t1 ... tn" is expected from which instantiations of the parameters
ballarin@14551
  1791
    are derived.  The command does not support old-style locales declared
ballarin@14551
  1792
    with "locale (open)".
ballarin@14551
  1793
      A few (very simple) examples can be found in FOL/ex/LocaleInst.thy.
ballarin@14175
  1794
ballarin@14175
  1795
* HOL: Tactic emulation methods induct_tac and case_tac understand static
ballarin@14175
  1796
  (Isar) contexts.
ballarin@14175
  1797
wenzelm@14556
  1798
kleing@14136
  1799
*** HOL ***
kleing@14136
  1800
kleing@14624
  1801
* Proof import: new image HOL4 contains the imported library from
kleing@14624
  1802
  the HOL4 system with about 2500 theorems. It is imported by
kleing@14624
  1803
  replaying proof terms produced by HOL4 in Isabelle. The HOL4 image
kleing@14624
  1804
  can be used like any other Isabelle image.  See
kleing@14624
  1805
  HOL/Import/HOL/README for more information.
kleing@14624
  1806
ballarin@14398
  1807
* Simplifier:
ballarin@14398
  1808
  - Much improved handling of linear and partial orders.
ballarin@14398
  1809
    Reasoners for linear and partial orders are set up for type classes
ballarin@14398
  1810
    "linorder" and "order" respectively, and are added to the default simpset
ballarin@14398
  1811
    as solvers.  This means that the simplifier can build transitivity chains
ballarin@14398
  1812
    to solve goals from the assumptions.
ballarin@14398
  1813
  - INCOMPATIBILITY: old proofs break occasionally.  Typically, applications
ballarin@14398
  1814
    of blast or auto after simplification become unnecessary because the goal
ballarin@14398
  1815
    is solved by simplification already.
ballarin@14398
  1816
wenzelm@14731
  1817
* Numerics: new theory Ring_and_Field contains over 250 basic numerical laws,
paulson@14389
  1818
    all proved in axiomatic type classes for semirings, rings and fields.
paulson@14389
  1819
paulson@14389
  1820
* Numerics:
paulson@14389
  1821
  - Numeric types (nat, int, and in HOL-Complex rat, real, complex, etc.) are
wenzelm@14731
  1822
    now formalized using the Ring_and_Field theory mentioned above.
paulson@14389
  1823
  - INCOMPATIBILITY: simplification and arithmetic behaves somewhat differently
paulson@14389
  1824
    than before, because now they are set up once in a generic manner.
wenzelm@14731
  1825
  - INCOMPATIBILITY: many type-specific arithmetic laws have gone.
paulson@14480
  1826
    Look for the general versions in Ring_and_Field (and Power if they concern
paulson@14480
  1827
    exponentiation).
paulson@14389
  1828
paulson@14401
  1829
* Type "rat" of the rational numbers is now available in HOL-Complex.
paulson@14389
  1830
schirmer@14255
  1831
* Records:
schirmer@14255
  1832
  - Record types are now by default printed with their type abbreviation
schirmer@14255
  1833
    instead of the list of all field types. This can be configured via
schirmer@14255
  1834
    the reference "print_record_type_abbr".
wenzelm@14731
  1835
  - Simproc "record_upd_simproc" for simplification of multiple updates added
schirmer@14255
  1836
    (not enabled by default).
schirmer@14427
  1837
  - Simproc "record_ex_sel_eq_simproc" to simplify EX x. sel r = x resp.
schirmer@14427
  1838
    EX x. x = sel r to True (not enabled by default).
schirmer@14255
  1839
  - Tactic "record_split_simp_tac" to split and simplify records added.
wenzelm@14731
  1840
kleing@14136
  1841
* 'specification' command added, allowing for definition by
skalberg@14224
  1842
  specification.  There is also an 'ax_specification' command that
skalberg@14224
  1843
  introduces the new constants axiomatically.
kleing@14136
  1844
nipkow@14375
  1845
* arith(_tac) is now able to generate counterexamples for reals as well.
nipkow@14375
  1846
ballarin@14399
  1847
* HOL-Algebra: new locale "ring" for non-commutative rings.
ballarin@14399
  1848
paulson@14243
  1849
* HOL-ex: InductiveInvariant_examples illustrates advanced recursive function
kleing@14610
  1850
  definitions, thanks to Sava Krsti\'{c} and John Matthews.
kleing@14610
  1851
wenzelm@14731
  1852
* HOL-Matrix: a first theory for matrices in HOL with an application of
kleing@14610
  1853
  matrix theory to linear programming.
kleing@14136
  1854
nipkow@14380
  1855
* Unions and Intersections:
nipkow@15119
  1856
  The latex output syntax of UN and INT has been changed
nipkow@15119
  1857
  from "\Union x \in A. B" to "\Union_{x \in A} B"
nipkow@15119
  1858
  i.e. the index formulae has become a subscript.
nipkow@15119
  1859
  Similarly for "\Union x. B", and for \Inter instead of \Union.
nipkow@14380
  1860
kleing@14418
  1861
* Unions and Intersections over Intervals:
wenzelm@14731
  1862
  There is new short syntax "UN i<=n. A" for "UN i:{0..n}. A". There is
wenzelm@14731
  1863
  also an x-symbol version with subscripts "\<Union>\<^bsub>i <= n\<^esub>. A"
kleing@14418
  1864
  like in normal math, and corresponding versions for < and for intersection.
kleing@14418
  1865
nipkow@15677
  1866
* HOL/List: Ordering "lexico" is renamed "lenlex" and the standard
nipkow@15677
  1867
  lexicographic dictonary ordering has been added as "lexord".
nipkow@15677
  1868
paulson@14401
  1869
* ML: the legacy theory structures Int and List have been removed. They had
paulson@14401
  1870
  conflicted with ML Basis Library structures having the same names.
nipkow@14380
  1871
webertj@14464
  1872
* 'refute' command added to search for (finite) countermodels.  Only works
webertj@14464
  1873
  for a fragment of HOL.  The installation of an external SAT solver is
webertj@14464
  1874
  highly recommended.  See "HOL/Refute.thy" for details.
webertj@14464
  1875
berghofe@14602
  1876
* 'quickcheck' command: Allows to find counterexamples by evaluating
berghofe@14602
  1877
  formulae under an assignment of free variables to random values.
berghofe@14602
  1878
  In contrast to 'refute', it can deal with inductive datatypes,
berghofe@14602
  1879
  but cannot handle quantifiers. See "HOL/ex/Quickcheck_Examples.thy"
berghofe@14602
  1880
  for examples.
