- From: <noah_mendelsohn@us.ibm.com>
- Date: Tue, 11 Feb 2003 10:14:44 -0500
- To: "David Orchard" <dorchard@bea.com>
- Cc: jones@research.att.com, xml-dist-app@w3.org
David Orchard asks:
>> Could you point to the reasoning why my suggestion on
>> requirement R6 was rejected? I can live without
>> (and prefer to) relative URIs, but I have a tough
>> time living without URI-References.
I think our intention was this. On the web, resources are named with
absolute URIs. For example, my web page might be named:
http://example.org/noah.html
A web resource is never named with a URI reference
http://example.org/noah.html#somePara
So, we believe parts should be named with absolute URIs, like any other
web resources.
As far as I know, URI References are, as the name implies, a mechanism
available when you REFER to a resource. THESE WE DO INTEND TO ALLOW, as
far as I know. So you can indeed when referring to my web page say:
http://example.org/noah.html#somePara
which means: "find the resource named http://example.org/noah.html,
determine its media-type, and use that to interpret #somepara." We do
intend to allow such references in a SOAP envelope, or anywhere else, but
the resource is named with the absolute URI. Similarly:
./noah.html#somePara
might resolve to the same document and paragraph, >if a suitable base URI
is known<. If the binding or other mechanism provides such a base, then
these references will be allowed too. Still, the name of the resource is
the absolute name. It's the reference that's richer. Thanks.
------------------------------------------------------------------
Noah Mendelsohn Voice: 1-617-693-4036
IBM Corporation Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------
Received on Tuesday, 11 February 2003 10:17:09 UTC