[Cross-spec] Namespace URI? Namespace Name? Namespace IRI?

• From: Michael Kay <mhk@mhk.me.uk>
• Date: Sat, 21 Feb 2004 10:55:56 -0000
• To: <public-qt-comments@w3.org>
• Message-ID: <001001c3f869$48ba2880$6401a8c0@pcukmka>

The term "namespace URI" is scattered liberally throughout our
documents. It is used in the names of some functions, including the
fn:namespace-uri() inherited from XPath 1.0. 

But XML Namespaces has always used the term "namespace name", and in 1.1
the namespace name is not a URI, it is an IRI. And just to complicate
the matter further, although you are supposed to use an IRI as the
namespace name, the Namespaces Rec says that the parser doesn't have to
check that it's a valid IRI, and our specs (certainly the XSLT spec) say
that you mustn't check it.

My inclination is to live with the inconsistency. We should just add a
note to the relevant specs, and most particularly to the data model,
pointing out that we are using the term "namespace URI" to mean
"namespace name", and that its value can be any IRI, or perhaps anything
you like except a zero-length string.

Michael Kay

Received on Saturday, 21 February 2004 05:55:14 UTC