Re: Namespace Comments

James Clark (jjc@jclark.com)
Tue, 04 Aug 1998 16:50:00 +0700


Don Park wrote:
>
> James,
>
> >Because it was felt undesirable that namespace processing should change
> >the structure of the element tree.
>
> Hmm. Namespace nodes would indeed look funny in a document displayed as a
> tree. I suppose the Namespace WG does not consider having extraneous
> attributes as change to the structure of the document.

Additional attributes do change the document but they are much less
likely to cause problems than additional container elements; they also
cause fewer problems for validation.

> What about having to change DTDs to use the namespace? Isn't that a bit
> harsh?

Any non-namespace aware validation facility is bound to be sub-optimal
with namespaces.

If you have only a single namespace, you can make a valid document use
namespaces just by changing the internal subset:

<!DOCTYPE doc SYSTEM "doc.dtd" [
<!ATTLIST doc xmlns CDATA "urn:...">
]>
<doc>...</doc>

> >Only names beginning with "xml" (any case) are reserved by the XML spec.
>
> Could we not reserve "namespace" as well? It seems quite restricting that
> only the XML spec itself can reserve names.

The namespace facility is intended eventually to become part of XML
proper, and can thus only make use of names already reserved by XML 1.0.

James