whew. hours worth of email! 95 new, nonspam emails!
> Many contributions consist of several packages, categories and domains, which
> are somehow related.
There are some clear relations. For instance, many files contain more than
one category, domain, or package (CDP).
The very highest level of thought I've given to the problem of
documenting the algebra (the same problem exists in the CATS (computer
algebra test suite) project) and amounts to this:
is there an existing categorization of math we can use?
there is a general purpose math categorization but it is too general
there appears to be a specific one for numeric mathematics done by NIST
there does not appear to be a specific categorization for
symbolic mathematics. we may need to invent one. in fact I
wrote up a grant proposal to NIST for exactly this effort
but they have $0 dollars for new research in this year's budget.
At the next level comes the axiom-specific question of how to organize
the algebra code. Bill and I spent a fair amount of time trying to
figure out how to display and organize the algebra hierarchy and, so
far, there is still no output on the problem.
Somehow I envisioned an attack on the problem that kept each CDP in
a separate, hyperref connected, literate document which would be a
\section in a larger \chapter in a larger \subject.
The algebra is a graph at its core and a lattice above the first 50-75
files. This does not lend itself well to any linear layout. Hyperlinking
helps but you REALLY need a good map of where you are. Bill has some
experience with a tool called TOUCHGRAPH.
Writing the algebra is going to first have to attack the "map"
and layout problems so we can each work on subgraphs and know
where we are.
t
_______________________________________________
Axiom-developer mailing list
Axiom-developer@nongnu.org
http://lists.nongnu.org/mailman/listinfo/axiom-developer
