- From: William F. Hammond <hammond@csc.albany.edu>
- Date: Tue, 17 Nov 1998 15:47:54 -0500 (EST)
- To: fiedorow@math.ohio-state.edu, rminer@geomtech.com
- Cc: www-math@w3.org
Robert Miner writes: : Finally, I certainly agree it would be a blow if a major player ... : entrenched a faulty implementation of MathML ... One way to make sure that the world establishes good MathML practice is for those interested in these questions to work together to come up with rigorously robust and efficient ways --- perhaps even at the risk of having those ways and their sponsors labeled "elitist" --- of generating good streams for MathML or whatever other good math-supporting XML-for-browser dialects that evolve. \major-aside{Please remember that I sometimes imagine billions and billions of SGML/XML dialects being processed and flowed in billions and billions of directions, all being optimally typeset for the printed page with TeX-based systems. Isn't that, after all, the XML youth-dream.} I may be able to help. But before I know how to move further in this particular direction than I have gone, I feel that I need to have comment on my drafty draft about the extant legacy of mathematical markup practice that is available at my URL http://www.albany.edu/~hammond/gellmu/notation . I myself may or may not choose to code for MathML specifically --- there are many other interesting coding targets, but I am interested and I could provide (a) guidance and (b) a starting point. I do think that the correct approach for authors is to begin with type definitions for each symbol, possibly relying on defaults like \anch{\ref{href=""}\show{Bourbaki's}} "\Z", "\Q", "\R", ... and perhaps relying on a context-dependent package of defaults, in document preambles, leaving it for processors to supply all the verbosity that MathML needs. Just keep in mind that we may not want to carry the semantics so far that we get entangled over the question of whether the axiom of choice applies. This only looks like it's not about engineering mathematics. The physicists certainly will understand. Even now the authors of some computer algebra systems will understand. At the very least it is about being able to handle that which is known to be involved in studying the solutions of finitely many polynomial equations (and sometimes "inequalities") in finitely many variables. I'm sorry that I find it necessary to be so elitist. -- Bill
Received on Tuesday, 17 November 1998 15:48:04 UTC