1 (* Title: build and test isac on Isabelle2013
2 Author: Walther Neuper, TU Graz, 100808
3 (c) due to copyright terms
5 For creating a heap image of isac see ~~/ROOT.
7 ATTENTION: no errors in this theory do not mean that there are no errors in Isac;
8 errors are rigorously detected when creating a heap.
11 header {* Loading the isac mathengine;
12 for debugging see text at begin (theory dependencies!) *}
19 use "ProgLang/termC.sml"
20 use "ProgLang/calculate.sml"
21 use "ProgLang/rewrite.sml"
22 "use_thy ProgLang/ListC"
23 "use_thy ProgLang/Tools"
24 "use_thy ProgLang/Script"
25 use "ProgLang/scrtools.sml"
26 *) "ProgLang/ProgLang"
28 (* use "Interpret/mstools.sml"
29 use "Interpret/ctree.sml"
30 use "Interpret/ptyps.sml"
31 use "Interpret/generate.sml"
32 use "Interpret/calchead.sml"
33 use "Interpret/appl.sml"
34 use "Interpret/rewtools.sml"
35 use "Interpret/script.sml"
36 use "Interpret/solve.sml"
37 use "Interpret/inform.sml"
38 use "Interpret/mathengine.sml"
39 *) "Interpret/Interpret"
41 (* use "xmlsrc/mathml.sml"
42 use "xmlsrc/datatypes.sml"
43 use "xmlsrc/pbl-met-hierarchy.sml"
44 use "xmlsrc/thy-hierarchy.sml"
45 use "xmlsrc/interface-xml.sml"
48 (* use "Frontend/messages.sml"
49 use "Frontend/states.sml"
50 use "Frontend/interface.sml"
53 *) "Frontend/Frontend"
55 "Knowledge/Build_Thydata" (*imports Isac.thy etc*)
60 show theory dependencies using the graph browser,
61 open "browser_info/HOL/Isac/session.graph"
62 and proceed from the ancestors towards the siblings.
65 section {*check presence of definitions from directories*}
67 ML {* is_reall_dsc; (*from "ProgLang/scrtools.sml" *) *}
68 ML {* me; (*from "Interpret/mathengine.sml"*) *}
69 ML {* contextthyOK2xml; (*"xmlsrc/interface-xml.sml"*) *}
70 ML {* print_exn_unit *}
71 ML {* list_rls (*from Atools.thy WN130615??? or ProgLang???*) *}
73 ML {* eval_occurs_in (*from Atools.thy*) *}
74 ML {* @{thm last_thmI} (*from Atools.thy*) *}
75 ML {*@{thm Querkraft_Belastung}*}
77 ML {* check_guhs_unique := false; *}
78 ML {* writeln "**** isac kernel = math-engine + Knowledge complete ******" *}
79 ML {* @{theory "Isac"} *}
80 ML {* CHECK_FILLPATS "at end of Build_Isac";
81 (*"CHECK_FILLPATS: diff_sin_chain exists, fillpat --> NOT empty !!!"*)*}
83 section {* State of approaching Isabelle by Isac *}
85 Mathias Lehnfeld gives the following list in his thesis in section
86 4.2.3 Relation to Ongoing Isabelle Development.
88 subsection {* (0) Survey on remaining Unsynchronized.ref *}
90 REMOVE WITHOUT REPLACEMENT (requires some efforts)
91 calcelems.sml:val theory' = Unsynchronized.ref ([]:(theory' * theory) list);
92 calcelems.sml:val isab_thm_thy = Unsynchronized.ref ([] : (thmDeriv * term) list);
93 calcelems.sml:val isabthys = Unsynchronized.ref ([] : theory list);
94 REPLACE BY KEStore... (has been overlooked)
95 calcelems.sml:val rew_ord' = Unsynchronized.ref ...
97 calcelems.sml:val trace_rewrite = Unsynchronized.ref false;
98 calcelems.sml:val depth = Unsynchronized.ref 99999;
99 calcelems.sml:val lim_rewrite = Unsynchronized.ref 99999;
100 calcelems.sml:val lim_deriv = Unsynchronized.ref 100;
101 Interpret/script.sml:val trace_script = Unsynchronized.ref false;
102 KEEP FOR EASIER DEVELOPMENT
103 calcelems.sml:val check_guhs_unique = Unsynchronized.ref true;
105 Knowledge/GCD_Poly_ML.thy: val trace_div = Unsynchronized.ref true;
106 Knowledge/GCD_Poly_ML.thy: val trace_div_invariant = Unsynchronized.ref false;
107 Knowledge/GCD_Poly_ML.thy: val trace_Euclid = Unsynchronized.ref true;
109 subsection {* (1) Exploit parallelism for concurrent session management *}
110 subsection {* (2) Make Isac’s programming language usable *}
111 subsection {* (3) Adopt Isabelle’s numeral computation for Isac *}
113 In 2002 isac already strived for floating point numbers. Since that time
114 isac represents numerals as "Free", see below (*1*). These experiments are
115 unsatisfactory with respect to logical soundness.
116 Since Isabelle now has started to care about floating point numbers, it is high
117 time to adopt these together with the other numerals. Isabelle2012/13's numerals
118 are different from Isabelle2011, see "test/Tools/isac/ProgLang/termC.sml".
120 The transition from "Free" to standard numerals is a task to be scheduled for
121 several weeks. The urgency of this task follows from the experience,
122 that (1.2) for "thehier" is very hard, because "num_str" seems to destroy
123 some of the long identifiers of theorems which would tremendously simplify
124 building a hierarchy of theorems according to (1.2), see (*2*) below.
126 ML {*(*1*) Free ("123.456", HOLogic.realT) *}
128 val unknown = filter ((curry op= "??.unknown") o fst) isacrlsthms';
129 unknown |> nth 1 |> snd |> term_to_string''' @{theory};
130 unknown |> nth 2 |> snd |> term_to_string''' @{theory};
131 (*but these seem ok:*)
132 Thm.get_name_hint @{thm add_0};
133 Thm.get_name_hint (num_str @{thm add_0});
135 subsection {* (4) Improve the efficiency of Isac’s rewrite-engine *}
136 subsection {* (5) Adopt Isabelle/jEdit for Isac *}