[4suite] default attribute values in DTD

[Uche Ogbuji]

Alexandre Fayolle wrote:
> 
> On Tue, 6 Feb 2001, Uche Ogbuji wrote:
> 
> > If I take your namespace document and process it using pyexpat rather
> > than xmlproc, suddenly I have a *compketely* different document under
> > XML NS 1.0, which means that DOM, XPath, XSLT, XPointer, XLink, RDF,
> > etc. all behave quite differently (usually by just breaking).
> >
> > This rather puts a big hole in XML's supposed interoperability.
> 
> Agreed, but this is not what I had in mind.
> 
> If 4DOM provides such a mechanism, it makes it easy for, say,
> xmltools.xmleditor to edit documents with namespaces, that are valid with
> regarding XMLNS 1.0. Requiring someone to insert a string such as
> "http://www.w3.org/1999/XSL/Transform" is both a real pain to remember and
> type without making any mistakes. Having the editor grabbing it and
> inserting it automagically from the DTD is certainly a better option.

Sure, but in this case, the DTD is nothing but a handy dictionary for
you.  Why not just make the dictionary a first-class construct, then? 

When we had a project with a similar need, we had a separate dialof box
from which people could pick namespace URIs which would be populated
into the text area (all Javascript stuff).

Basically, I don't see that this requires DTD.

> Besides, when the document is serialized with Print or PrettyPrint,
> extraneous xmlns attributes are removed and the whole thing is still
> readable with any non validating processor.

I don't follow.  If xmlns attributes are removed, how does it work with
non-validating processors?

-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji at fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python