webertj@14464
  1881
wenzelm@14606
  1882
oheimb@14536
  1883
*** HOLCF ***
oheimb@14536
  1884
oheimb@14536
  1885
* Streams now come with concatenation and are part of the HOLCF image
oheimb@14536
  1886
wenzelm@14572
  1887
wenzelm@14572
  1888
kleing@14136
  1889
New in Isabelle2003 (May 2003)
wenzelm@14606
  1890
------------------------------
kleing@14136
  1891
wenzelm@13280
  1892
*** General ***
wenzelm@13280
  1893
berghofe@13618
  1894
* Provers/simplifier:
berghofe@13618
  1895
nipkow@13781
  1896
  - Completely reimplemented method simp (ML: Asm_full_simp_tac):
berghofe@13618
  1897
    Assumptions are now subject to complete mutual simplification,
berghofe@13618
  1898
    not just from left to right. The simplifier now preserves
berghofe@13618
  1899
    the order of assumptions.
berghofe@13618
  1900
berghofe@13618
  1901
    Potential INCOMPATIBILITY:
berghofe@13618
  1902
nipkow@13781
  1903
    -- simp sometimes diverges where the old version did
nipkow@13781
  1904
       not, e.g. invoking simp on the goal
berghofe@13618
  1905
berghofe@13618
  1906
        [| P (f x); y = x; f x = f y |] ==> Q
berghofe@13618
  1907
nipkow@13781
  1908
       now gives rise to the infinite reduction sequence
nipkow@13781
  1909
nipkow@13781
  1910
        P(f x) --(f x = f y)--> P(f y) --(y = x)--> P(f x) --(f x = f y)--> ...
nipkow@13781
  1911
nipkow@13781
  1912
       Using "simp (asm_lr)" (ML: Asm_lr_simp_tac) instead often solves this
nipkow@13781
  1913
       kind of problem.
nipkow@13781
  1914
nipkow@13781
  1915
    -- Tactics combining classical reasoner and simplification (such as auto)
nipkow@13781
  1916
       are also affected by this change, because many of them rely on
nipkow@13781
  1917
       simp. They may sometimes diverge as well or yield a different numbers
nipkow@13781
  1918
       of subgoals. Try to use e.g. force, fastsimp, or safe instead of auto
nipkow@13781
  1919
       in case of problems. Sometimes subsequent calls to the classical
nipkow@13781
  1920
       reasoner will fail because a preceeding call to the simplifier too
nipkow@13781
  1921
       eagerly simplified the goal, e.g. deleted redundant premises.
berghofe@13618
  1922
berghofe@13618
  1923
  - The simplifier trace now shows the names of the applied rewrite rules
berghofe@13618
  1924
nipkow@13829
  1925
  - You can limit the number of recursive invocations of the simplifier
nipkow@13829
  1926
    during conditional rewriting (where the simplifie tries to solve the
nipkow@13829
  1927
    conditions before applying the rewrite rule):
nipkow@13829
  1928
    ML "simp_depth_limit := n"
nipkow@13829
  1929
    where n is an integer. Thus you can force termination where previously
nipkow@13829
  1930
    the simplifier would diverge.
nipkow@13829
  1931
ballarin@13835
  1932
  - Accepts free variables as head terms in congruence rules.  Useful in Isar.
nipkow@13829
  1933
ballarin@13938
  1934
  - No longer aborts on failed congruence proof.  Instead, the
ballarin@13938
  1935
    congruence is ignored.
ballarin@13938
  1936
berghofe@14008
  1937
* Pure: New generic framework for extracting programs from constructive
berghofe@14008
  1938
  proofs. See HOL/Extraction.thy for an example instantiation, as well
berghofe@14008
  1939
  as HOL/Extraction for some case studies.
berghofe@14008
  1940
nipkow@13868
  1941
* Pure: The main goal of the proof state is no longer shown by default, only
nipkow@13868
  1942
the subgoals. This behaviour is controlled by a new flag.
ballarin@13835
  1943
   PG menu: Isabelle/Isar -> Settings -> Show Main Goal
nipkow@13815
  1944
(ML: Proof.show_main_goal).
nipkow@13815
  1945
nipkow@13815
  1946
* Pure: You can find all matching introduction rules for subgoal 1, i.e. all
nipkow@13815
  1947
rules whose conclusion matches subgoal 1:
nipkow@13815
  1948
      PG menu: Isabelle/Isar -> Show me -> matching rules
nipkow@13815
  1949
The rules are ordered by how closely they match the subgoal.
nipkow@13815
  1950
In particular, rules that solve a subgoal outright are displayed first
nipkow@13815
  1951
(or rather last, the way they are printed).
nipkow@13815
  1952
(ML: ProofGeneral.print_intros())
nipkow@13815
  1953
nipkow@13815
  1954
* Pure: New flag trace_unify_fail causes unification to print
nipkow@13781
  1955
diagnostic information (PG: in trace buffer) when it fails. This is
nipkow@13781
  1956
useful for figuring out why single step proofs like rule, erule or
nipkow@13781
  1957
assumption failed.
nipkow@13781
  1958
nipkow@13815
  1959
* Pure: Locale specifications now produce predicate definitions
wenzelm@13410
  1960
according to the body of text (covering assumptions modulo local
wenzelm@13410
  1961
definitions); predicate "loc_axioms" covers newly introduced text,
wenzelm@13410
  1962
while "loc" is cumulative wrt. all included locale expressions; the
wenzelm@13410
  1963
latter view is presented only on export into the global theory
wenzelm@13410
  1964
context; potential INCOMPATIBILITY, use "(open)" option to fall back
wenzelm@13410
  1965
on the old view without predicates;
wenzelm@13410
  1966
wenzelm@13459
  1967
* Pure: predefined locales "var" and "struct" are useful for sharing
wenzelm@13459
  1968
parameters (as in CASL, for example); just specify something like
wenzelm@13459
  1969
``var x + var y + struct M'' as import;
wenzelm@13459
  1970
wenzelm@13463
  1971
* Pure: improved thms_containing: proper indexing of facts instead of
wenzelm@13463
  1972
raw theorems; check validity of results wrt. current name space;
wenzelm@13463
  1973
include local facts of proof configuration (also covers active
wenzelm@13541
  1974
locales), cover fixed variables in index; may use "_" in term
wenzelm@13541
  1975
specification; an optional limit for the number of printed facts may
wenzelm@13541
  1976
be given (the default is 40);
wenzelm@13541
  1977
wenzelm@13541
  1978
* Pure: disallow duplicate fact bindings within new-style theory files
wenzelm@13541
  1979
(batch-mode only);
wenzelm@13540
  1980
wenzelm@13463
  1981
* Provers: improved induct method: assumptions introduced by case
wenzelm@13463
  1982
"foo" are split into "foo.hyps" (from the rule) and "foo.prems" (from
wenzelm@13463
  1983
the goal statement); "foo" still refers to all facts collectively;
wenzelm@13463
  1984
paulson@13550
  1985
* Provers: the function blast.overloaded has been removed: all constants
paulson@13550
  1986
are regarded as potentially overloaded, which improves robustness in exchange
paulson@13550
  1987
for slight decrease in efficiency;
paulson@13550
  1988
nipkow@13781
  1989
* Provers/linorder: New generic prover for transitivity reasoning over
nipkow@13781
  1990
linear orders.  Note: this prover is not efficient!
