> ...agreed but the author of the XML should still leave the namespace
> declarations in the XML
yes I agree (it even says that in the namespace REC;-) but as a
practical matter given the state of most (x)html authoring, if people
get as far having a validator confirm that their document is dtd-valid,
it's not surprising that people think it's OK.
We (I) took this defaulting out of MathML+XHTML (somewhere between
mathml 1.1 and mathml 2, I think) but with some fear. Yes it forces
people to be more explicit, and it incidentally enables the msxml parser
in IE to parse the document (it rejects anydocument that references a
DTD that defaults a namespace, even if the namespace is explict in the
instance) BUT it also means that <math xmlns="any old rubbish">... is
also DTD-valid.
David
