Namespaces does *not* formally introduce "global attributes"

Ronald Bourret rbourret at
Thu Feb 4 16:53:26 GMT 1999

David Megginson wrote:

> > AHA! So the namespaces proposal _DOES_ go beyond simply qualifying
> > names with URIs!
> No it doesn't.

Mea culpa.  David is right.  Global attributes are no longer a normative 
part of the spec.

The namespaces spec says the following about non-xmlns attributes:

a) They can be prefixed or unprefixed
b) The default namespace does not apply to unprefixed attributes
c) Applications should use the namespace name, not the prefix, in 
constructing qualified names
d) No tag can contain two attributes whose names are identical or whose 
names resolve to the same qualified name; this clarifies what is meant by 
"name" in the Unique Att Spec validity constraint in XML

It does not tell us:

e) What the difference is between prefixed and unprefixed attribute names 
(other than the existence of the prefix) -- that is, the semantic 
difference (if any) between prefixed and unprefixed attributes
f) How to process prefixed or unprefixed attributes, except as noted in (c)

Two further comments:
1) Tim Bray's statement that unprefixed attributes do not belong to an XML 
namespace derives from (b).  Since there is no prefix to associate them 
with an XML namespace, and we can't the default XML namespace doesn't 
apply, there is simply no association.  However, as A.2 points out, many 
applications are likely to use traditional, per-element namespaces for 
unprefixed attributes. Note that these are a consequence of the Unique Att 
Spec validity constraint in the XML spec, not anything in the namespaces 

2) As far as I can tell, there is nothing in the normative part of the spec 
that would lead us to conclude the existence of a global attribute 
partition that is separate from the per-element-type partitions, as is 
described in A.2.  That is, the namespaces spec gives us no reason to 
believe that ns:a attributes in the following elements are in any way 
related, any more than the b attributes are related:

   <foo ns:a="1" b="1"/>
   <bar ns:a="1" b="1"/>

Is this correct?

-- Ron Bourret

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 (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