SAX2 Namespace Support

David Brownell david-b at
Tue Dec 21 14:59:20 GMT 1999

David Megginson wrote:
> Ray Waldin <rwaldin at> writes:
> > Overall, very nice!
> >
> > My only concern is that LexicalHandler should be required by all parsers.
> > Otherwise you will get "fully compliant" SAX2 parsers which cannot be used to
> > resolve QNames found in attribute values against in-scope namespace
> > declarations. There are many examples where that's critical:
> >
> > - evaluate XPath expression (XLST, XPointer, etc.)
> > - resolve XLink locator role (see
> > - follow XML Schema references (see
> >
> >
> > and probably more to come...
> >
> > IMHO, LexicalHandler must be supported by all SAX2 parsers.
> Or else we can put the callbacks back into a separate
> NamespaceHandler, so that parsers are not forced to report comments,
> CDATA section boundaries, and other noise as well.

I certainly prefer to see marginally relevant stuff like
comments and CDATA boundaries remain marginal.

The core data models of XML are elements, text, and (for
some) PIs.  Namespace-aware element (and attribute) processing
won't change that; most of the rest is noise.

> Do others agree that the scope of NS declarations is essential
> (i.e. shouldn't be optional)?  I knew that XSLT needed it, but I
> hadn't realized that so many other apps were now relying on resolving
> prefixes in attribute values and character data -- I need to keep more
> up to date on the specs.

Those three specs listed above seem to be a convincing argument
for exposing the capability if it's already in the parser.

- Dave

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list