W3C home > Mailing lists > Public > xml-names-editor@w3.org > June 1999

RE: namespace for unqualified attributes

From: Andrew Layman <andrewl@microsoft.com>
Date: Fri, 4 Jun 1999 12:30:53 -0700
Message-ID: <5BF896CAFE8DD111812400805F1991F708AAF682@RED-MSG-08>
To: "'Paul Rabin'" <prabin@odi.com>, xml-names-editor@w3.org
All of those alternatives listed are incorrect.  It would be more accurate
to say something like: Unqualified attributes belong to a small namespace
associated with the containing element type.  However, the exactly ideal
wording has never been worked out.  I recommend you look closely at the
examples in th excellent Appendix 3 written by Tim Bray.  It should make the
intention more clear.

-----Original Message-----
From: Paul Rabin [mailto:prabin@odi.com]
Sent: Friday, June 04, 1999 11:51 AM
To: xml-names-editor@w3.org
Subject: namespace for unqualified attributes


The "Namespaces in XML" specification should, but appears not to
normatively specify which of the following is true for unqualified
attributes, that is, attributes whose name matches NCName rather than QName:

a) they belong to no namespace (or to an unnamed namespace associated with
the document), or

b) they belong to the same namespace as the element in whose start tag they
occur, or

c) it is unspecified which of a) or b) is the case.

I've found three parts of the specification that bear on this point:

Section 5.2 says:

	"Note that default namespaces do not apply directly to attributes."

Section 5.3 says:

	"... the default namespace does not apply to attribute names."

Appendix A.3 says:

	"... the names of all unqualified attributes are assigned to the
appropriate per-element-type partition."

Regarding the text cited above from section 5.2, a) it is not clear whether
it is normative, since it is a note; and b) it suggests that default
namespaces might apply to attributes in some indirect manner.

It might also be doubted whether the text cited above from section 5.3 is
normative, since it describes an example, and draws conclusions from
requirements purportedly given elsewhere in the specification rather than
stating requirements directly.  The example, however, is not consistent
with a requirement that unqualified attributes belong to the namespace of
their element, since in this case, the namespace of the element is the
default namespace.

The text cited from appendix A.3 is, of course, non-normative, but seems to
imply that unqualified attributes do belong to the namespace of their

If there is other text that bears on this question, I would be most
grateful if it were brought to my attention.

	Sincerely yours,

	Paul Rabin
Received on Friday, 4 June 1999 15:31:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:13:22 UTC