nipkow@13781
  1991
wenzelm@13522
  1992
* Isar: preview of problems to finish 'show' now produce an error
wenzelm@13522
  1993
rather than just a warning (in interactive mode);
wenzelm@13522
  1994
wenzelm@13280
  1995
nipkow@13158
  1996
*** HOL ***
nipkow@13158
  1997
nipkow@13899
  1998
* arith(_tac)
nipkow@13899
  1999
nipkow@13899
  2000
 - Produces a counter example if it cannot prove a goal.
nipkow@13899
  2001
   Note that the counter example may be spurious if the goal is not a formula
nipkow@13899
  2002
   of quantifier-free linear arithmetic.
nipkow@13899
  2003
   In ProofGeneral the counter example appears in the trace buffer.
nipkow@13899
  2004
nipkow@13899
  2005
 - Knows about div k and mod k where k is a numeral of type nat or int.
nipkow@13899
  2006
nipkow@13899
  2007
 - Calls full Presburger arithmetic (by Amine Chaieb) if quantifier-free
nipkow@13899
  2008
   linear arithmetic fails. This takes account of quantifiers and divisibility.
wenzelm@14731
  2009
   Presburger arithmetic can also be called explicitly via presburger(_tac).
nipkow@13899
  2010
nipkow@13899
  2011
* simp's arithmetic capabilities have been enhanced a bit: it now
nipkow@13899
  2012
takes ~= in premises into account (by performing a case split);
nipkow@13899
  2013
nipkow@13899
  2014
* simp reduces "m*(n div m) + n mod m" to n, even if the two summands
nipkow@13899
  2015
are distributed over a sum of terms;
nipkow@13899
  2016
ballarin@13735
  2017
* New tactic "trans_tac" and method "trans" instantiate
ballarin@13735
  2018
Provers/linorder.ML for axclasses "order" and "linorder" (predicates
wenzelm@14731
  2019
"<=", "<" and "=").
wenzelm@14731
  2020
wenzelm@14731
  2021
* function INCOMPATIBILITIES: Pi-sets have been redefined and moved from main
paulson@13587
  2022
HOL to Library/FuncSet; constant "Fun.op o" is now called "Fun.comp";
paulson@13587
  2023
wenzelm@13443
  2024
* 'typedef' command has new option "open" to suppress the set
wenzelm@13443
  2025
definition;
wenzelm@13443
  2026
wenzelm@13522
  2027
* functions Min and Max on finite sets have been introduced (theory
wenzelm@13522
  2028
Finite_Set);
nipkow@13492
  2029
wenzelm@13443
  2030
* attribute [symmetric] now works for relations as well; it turns
wenzelm@13443
  2031
(x,y) : R^-1 into (y,x) : R, and vice versa;
wenzelm@13443
  2032
nipkow@13613
  2033
* induct over a !!-quantified statement (say !!x1..xn):
nipkow@13613
  2034
  each "case" automatically performs "fix x1 .. xn" with exactly those names.
nipkow@13613
  2035
nipkow@13899
  2036
* Map: `empty' is no longer a constant but a syntactic abbreviation for
nipkow@13899
  2037
%x. None. Warning: empty_def now refers to the previously hidden definition
nipkow@13899
  2038
of the empty set.
nipkow@13899
  2039
ballarin@14018
  2040
* Algebra: formalization of classical algebra.  Intended as base for
ballarin@14018
  2041
any algebraic development in Isabelle.  Currently covers group theory
ballarin@14018
  2042
(up to Sylow's theorem) and ring theory (Universal Property of
ballarin@14018
  2043
Univariate Polynomials).  Contributions welcome;
paulson@13960
  2044
paulson@13960
  2045
* GroupTheory: deleted, since its material has been moved to Algebra;
paulson@13960
  2046
wenzelm@14731
  2047
* Complex: new directory of the complex numbers with numeric constants,
wenzelm@14731
  2048
nonstandard complex numbers, and some complex analysis, standard and
paulson@13966
  2049
nonstandard (Jacques Fleuriot);
paulson@13966
  2050
paulson@13966
  2051
* HOL-Complex: new image for analysis, replacing HOL-Real and HOL-Hyperreal;
paulson@13966
  2052
wenzelm@14731
  2053
* Hyperreal: introduced Gauge integration and hyperreal logarithms (Jacques
paulson@13966
  2054
Fleuriot);
paulson@13960
  2055
wenzelm@13549
  2056
* Real/HahnBanach: updated and adapted to locales;
wenzelm@13549
  2057
ballarin@13995
  2058
* NumberTheory: added Gauss's law of quadratic reciprocity (by Avigad,
ballarin@13995
  2059
Gray and Kramer);
paulson@13872
  2060
paulson@13872
  2061
* UNITY: added the Meier-Sanders theory of progress sets;
paulson@13872
  2062
kleing@14011
  2063
* MicroJava: bytecode verifier and lightweight bytecode verifier
kleing@14011
  2064
as abstract algorithms, instantiated to the JVM;
kleing@14011
  2065
schirmer@14010
  2066
* Bali: Java source language formalization. Type system, operational
schirmer@14010
  2067
semantics, axiomatic semantics. Supported language features:
schirmer@14010
  2068
classes, interfaces, objects,virtual methods, static methods,
schirmer@14010
  2069
static/instance fields, arrays, access modifiers, definite
schirmer@14010
  2070
assignment, exceptions.
