Re: SAX 2.0 enhancement proposal

[David Brownell <david-b@pacbell.net>]

> You seem to have chosen to ignore the 'Unless' qualifier at the beginning of
> the sentence.  It is clear from the spec that the relative URIs can mean
> whatever the particular application wants them to mean.

See my response to Richard Tobin on that point.  Oh, and my original
comment about relative URIs and trouble ... :)

I wasn't ignoring the "unless", but there sure appears to be a mismatch
between semantics discussed in the first half of that sentence, applied
by applications (like xml:base as applied to xhtml:a hrefs, which is like
one of those motivating examples) and those that are part of XML itself
(the second half of the sentence).

It's a truism that application level semantics are out of scope for the XML
spec ... saying to me there was no need for the first half of that sentence.

So I'd agree that's a place where the XML spec is strange, and where you
could probably cause some erratum to make XML mean what you want.


>      The common case is
> that they are relative to the containing entity, but this is not the only
> possible case.

Yeah, you could read that to say that if someone puts a particular
PI in the document epilogue (!), then it can redefine every entity that
was declared in the DTD and, presumably, read already.  That's
the kind of interpretation that you're advocating as a good thing??

- Dave