Another errata?

james anderson James.Anderson at
Sat Jan 30 15:45:32 GMT 1999

Tim Bray wrote:
> No.  Adam Donahue is correct on this one.  The spec says explicitly,
> and then belabors the point with examples; defaults don't apply to
> attributes.  The only attribute that has a namespace is one that
> has a prefix.
> ...
> Now clearly, an application processing an unprefixed attribute will
> know the namespace (if any) of the element to which it's attached,
> and may well use that information.  For example, in
>   <html:a href="...">
> an HTML processor may (& probably should) choose to process that href
> attribute in an HTML-ish way; BUT, the href attribute is not, in the
> formal terminology of the spec, in the html or any other namespace. -T.


If this is true, then what is the "Per-Element-Type Partition" described at
great length on the appendix on the "internal structure" of xml namespaces?
The spec would appear to assert some form of relation between the namespace
which comprises an unqualified attribute name and the namespace of the element identifier.

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe 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