wenzelm@13549
  2071
kleing@14011
  2072
wenzelm@13549
  2073
*** ZF ***
wenzelm@13549
  2074
webertj@15154
  2075
* ZF/Constructible: consistency proof for AC (Gdel's constructible
wenzelm@13549
  2076
universe, etc.);
wenzelm@13549
  2077
paulson@13872
  2078
* Main ZF: virtually all theories converted to new-style format;
nipkow@13518
  2079
wenzelm@13280
  2080
wenzelm@13478
  2081
*** ML ***
wenzelm@13478
  2082
wenzelm@13478
  2083
* Pure: Tactic.prove provides sane interface for internal proofs;
wenzelm@13478
  2084
omits the infamous "standard" operation, so this is more appropriate
wenzelm@13478
  2085
than prove_goalw_cterm in many situations (e.g. in simprocs);
wenzelm@13478
  2086
wenzelm@13478
  2087
* Pure: improved error reporting of simprocs;
wenzelm@13478
  2088
wenzelm@13478
  2089
* Provers: Simplifier.simproc(_i) provides sane interface for setting
wenzelm@13478
  2090
up simprocs;
wenzelm@13478
  2091
wenzelm@13478
  2092
kleing@13953
  2093
*** Document preparation ***
kleing@13953
  2094
kleing@13953
  2095
* uses \par instead of \\ for line breaks in theory text. This may
kleing@13953
  2096
shift some page breaks in large documents. To get the old behaviour
kleing@13953
  2097
use \renewcommand{\isanewline}{\mbox{}\\\mbox{}} in root.tex.
kleing@13953
  2098
wenzelm@14731
  2099
* minimized dependencies of isabelle.sty and isabellesym.sty on
kleing@13953
  2100
other packages
kleing@13953
  2101
kleing@13953
  2102
* \<euro> now needs package babel/greek instead of marvosym (which
kleing@13953
  2103
broke \Rightarrow)
kleing@13953
  2104
wenzelm@14731
  2105
* normal size for \<zero>...\<nine> (uses \mathbf instead of
kleing@13954
  2106
textcomp package)
kleing@13953
  2107
wenzelm@13280
  2108
wenzelm@14572
  2109
wenzelm@12984
  2110
New in Isabelle2002 (March 2002)
wenzelm@12984
  2111
--------------------------------
wenzelm@11474
  2112
wenzelm@11572
  2113
*** Document preparation ***
wenzelm@11572
  2114
wenzelm@11842
  2115
* greatly simplified document preparation setup, including more
wenzelm@11842
  2116
graceful interpretation of isatool usedir -i/-d/-D options, and more
wenzelm@11842
  2117
instructive isatool mkdir; users should basically be able to get
wenzelm@12899
  2118
started with "isatool mkdir HOL Test && isatool make"; alternatively,
wenzelm@12899
  2119
users may run a separate document processing stage manually like this:
wenzelm@12899
  2120
"isatool usedir -D output HOL Test && isatool document Test/output";
wenzelm@11842
  2121
wenzelm@11842
  2122
* theory dependency graph may now be incorporated into documents;
wenzelm@11842
  2123
isatool usedir -g true will produce session_graph.eps/.pdf for use
wenzelm@11842
  2124
with \includegraphics of LaTeX;
wenzelm@11842
  2125
wenzelm@11864
  2126
* proper spacing of consecutive markup elements, especially text
wenzelm@11864
  2127
blocks after section headings;
wenzelm@11864
  2128
wenzelm@11572
  2129
* support bold style (for single symbols only), input syntax is like
wenzelm@11572
  2130
this: "\<^bold>\<alpha>" or "\<^bold>A";
wenzelm@11572
  2131
wenzelm@11814
  2132
* \<bullet> is now output as bold \cdot by default, which looks much
wenzelm@11572
  2133
better in printed text;
wenzelm@11572
  2134
wenzelm@11712
  2135
* added default LaTeX bindings for \<tturnstile> and \<TTurnstile>;
wenzelm@11712
  2136
note that these symbols are currently unavailable in Proof General /
wenzelm@12769
  2137
X-Symbol; new symbols \<zero>, \<one>, ..., \<nine>, and \<euro>;
wenzelm@12690
  2138
wenzelm@12690
  2139
* isatool latex no longer depends on changed TEXINPUTS, instead
wenzelm@12690
  2140
isatool document copies the Isabelle style files to the target
wenzelm@12690
  2141
location;
wenzelm@11712
  2142
wenzelm@11572
  2143
wenzelm@11633
  2144
*** Isar ***
wenzelm@11633
  2145
wenzelm@12312
  2146
* Pure/Provers: improved proof by cases and induction;
wenzelm@12280
  2147
  - 'case' command admits impromptu naming of parameters (such as
wenzelm@12280
  2148
    "case (Suc n)");
wenzelm@12280
  2149
  - 'induct' method divinates rule instantiation from the inductive
wenzelm@12280
  2150
    claim; no longer requires excessive ?P bindings for proper
wenzelm@12280
  2151
    instantiation of cases;
wenzelm@12280
  2152
  - 'induct' method properly enumerates all possibilities of set/type
wenzelm@12280
  2153
    rules; as a consequence facts may be also passed through *type*
wenzelm@12280
  2154
    rules without further ado;
wenzelm@12280
  2155
  - 'induct' method now derives symbolic cases from the *rulified*
wenzelm@12280
  2156
    rule (before it used to rulify cases stemming from the internal
wenzelm@12280
  2157
    atomized version); this means that the context of a non-atomic
wenzelm@12280
  2158
    statement becomes is included in the hypothesis, avoiding the
wenzelm@12280
  2159
    slightly cumbersome show "PROP ?case" form;
wenzelm@12280
  2160
  - 'induct' may now use elim-style induction rules without chaining
wenzelm@12280
  2161
    facts, using ``missing'' premises from the goal state; this allows
wenzelm@12280
  2162
    rules stemming from inductive sets to be applied in unstructured
wenzelm@12280
  2163
    scripts, while still benefitting from proper handling of non-atomic
wenzelm@12280
  2164
    statements; NB: major inductive premises need to be put first, all
wenzelm@12280
  2165
    the rest of the goal is passed through the induction;
wenzelm@12280
  2166
  - 'induct' proper support for mutual induction involving non-atomic
wenzelm@12280
  2167
    rule statements (uses the new concept of simultaneous goals, see
wenzelm@12280
  2168
    below);
wenzelm@12853
  2169
  - append all possible rule selections, but only use the first
wenzelm@12853
  2170
    success (no backtracking);
wenzelm@11995
  2171
  - removed obsolete "(simplified)" and "(stripped)" options of methods;
wenzelm@12754
  2172
  - undeclared rule case names default to numbers 1, 2, 3, ...;
