[Axiom-developer] Axiom Conference Call Sept 18, 2006

[root]

Participants: Kai Kaminski, Gabriel Dos Reis, Bill Page, Tim Daly
Start: 1pm
End: 2pm

1) Bill looked into the details of SVN/SVK on axiom-developer.org
   but has not yet succeeded in building an installation. 

2) Tim looked at porting Axiom to the MAC. 

   There is an outstanding GCL issue. GCL save-system cannot save
   images that will execute successfully. 

   Bill suggested trying to use compiler-link instead.
   This will be investigated.

3) Tim spent time this weekend looking into Martin's build issues
   without success.

   Suggestion that Jose send a diff of his changes to --patch-50
   so Tim can see what approach he was taking

4) The multiple-image-per-page issue with the original book was discussed.
   Kai suggested fixing the postscript files so they all contain the
   bounding box. Tim didn't think that was going to affect the floating
   issue in Latex but Tim was going to review it anyway.

5) Indefinites. 

   Tim mentioned that indefinites could be done using provisos.

   Bill objected that provisos was a proof-theoretic approach
   and that an algebraic approach might be more axiom-like.
   Thus, express Indefinite(Integer) as a Poly(Int) domain if there
   was sufficient coverage.

   Kai thought that both approaches were needed.

6) Kai has a student who wants to work on provisos.
   Tim agreed to share work with the student.

7) Tim did not yet recover the solaris changes.

   Bill pointed out that the solaris port could happen in two
   different environments, a standard solaris and a solaris with
   the GNU tool chain. He feels that the Build-improvements path
   would be more fine-grained rather than the global decision
   approach used by Gold. Thus the solaris port work is better
   left to Build-improvments.

8) Bill tried to build GCL-2.6.8pre on Windows without success.
   He points out that the older versions of GCL will no longer
   build and the fault is likely somewhere in his build setup.

9) Google objects to having binaries in the zips directory.
   It was agreed that since Google was only a method of getting
   around the breakage of Sourcforge SVN and that we plan to put
   SVN on axiom-developer therefore we should ignore Google going
   forward.

10) Gaby pointed out that SVK was more efficient than SVN
    Bill will look at SVK in more detail. There appears to be a
    Perl dependency that is at issue.

    Bill and Gaby will pursue this offline.

Tim will be generally unavailable until next monday evening.
The next scheduled conference call will be announced later.

Tim, for the group