RE: XSLT, XHTML, and default attribute values [somewha

> > Sure, but the processor must still fetch the defaulted
> values whether
> > it validates the content or not, otherwise it is
> potentially leaving
> > out data.
>
> but that's what the xml spec allows them to do. (Mozilla for
> instance never fetches any external dtd file except those
> that have copies installed into mozilla's own directory tree.
> So if you default attribute values, mozilla doesn't usually
> see the defaults.

Are you sure?  Even if Mozilla allows it, is it still this case?  Don't
get me wrong, I think defaulted values in DTDs are awful (a weird mix of
content and validation), but I've never had any luck ignoring them using
Xerces in Java without implementing a custom entity resolver.  It's
possible Mozilla implements its own custom entity resolver in the event
it cannot reach a referenced DTD, giving the appearance it has ignored
the defaulted values.  Either way, I'm speculating wildly here :)

cheers
andrew