RE: URIs versus wannabe-URIs

> 
> Do we need to do anything beyond reference the following text from the
> XML Schema spec? It already mentions XLink escaping, and the
> requirement that the result be a valid URI:

That's part of the solution. The difficulty is finding all the places where
the specs are currenly loose. For example, the specification of fn:doc() in
F+O says that the argument must be an xs:anyURI, but the summary in the
language book (2.3.4) assumes that this xs:anyURI will be a URI.

We could of course try to fix it with a blanket disclaimer of the form "when
we say URI, we really mean xs:anyURI" but I suspect that will leave many
ambiguities.

Michael Kay

Received on Wednesday, 4 August 2004 09:32:38 UTC