Re: semantics, extensibility

>I hate to be so blunt, but you will find your proof at
>http://www.lfw.org/math/.  Though a few things have been
>evolving, it's been sitting there for weeks.

And I do intend to look at it when I have time. But as I understand it,
this is part of your MINCE system -- is it completely obvious how it would
fit in to the current HTML-Math proposal? Maybe you should write up
a version of your system of extensibility (as concisely as possible
whilke being complete) using the terminology and notation of the
current HTML-Math proposal (i.e. my letter of around June 1),
and send it to the group as a proposed addendum to that letter.

Or, send a letter to the group briefly outlining the issues it solves
and how it solves them -- for example, do you distinguish between
"semantic" macro expansions (which might sensibly be overridden by reader's
style preferences), and "abbreviation" expansions (which might not)?
What parts of expanded forms are further expanded? What parts of
expansion rules themselves are expanded using prior rules?
When text is copied, is it copied in unexpanded form with applicable
expansion rules included, and if so, exactly how does this work?
How do you match sequences of arguments within variable length forms?
Can a rule match any pattern or just a certain form which "looks like
an instance of a macro use"?