Re: [xml-dev] RE: The Best Technologies Don't Win

[David Carlisle <davidc@nag.co.uk>]

How did this thread get hijacked onto MathML?, still...

> BTW, I agreed with Hixie that making MathML serializable as HTML5  
> probably makes sense.

It's easy to make a version of the MathML DTD compatible with the SGML
declaration for HTML rather than the standard one which is compatible
with (the SGML declaration for) XML. One could choose while doing that
whether to make some of the end tags (or even start tags) optional if
that was thought desirable. I did this back in 1998 or so to make it a
bit easier to work with sp/jade without having to keep supplying the sgml 
declaration for xml, however one consequence of that would be that all
the empty elements (of which MathML has several) would _have_ to be
marked up as <log> not <log/>. Although given the confusion surrounding
<br> versus <br/> it would perhaps make sense for HTML5 generally to use
an SGML declaration such that if an element is declared EMPTY, then
<foo> and >foo/> are both legal and equivalent (I assume that's
possible, been a while since I tinkered with SGML declarations).

David