wenzelm@12754
  2173
  - added 'print_induct_rules' (covered by help item in recent Proof
wenzelm@12754
  2174
    General versions);
wenzelm@11995
  2175
  - moved induct/cases attributes to Pure, methods to Provers;
wenzelm@11995
  2176
  - generic method setup instantiated for FOL and HOL;
wenzelm@11986
  2177
wenzelm@12163
  2178
* Pure: support multiple simultaneous goal statements, for example
wenzelm@12163
  2179
"have a: A and b: B" (same for 'theorem' etc.); being a pure
wenzelm@12163
  2180
meta-level mechanism, this acts as if several individual goals had
wenzelm@12163
  2181
been stated separately; in particular common proof methods need to be
wenzelm@12163
  2182
repeated in order to cover all claims; note that a single elimination
wenzelm@12163
  2183
step is *not* sufficient to establish the two conjunctions, so this
wenzelm@12163
  2184
fails:
wenzelm@12163
  2185
wenzelm@12163
  2186
  assume "A & B" then have A and B ..   (*".." fails*)
wenzelm@12163
  2187
wenzelm@12163
  2188
better use "obtain" in situations as above; alternative refer to
wenzelm@12163
  2189
multi-step methods like 'auto', 'simp_all', 'blast+' etc.;
wenzelm@12163
  2190
wenzelm@12078
  2191
* Pure: proper integration with ``locales''; unlike the original
webertj@15154
  2192
version by Florian Kammller, Isar locales package high-level proof
wenzelm@12078
  2193
contexts rather than raw logical ones (e.g. we admit to include
wenzelm@12280
  2194
attributes everywhere); operations on locales include merge and
wenzelm@12964
  2195
rename; support for implicit arguments (``structures''); simultaneous
wenzelm@12964
  2196
type-inference over imports and text; see also HOL/ex/Locales.thy for
wenzelm@12964
  2197
some examples;
wenzelm@12078
  2198
wenzelm@12707
  2199
* Pure: the following commands have been ``localized'', supporting a
wenzelm@12707
  2200
target locale specification "(in name)": 'lemma', 'theorem',
wenzelm@12707
  2201
'corollary', 'lemmas', 'theorems', 'declare'; the results will be
wenzelm@12707
  2202
stored both within the locale and at the theory level (exported and
wenzelm@12707
  2203
qualified by the locale name);
wenzelm@12707
  2204
wenzelm@12964
  2205
* Pure: theory goals may now be specified in ``long'' form, with
wenzelm@12964
  2206
ad-hoc contexts consisting of arbitrary locale elements. for example
wenzelm@12964
  2207
``lemma foo: fixes x assumes "A x" shows "B x"'' (local syntax and
wenzelm@12964
  2208
definitions may be given, too); the result is a meta-level rule with
wenzelm@12964
  2209
the context elements being discharged in the obvious way;
wenzelm@12964
  2210
wenzelm@12964
  2211
* Pure: new proof command 'using' allows to augment currently used
wenzelm@12964
  2212
facts after a goal statement ('using' is syntactically analogous to
wenzelm@12964
  2213
'apply', but acts on the goal's facts only); this allows chained facts
wenzelm@12964
  2214
to be separated into parts given before and after a claim, as in
wenzelm@12964
  2215
``from a and b have C using d and e <proof>'';
wenzelm@12078
  2216
wenzelm@11722
  2217
* Pure: renamed "antecedent" case to "rule_context";
wenzelm@11722
  2218
wenzelm@12964
  2219
* Pure: new 'judgment' command records explicit information about the
wenzelm@12964
  2220
object-logic embedding (used by several tools internally); no longer
wenzelm@12964
  2221
use hard-wired "Trueprop";
wenzelm@12964
  2222
wenzelm@11738
  2223
* Pure: added 'corollary' command;
wenzelm@11738
  2224
wenzelm@11722
  2225
* Pure: fixed 'token_translation' command;
wenzelm@11722
  2226
wenzelm@11899
  2227
* Pure: removed obsolete 'exported' attribute;
wenzelm@11899
  2228
wenzelm@11933
  2229
* Pure: dummy pattern "_" in is/let is now automatically lifted over
wenzelm@11933
  2230
bound variables: "ALL x. P x --> Q x" (is "ALL x. _ --> ?C x")
wenzelm@11899
  2231
supersedes more cumbersome ... (is "ALL x. _ x --> ?C x");
wenzelm@11899
  2232
wenzelm@11952
  2233
* Pure: method 'atomize' presents local goal premises as object-level
wenzelm@11952
  2234
statements (atomic meta-level propositions); setup controlled via
wenzelm@11952
  2235
rewrite rules declarations of 'atomize' attribute; example
wenzelm@11952
  2236
application: 'induct' method with proper rule statements in improper
wenzelm@11952
  2237
proof *scripts*;
wenzelm@11952
  2238
wenzelm@12106
  2239
* Pure: emulation of instantiation tactics (rule_tac, cut_tac, etc.)
wenzelm@12106
  2240
now consider the syntactic context of assumptions, giving a better
wenzelm@12106
  2241
chance to get type-inference of the arguments right (this is
wenzelm@12106
  2242
especially important for locales);
wenzelm@12106
  2243
wenzelm@12312
  2244
* Pure: "sorry" no longer requires quick_and_dirty in interactive
wenzelm@12312
  2245
mode;
wenzelm@12312
  2246
wenzelm@12405
  2247
* Pure/obtain: the formal conclusion "thesis", being marked as
wenzelm@12405
  2248
``internal'', may no longer be reference directly in the text;
wenzelm@12405
  2249
potential INCOMPATIBILITY, may need to use "?thesis" in rare
wenzelm@12405
  2250
situations;
wenzelm@12405
  2251
wenzelm@12405
  2252
* Pure: generic 'sym' attribute which declares a rule both as pure
wenzelm@12405
  2253
'elim?' and for the 'symmetric' operation;
wenzelm@12405
  2254
wenzelm@12877
  2255
* Pure: marginal comments ``--'' may now occur just anywhere in the
wenzelm@12877
  2256
text; the fixed correlation with particular command syntax has been
wenzelm@12877
  2257
discontinued;
wenzelm@12877
  2258
berghofe@13023
  2259
* Pure: new method 'rules' is particularly well-suited for proof
berghofe@13023
  2260
search in intuitionistic logic; a bit slower than 'blast' or 'fast',
berghofe@13023
  2261
but often produces more compact proof terms with less detours;
berghofe@13023
  2262
wenzelm@12364
  2263
* Pure/Provers/classical: simplified integration with pure rule
wenzelm@12364
  2264
attributes and methods; the classical "intro?/elim?/dest?"
