1 0.5.8 - ... - toward the 1.x series
2 * Complete rewriting of paramodulation code (thanks to Maxime Denes),
3 that is abstract over the data type embedded in the fisrt order
4 theory the procedure is able to handle.
5 Can be used as a stand-alone prover, and partecipated
6 to the CASC22 beating many older systems ;-)
7 * new (compact) kernel:
9 * height (inter object dependency) driven conversion
10 * compact metavariable representation
11 * experimental irrelevant arguments handling in convesion
12 (in the spirit of proof irrelevance)
13 * explicit predicative universes
15 * bi-directional type inference
17 * implicit vector (expands to the tight number of implicits)
19 * new type for tactics, with a global view on the ongoing proof
20 * heavily based on the refiner
21 * work with terms labelled with their contexts
22 * new auto (prolog like) proof search tactic:
23 * works on a set of goals
24 * non uniform context support:
25 * arbitrary tactics can be called to attack a goal (hardcoded
26 at the OCaml level), for now just apply and intro
28 0.5.7 - 15/02/2009 - PĂ doa release
29 * are_convertible bug solved, arguments of application where
30 compared allowing cumulativity. This allowed to prove Type0=Type1.
31 * cases tactic speedup in the simplest case of an inductive type
32 hose right parameters have all to be abstracted when the outtype is
34 * maction support added to output notation (adopted for "=" that
35 expands to "= \sub T" to show the equality type)
36 * generation of derived lemmas rewritten to be based on hooks that
37 are triggered every definition
38 * composition of coercions fixed, more composite are generated
39 * undo mechanism for coercions remade, should work better
40 * new command "prefer coercion foo" to reorder coercions.
41 * UTF-8 eq classes and virtuals described in the manual and
42 consistently printed in the TeX/UTF-8 table
43 * added a memory system for UTF-8 equivalence classes, so that
44 last choice is remembered and attempted first next time the same
45 eq class is walked trough
46 * some core notation fixed to allow a better setoid-path
47 representation (foo ^-1 changed precedence)
48 * matita manual updated with a section regarding the usage of
49 the liveCD with the VirtualBox emulator
50 * liveCD fixed to allow to easily mount/unmount real host folders
52 * ttf-freefont package added to the liveCD together with gucharmap
53 * coercions from convertible types are not inserted for real, they
54 just behave as hints for the unification algorithm
55 * virtuals (eq classes) fixed w.r.t. greek letters
56 * some virtuals added to ease the writing of diamond (<>), box ([])
58 * code that generates eliminators fixed to use whd ~delta set to true
59 when counting products, this allows to used definitions as inductive
61 * match ... with rendering fixed to break the box containing
62 constructors arguments if needed (i.e. record projections
63 are now almost readable)
64 * fixed pattern matching over ast terms, Ast.Case construct can be
65 matched correctly. Look the o-algebra example if you are
67 * Mpadded MathML tag can be used in notation, it allows to overlap
68 symbols reducing the bounding box size
70 0.5.6 - 1/12/2008 - bugfix release
71 * more abstract disambiguation algorithm, simpler instantiation
72 to a different CIC/refiner
73 * natural deduction support improved in the first order case
74 * natural deduction lem rule does now support lemmas
75 with (up to) 3 premises (multicut rule, displayed as
78 0.5.5 - 17/11/2008 - bugfix release with students in mind
79 * by ... we proved fixed to use only the specified lemmas but
80 using full unification inside auto.
81 * new apply rule tactic, that exploits the goal type to
82 disambiguate the input term.
83 * new didactic/ library directory, with support for natural deduction
86 0.5.4 - 19/10/2008 - bugfix release
87 * When a file is opened, the cursor is placed at the begin of the
88 buffer and not atthe end as before
90 * More code in the direction of a fully functional matita status, that
91 improved undo reliability in the parser/notation modules
92 * matitac was seldom compiling up-to-date files, fixed
93 * Memory consumption durin proof construction cut down using Lazy.t
95 * mstyle support in notation for text color, font size
96 * AutoGui now scales fonts to the correct user-requested size
97 * Non linear pattern matching from the level of terms to the
98 one of content in interpretation command (if the same variable name
99 is used, the two captured terms must be alpha equivalent to match)
101 0.5.3 - 23/7/2008 - bugfix release
102 * many fixes concerning the CProp hiearchy
103 * coercion database simplified
104 * coercion hiding now works properly for coercions to funclass
105 * triangular pullback works better with coercions to the function space
106 * notation for lists fixed to add a break point after the separator
107 * notation for the existential is now user definable
108 * \infrule layout added, allows to display readable fractions
109 * better window for terms grammar and TeX/Unicode
110 * fixed a bug in the positivity check not considering some subterms
111 * fixed some GUI glitches thanks to glade-3
113 0.5.2 - 2/7/2008 - better-usability-for-the-working-constructivist release
114 * refinement of match fixed to prevent useless unfolding,
115 head_beta_reduce is used instead of whd ~delta:true
116 * CProp hierarchy, interleaved with type (used to be a single universe)
117 * Notation now allows to set the precedence level of subterms, avoiding
118 unnecessary parenthese
119 * UTF-8 / TeX conversion table is now available in the Help menu
120 * Notation subsystem described in the manual
121 * EBNF term grammar is now available in the View menu (and is in sync with
122 the currently loaded notation)
123 * Rewrite tactic (in hypothesis) now computes the right sort and
124 uses the right equality elimination principle
125 * URI free interpretation and coercion statements (scripts are almost
127 * elim and cases tactic now support a pattern to specify hypotheses that
128 are generalized before the elimination.
130 0.5.1 - 29/5/2008 - minor bug fix release
131 * visualization of inductive types reports the number of fixed parameters
132 * a wrong context was used to refine fixpoints arguments
133 when trying to optimize out the letin (grep for `AvoidLetIn):
134 (let rec f x = Fix... in f t ---> Fix... t)
135 * auto fixed to prefer goals with metavariables to closed ones,
136 added new syntax to specify the universe "auto by t1, t2, ...",
137 updated documentation describing all auto parameters
138 * declarative language syntax and documentation ported to the new
141 0.5.0 - 9/5/2008 - bugfix release
142 * first release not considered experimental
144 0.4.98 - ??/11/2007 - bugfix release
145 * compiles against camlp5 >= 5.0
146 * changed lablgtksourceview module name since it is now part of lablgtk2
148 0.4.97 - 16/11/2007 - initial release