4 <!ELEMENT minutes (about,participant*,presentation*,entry*)>
6 <!ELEMENT about (#PCDATA)>
8 <!ELEMENT entry (#PCDATA)>
10 author CDATA #IMPLIED>
12 <!ELEMENT presentation (#PCDATA|ul)*>
13 <!ATTLIST presentation
14 media (blackboard|slides) #REQUIRED
17 author CDATA #REQUIRED>
19 <!ELEMENT participant (#PCDATA)>
26 <!ELEMENT li (#PCDATA|ul)*>
30 <about>the Kick-Off</about>
31 <participant site="bologna" file="asperti"/>
32 <participant site="bologna" file="sacerdoti"/>
33 <participant site="inria" file="herbelin"/>
34 <participant site="inria" file="rideau"/>
35 <participant site="inria" file="pottier"/>
36 <participant site="inria" file="werner"/>
37 <participant site="nijmegen" file="geuvers"/>
38 <participant site="nijmegen" file="wiedijk"/>
39 <participant site="dfki" file="moschner"/>
40 <participant site="trusted-logic" file="gimenez"/>
41 <participant site="aei" file="schutz"/>
42 <participant site="aei" file="wegner"/>
43 <participant site="aei" file="velden"/>
44 <participant site="aei" file="kelley"/>
45 <participant site="aei" file="weyher"/>
46 <participant site="aei" file="pollney"/>
47 <presentation site="bologna" author="asperti" media="blackboard">
48 A comparison of the aims and approaches of HELM and OpenMath.
50 <presentation site="inria" author="pottier"
51 media="slides" slides="kick-off/lemme2.ppt">
52 Presentation of the work of Sophia-Antipolis: Project Lemme;
53 PCoq (an interface to the Coq system based on Aioli [for tree management],
54 PPML [a sort of stylesheet language] and Figue [rendering engine]).
56 <presentation site="inria" author="herbelin" media="slides">
57 Presentation of the work of Rocquencourt: Coq (proof assistant) and its
60 <li>Classification of theories:
62 <li>Actually based only on the affiliation of the author.</li>
63 <li>We should have classifications based on subject, theme, author, etc.</li>
68 <li>By pattern. Problem: the patterns (n < m+1) and (n <= m) are
69 differents but denote the same thing.</li>
70 <li>By isomorphisms. Problem: we can capture just a few of them.</li>
73 <li>Proof Rendering in Natural Language: still too many detailed.
74 Views at different levels of detail can probably help.
79 <presentation site="nijmegen" author="geuvers" media="slides">
80 Presentation of the work of Nijmegen; why Nijmegen is interested in MOWGLI;
81 planned contributions to MOWGLI.
83 <presentation site="aei" author="wegner" media="blackboard">
84 Presentation of the projects in which he is envolved or that he is
85 coordinating; interest in metadata; problems related to having papers
86 in electronic form. His main contribution will be providing links to
89 <presentation site="aei" author="schutz" media="slides">
90 Presentation of the work done at the Max Planck Institute for Gravitational
91 Physics; presentation of Living Reviews in Relativity and expectations
92 from MOWGLI (i.e. searching, rendering, interoperability). He points out
93 how Living Reviews in Relativity already provide a notion of versioning;
94 how hyperlinks are managed (often opening pop-ups). He finally presents
95 some data that show that on-line browsing of papers really happens.
97 <presentation site="trusted-logic" author="gimenez" media="blackboard">
98 Presentation of the work done at Trusted Logic. Brief introduction to
99 the Common Criteria Software: lot of documentation must be produced for
100 third (and fourth!) parties evaluation; formal evaluation is one goal
101 (not yet reached). The main problems Trusted Logic meets are:
103 <li>Presentation</li>
104 <li>Managing thousands of definitions/theorems and links between them</li>
105 <li>Evaluators needs: hiding/displaying information; different views on the
106 same proofs/definitions; metadata; backpointers (which lemmas are
107 used in a theorem)</li>
108 <li>Interoperability with other software tools</li>
109 <li>Proofs mantainance</li>
111 A final remark is that Trusted Logic is just interested in provability
112 (and proof-scripts) and not in proofs (i.e. lambda-terms or natural
113 language description of them).
115 <presentation site="dfki" author="moschner" media="slides">
116 Presentation of the work done at DFKI. Contributions to MOWGLI: OMDoc
117 (to encode mathematical documents) and/or MBase (to distribute mathematical
118 documents) and metadata.
120 <entry author="aei/wegner">
121 Whenever someone is going to make a talk, he must report this to him.
124 To reach an agreement on the consortium agreement, Trusted Logic will
125 send a completely filled-in, light version of the agreement to everybody.
126 If the agreement will not satisfy everybody up to minor modifications,
127 we will go for the heavy proposal.
129 <entry author="nijmegen/geuvers">
130 The members of the PCC are responsible to communicate the name of the
131 site responsible for every WP.
134 There will be two mailing lists. The first one (for everybody) is
135 the one already created. The second one will be an administrative
139 All the meetings have already been scheduled: the first one will be held
140 in Nijmegen from the 17th to the 19th of July. The others will be
141 in Bertinoro (just after MKM03), Sophia (October 2003), Saarbrucken
144 <entry author="trusted-logic/gimenez">
145 It will be simpler to have PCC meetings by phones.
147 <entry author="aei/wegner">
148 During the plenary meetings, sub-meetings will be organized to discuss
151 <entry author="bologna/asperti">
152 Every Package Leader must submit a contribution to the deliverable 0.a.
154 <entry author="aei/wegner">
155 It is better to use a task-force of external experts to comment on
158 <entry author="aei/wegner">
159 The set of requirements must be ordered by importance (useful if some of
160 them are not fulfilled).
163 A long discussion on the topic of deliverable 1.c showed that there is
164 some mismatch in the vocabulary of the participants.
166 <entry author="bologna/asperti">
167 The choice between MathML and OpenMath is quite difficult, because
168 there are really no strong pro and cons in favor or against any of them.
170 <entry author="aei/schutz">
171 Maybe not having browsers supporting MathML is not a huge problem.