wenzelm@12364
  2265
declarations coincide with the pure ones; the "rule" method no longer
wenzelm@12364
  2266
includes classically swapped intros; "intro" and "elim" methods no
wenzelm@12364
  2267
longer pick rules from the context; also got rid of ML declarations
wenzelm@12364
  2268
AddXIs/AddXEs/AddXDs; all of this has some potential for
wenzelm@12364
  2269
INCOMPATIBILITY;
wenzelm@12364
  2270
wenzelm@12405
  2271
* Provers/classical: attribute 'swapped' produces classical inversions
wenzelm@12405
  2272
of introduction rules;
wenzelm@12405
  2273
wenzelm@12364
  2274
* Provers/simplifier: 'simplified' attribute may refer to explicit
wenzelm@12364
  2275
rules instead of full simplifier context; 'iff' attribute handles
wenzelm@12364
  2276
conditional rules;
wenzelm@11936
  2277
wenzelm@11745
  2278
* HOL: 'typedef' now allows alternative names for Rep/Abs morphisms;
wenzelm@11745
  2279
wenzelm@11690
  2280
* HOL: 'recdef' now fails on unfinished automated proofs, use
wenzelm@11633
  2281
"(permissive)" option to recover old behavior;
wenzelm@11633
  2282
wenzelm@11933
  2283
* HOL: 'inductive' no longer features separate (collective) attributes
wenzelm@11933
  2284
for 'intros' (was found too confusing);
wenzelm@11933
  2285
wenzelm@12405
  2286
* HOL: properly declared induction rules less_induct and
wenzelm@12405
  2287
wf_induct_rule;
wenzelm@12405
  2288
kleing@11788
  2289
wenzelm@11474
  2290
*** HOL ***
wenzelm@11474
  2291
wenzelm@11702
  2292
* HOL: moved over to sane numeral syntax; the new policy is as
wenzelm@11702
  2293
follows:
wenzelm@11702
  2294
wenzelm@11702
  2295
  - 0 and 1 are polymorphic constants, which are defined on any
wenzelm@11702
  2296
  numeric type (nat, int, real etc.);
wenzelm@11702
  2297
wenzelm@11702
  2298
  - 2, 3, 4, ... and -1, -2, -3, ... are polymorphic numerals, based
wenzelm@11702
  2299
  binary representation internally;
wenzelm@11702
  2300
wenzelm@11702
  2301
  - type nat has special constructor Suc, and generally prefers Suc 0
wenzelm@11702
  2302
  over 1::nat and Suc (Suc 0) over 2::nat;
wenzelm@11702
  2303
wenzelm@12364
  2304
This change may cause significant problems of INCOMPATIBILITY; here
wenzelm@12364
  2305
are some hints on converting existing sources:
wenzelm@11702
  2306
wenzelm@11702
  2307
  - due to the new "num" token, "-0" and "-1" etc. are now atomic
wenzelm@11702
  2308
  entities, so expressions involving "-" (unary or binary minus) need
wenzelm@11702
  2309
  to be spaced properly;
wenzelm@11702
  2310
wenzelm@11702
  2311
  - existing occurrences of "1" may need to be constraint "1::nat" or
wenzelm@11702
  2312
  even replaced by Suc 0; similar for old "2";
wenzelm@11702
  2313
wenzelm@11702
  2314
  - replace "#nnn" by "nnn", and "#-nnn" by "-nnn";
wenzelm@11702
  2315
wenzelm@11702
  2316
  - remove all special provisions on numerals in proofs;
wenzelm@11702
  2317
wenzelm@13042
  2318
* HOL: simp rules nat_number expand numerals on nat to Suc/0
wenzelm@12837
  2319
representation (depends on bin_arith_simps in the default context);
wenzelm@12837
  2320
wenzelm@12736
  2321
* HOL: symbolic syntax for x^2 (numeral 2);
wenzelm@12736
  2322
wenzelm@12335
  2323
* HOL: the class of all HOL types is now called "type" rather than
wenzelm@12335
  2324
"term"; INCOMPATIBILITY, need to adapt references to this type class
wenzelm@12335
  2325
in axclass/classes, instance/arities, and (usually rare) occurrences
wenzelm@12335
  2326
in typings (of consts etc.); internally the class is called
wenzelm@12335
  2327
"HOL.type", ML programs should refer to HOLogic.typeS;
wenzelm@12335
  2328
wenzelm@12280
  2329
* HOL/record package improvements:
wenzelm@12280
  2330
  - new derived operations "fields" to build a partial record section,
wenzelm@12280
  2331
    "extend" to promote a fixed record to a record scheme, and
wenzelm@12280
  2332
    "truncate" for the reverse; cf. theorems "xxx.defs", which are *not*
wenzelm@12280
  2333
    declared as simp by default;
wenzelm@12587
  2334
  - shared operations ("more", "fields", etc.) now need to be always
wenzelm@12587
  2335
    qualified) --- potential INCOMPATIBILITY;
wenzelm@12280
  2336
  - removed "make_scheme" operations (use "make" with "extend") --
wenzelm@12280
  2337
    INCOMPATIBILITY;
wenzelm@11937
  2338
  - removed "more" class (simply use "term") -- INCOMPATIBILITY;
wenzelm@12253
  2339
  - provides cases/induct rules for use with corresponding Isar
wenzelm@12253
  2340
    methods (for concrete records, record schemes, concrete more
wenzelm@12280
  2341
    parts, and schematic more parts -- in that order);
wenzelm@11930
  2342
  - internal definitions directly based on a light-weight abstract
wenzelm@11930
  2343
    theory of product types over typedef rather than datatype;
wenzelm@11930
  2344
berghofe@13023
  2345
* HOL: generic code generator for generating executable ML code from
berghofe@13023
  2346
specifications; specific support for HOL constructs such as inductive
berghofe@13023
  2347
datatypes and sets, as well as recursive functions; can be invoked
berghofe@13023
  2348
via 'generate_code' theory section;
berghofe@13023
  2349
wenzelm@11933
  2350
* HOL: canonical cases/induct rules for n-tuples (n = 3..7);
wenzelm@11933
  2351
paulson@13824
  2352
* HOL: consolidated and renamed several theories.  In particular:
wenzelm@14731
  2353
        Ord.thy has been absorbed into HOL.thy
wenzelm@14731
  2354
