- From: Simon St.Laurent <simonstl@simonstl.com>
- Date: Tue, 23 May 2000 23:32:42 -0400
- To: <xml-uri@w3.org>
At 10:33 PM 5/23/00 -0400, Jonathan Borden wrote: >Thanks for the nice summary. After careful consideration, and my initial >inkling that absolutization was the way to go, I now conclude that the >current namespace spec is correct and that namespace URIs ***including >relative URIs*** ought be compared literally (char-by-char). > >This is most consistent with other URI usage where 'path' normalization is >done at the server's discression though the URIs themselves are not >'identical' e.g. >http://www.x.com/a/./b/c and http://www.x.com/a/b/c are not identical URIs >though resolve to the same content/document on most servers. > >Usage of relative URIs will have the well identified pitfalls. So be it. You >makes your choices you gets your consequences. I started out thinking that relative URI references were evil creatures that should be disallowed entirely in namespaces, but wind up at a similar position: compare namespace URI references (including relative ones) literally, char-by-char, and let higher levels handle absolutization when needed. Relative URI references need warning labels, but I think that can be handled as best practices rather than specs. Simon St.Laurent XML Elements of Style / XML: A Primer, 2nd Ed. Building XML Applications Inside XML DTDs: Scientific and Technical Cookies / Sharing Bandwidth http://www.simonstl.com
Received on Tuesday, 23 May 2000 23:30:50 UTC