Re: xml:* attributes

On 15/08/2012 05:00, John Cowan wrote:
> It does, however, impair the ability of processors that don't know
> the document type to extract certain kinds of information from the
> document. This may or may not be a worthwhile goal, but it has been
> pursued not only in the xml: namespace, but for essentially all
> namespaced attributes. For example, the xslt:version attribute serves
> to identify documents that are XSLT stylesheets even when their
> elements are not in the XSLT namespace.


This is true of course, but the only reason XSLT gets itself into a
position of having elements that it doesn't know about (so benefiting
from having an xsl:version attribute that it does know about) is because
XSLT (more than just about any other system) makes essential use of
_elements_ in other namespaces.

Thus I see very little use in general for namespaced attributes if there
are no namespaced elements.

I think taking a position that micro-xml need only be compatible with
xml 1.0 not xml 1.0+NS will just lead to confusion. People get confused
enough with namespace prefixes already without introducing a spec where
you can use them but they mean something different.

So I would vote for James' choice A: do not allow colon in names. (And
also I'd ban (or otherwise discourage) the xmlns attribute, so avoiding
namespaces altogether)

David



________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 
________________________________________________________________________

Received on Wednesday, 15 August 2012 09:22:14 UTC