        String.thy has been absorbed into List.thy
wenzelm@14731
  2355
wenzelm@11802
  2356
* HOL: concrete setsum syntax "\<Sum>i:A. b" == "setsum (%i. b) A"
wenzelm@11802
  2357
(beware of argument permutation!);
wenzelm@11802
  2358
wenzelm@11657
  2359
* HOL: linorder_less_split superseded by linorder_cases;
wenzelm@11657
  2360
wenzelm@12917
  2361
* HOL/List: "nodups" renamed to "distinct";
nipkow@12889
  2362
wenzelm@11633
  2363
* HOL: added "The" definite description operator; move Hilbert's "Eps"
paulson@13824
  2364
to peripheral theory "Hilbert_Choice"; some INCOMPATIBILITIES:
paulson@13824
  2365
  - Ex_def has changed, now need to use some_eq_ex
wenzelm@11437
  2366
wenzelm@11572
  2367
* HOL: made split_all_tac safe; EXISTING PROOFS MAY FAIL OR LOOP, so
wenzelm@11572
  2368
in this (rare) case use:
wenzelm@11572
  2369
wenzelm@11572
  2370
  delSWrapper "split_all_tac"
wenzelm@11572
  2371
  addSbefore ("unsafe_split_all_tac", unsafe_split_all_tac)
wenzelm@11572
  2372
wenzelm@11572
  2373
* HOL: added safe wrapper "split_conv_tac" to claset; EXISTING PROOFS
wenzelm@11474
  2374
MAY FAIL;
nipkow@11361
  2375
wenzelm@11572
  2376
* HOL: introduced f^n = f o ... o f; warning: due to the limits of
wenzelm@11572
  2377
Isabelle's type classes, ^ on functions and relations has too general
wenzelm@11572
  2378
a domain, namely ('a * 'b) set and 'a => 'b; this means that it may be
wenzelm@11572
  2379
necessary to attach explicit type constraints;
nipkow@11307
  2380
wenzelm@12917
  2381
* HOL/Relation: the prefix name of the infix "O" has been changed from
wenzelm@12917
  2382
"comp" to "rel_comp"; INCOMPATIBILITY: a few theorems have been
wenzelm@12917
  2383
renamed accordingly (eg "compI" -> "rel_compI").
nipkow@12489
  2384
wenzelm@11487
  2385
* HOL: syntax translations now work properly with numerals and records
wenzelm@11487
  2386
expressions;
wenzelm@11474
  2387
wenzelm@12457
  2388
* HOL: bounded abstraction now uses syntax "%" / "\<lambda>" instead
wenzelm@12457
  2389
of "lam" -- INCOMPATIBILITY;
wenzelm@11474
  2390
wenzelm@11933
  2391
* HOL: got rid of some global declarations (potential INCOMPATIBILITY
wenzelm@11933
  2392
for ML tools): const "()" renamed "Product_Type.Unity", type "unit"
wenzelm@11933
  2393
renamed "Product_Type.unit";
wenzelm@11611
  2394
nipkow@12564
  2395
* HOL: renamed rtrancl_into_rtrancl2 to converse_rtrancl_into_rtrancl
nipkow@12564
  2396
wenzelm@12924
  2397
* HOL: removed obsolete theorem "optionE" (use "option.exhaust", or
wenzelm@12924
  2398
the "cases" method);
wenzelm@12924
  2399
wenzelm@12597
  2400
* HOL/GroupTheory: group theory examples including Sylow's theorem (by
webertj@15154
  2401
Florian Kammller);
wenzelm@12597
  2402
wenzelm@12608
  2403
* HOL/IMP: updated and converted to new-style theory format; several
wenzelm@12608
  2404
parts turned into readable document, with proper Isar proof texts and
wenzelm@12608
  2405
some explanations (by Gerwin Klein);
wenzelm@12597
  2406
wenzelm@12734
  2407
* HOL-Real: added Complex_Numbers (by Gertrud Bauer);
wenzelm@12734
  2408
wenzelm@12690
  2409
* HOL-Hyperreal is now a logic image;
wenzelm@12690
  2410
wenzelm@11611
  2411
wenzelm@12022
  2412
*** HOLCF ***
wenzelm@12022
  2413
wenzelm@12622
  2414
* Isar: consts/constdefs supports mixfix syntax for continuous
wenzelm@12622
  2415
operations;
wenzelm@12622
  2416
wenzelm@12622
  2417
* Isar: domain package adapted to new-style theory format, e.g. see
wenzelm@12622
  2418
HOLCF/ex/Dnat.thy;
wenzelm@12622
  2419
wenzelm@12622
  2420
* theory Lift: proper use of rep_datatype lift instead of ML hacks --
wenzelm@12280
  2421
potential INCOMPATIBILITY; now use plain induct_tac instead of former
wenzelm@12280
  2422
lift.induct_tac, always use UU instead of Undef;
wenzelm@12022
  2423
wenzelm@12597
  2424
* HOLCF/IMP: updated and converted to new-style theory;
wenzelm@12597
  2425
wenzelm@12022
  2426
wenzelm@11474
  2427
*** ZF ***
wenzelm@11474
  2428
wenzelm@12622
  2429
* Isar: proper integration of logic-specific tools and packages,
wenzelm@12622
  2430
including theory commands '(co)inductive', '(co)datatype',
wenzelm@12622
  2431
'rep_datatype', 'inductive_cases', as well as methods 'ind_cases',
wenzelm@12622
  2432
'induct_tac', 'case_tac', and 'typecheck' (with attribute 'TC');
wenzelm@12622
  2433
wenzelm@12622
  2434
* theory Main no longer includes AC; for the Axiom of Choice, base
wenzelm@12622
  2435
your theory on Main_ZFC;
wenzelm@12622
  2436
wenzelm@12622
  2437
* the integer library now covers quotients and remainders, with many
wenzelm@12622
  2438
laws relating division to addition, multiplication, etc.;
paulson@12563
  2439
wenzelm@12280
  2440
* ZF/UNITY: Chandy and Misra's UNITY is now available in ZF, giving a
wenzelm@12280
  2441
typeless version of the formalism;
wenzelm@12280
  2442
wenzelm@13025
  2443
* ZF/AC, Coind, IMP, Resid: updated and converted to new-style theory
wenzelm@13025
  2444
format;
wenzelm@12608
  2445
wenzelm@12280
  2446
* ZF/Induct: new directory for examples of inductive definitions,
wenzelm@12608
  2447
including theory Multiset for multiset orderings; converted to
wenzelm@12608
  2448
new-style theory format;
wenzelm@12177
  2449
wenzelm@13025
  2450
* ZF: many new theorems about lists, ordinals, etc.;
paulson@12850
  2451
wenzelm@11474
  2452
wenzelm@11474
  2453
*** General ***
wenzelm@11474
  2454
wenzelm@12280
  2455
* Pure/kernel: meta-level proof terms (by Stefan Berghofer); reference
wenzelm@12280
  2456
variable proof controls level of detail: 0 = no proofs (only oracle
wenzelm@12280
  2457
dependencies), 1 = lemma dependencies, 2 = compact proof terms; see
wenzelm@12280
  2458
also ref manual for further ML interfaces;
wenzelm@12280
  2459
wenzelm@12280
  2460
* Pure/axclass: removed obsolete ML interface
wenzelm@12280
  2461
goal_subclass/goal_arity;
wenzelm@12280
  2462
wenzelm@12280
  2463
* Pure/syntax: new token syntax "num" for plain numerals (without "#"
wenzelm@12280
  2464
of "xnum"); potential INCOMPATIBILITY, since -0, -1 etc. are now
wenzelm@12280
  2465
separate tokens, so expressions involving minus need to be spaced
wenzelm@12280
  2466
properly;
wenzelm@12280
  2467
wenzelm@12312
  2468
* Pure/syntax: support non-oriented infixes, using keyword "infix"
wenzelm@12312
  2469
rather than "infixl" or "infixr";
wenzelm@12312
  2470
wenzelm@12312
  2471
* Pure/syntax: concrete syntax for dummy type variables admits genuine
wenzelm@12312
  2472
sort constraint specifications in type inference; e.g. "x::_::foo"
wenzelm@12312
  2473
ensures that the type of "x" is of sort "foo" (but not necessarily a
wenzelm@12312
  2474
type variable);
wenzelm@12280
  2475
wenzelm@12280
  2476
* Pure/syntax: print modes "type_brackets" and "no_type_brackets"
wenzelm@12280
  2477
control output of nested => (types); the default behavior is
wenzelm@12280
  2478
"type_brackets";
wenzelm@12280
  2479
wenzelm@12280
  2480
* Pure/syntax: builtin parse translation for "_constify" turns valued
wenzelm@11817
  2481
tokens into AST constants;
wenzelm@11474
  2482
wenzelm@12280
  2483
* Pure/syntax: prefer later declarations of translations and print
wenzelm@12280
  2484
translation functions; potential INCOMPATIBILITY: need to reverse
wenzelm@12280
  2485
multiple declarations for same syntax element constant;
wenzelm@12280
  2486
wenzelm@12832
  2487
* Pure/show_hyps reset by default (in accordance to existing Isar
wenzelm@12832
  2488
practice);
wenzelm@12832
  2489
wenzelm@12280
  2490
* Provers/classical: renamed addaltern to addafter, addSaltern to
wenzelm@12280
  2491
addSafter;
wenzelm@12280
  2492
wenzelm@12280
  2493
* Provers/clasimp: ``iff'' declarations now handle conditional rules
wenzelm@12280
  2494
as well;
wenzelm@12253
  2495
wenzelm@12538
  2496
* system: tested support for MacOS X; should be able to get Isabelle +
wenzelm@12538
  2497
Proof General to work in a plain Terminal after installing Poly/ML
wenzelm@12538
  2498
(e.g. from the Isabelle distribution area) and GNU bash alone
wenzelm@12538
  2499
(e.g. from http://www.apple.com); full X11, XEmacs and X-Symbol
wenzelm@12538
  2500
support requires further installations, e.g. from
wenzelm@12538
  2501
http://fink.sourceforge.net/);
wenzelm@12538
  2502
wenzelm@12280
  2503
* system: support Poly/ML 4.1.1 (able to manage larger heaps);
wenzelm@11551
  2504
wenzelm@12753
  2505
* system: reduced base memory usage by Poly/ML (approx. 20 MB instead
wenzelm@12753
  2506
of 40 MB), cf. ML_OPTIONS;
wenzelm@12753
  2507
wenzelm@11633
  2508
* system: Proof General keywords specification is now part of the
wenzelm@11633
  2509
Isabelle distribution (see etc/isar-keywords.el);
wenzelm@11633
  2510
wenzelm@12728
  2511
* system: support for persistent Proof General sessions (refrain from
wenzelm@12728
  2512
outdating all loaded theories on startup); user may create writable
wenzelm@12728
  2513
logic images like this: ``isabelle -q HOL Test'';
wenzelm@12597
  2514
wenzelm@11551
  2515
* system: smart selection of Isabelle process versus Isabelle
wenzelm@11572
  2516
interface, accommodates case-insensitive file systems (e.g. HFS+); may
wenzelm@11572
  2517
run both "isabelle" and "Isabelle" even if file names are badly
wenzelm@11572
  2518
damaged (executable inspects the case of the first letter of its own
wenzelm@11572
  2519
name); added separate "isabelle-process" and "isabelle-interface";
wenzelm@11551
  2520
wenzelm@12472
  2521
* system: refrain from any attempt at filtering input streams; no
wenzelm@12472
  2522
longer support ``8bit'' encoding of old isabelle font, instead proper
wenzelm@12472
  2523
iso-latin characters may now be used; the related isatools
wenzelm@12472
  2524
"symbolinput" and "nonascii" have disappeared as well;
wenzelm@12472
  2525
wenzelm@12472
  2526
* system: removed old "xterm" interface (the print modes "xterm" and
wenzelm@12472
  2527
"xterm_color" are still available for direct use in a suitable
wenzelm@12472
  2528
terminal);
wenzelm@12472
  2529
paulson@11314
  2530
oheimb@11169
  2531
wenzelm@11062
  2532
New in Isabelle99-2 (February 2001)
wenzelm@11062
  2533
-----------------------------------
wenzelm@11062
  2534
nipkow@10224
  2535
*** Overview of INCOMPATIBILITIES ***
nipkow@10224
  2536
paulson@11241
  2537
* HOL: please note that theories in the Library and elsewhere often use the
paulson@11241
  2538
new-style (Isar) format; to refer to their theorems in an ML script you must
wenzelm@12622
  2539
bind them to ML identifers by e.g.      val thm_name = thm "thm_name";
paulson@11241
  2540
wenzelm@11043
  2541
* HOL: inductive package no longer splits induction rule aggressively,
wenzelm@11043
  2542
but only as far as specified by the introductions given; the old
oheimb@11130
  2543
format may be recovered via ML function complete_split_rule or attribute
wenzelm@11043
  2544
'split_rule (complete)';
wenzelm@11043
  2545
wenzelm@10998