- From: Dan Connolly <connolly@w3.org>
- Date: Fri, 19 May 2000 09:47:01 -0500
- To: John Cowan <jcowan@reutershealth.com>
- CC: "xml-uri@w3.org" <xml-uri@w3.org>
John Cowan wrote: > > Dan Connolly wrote: [...] > > The only answer that I can see that's consistent with > > both the XPath implementations and Web Architecture is: "don't do that." > > Well and good. But what is to be done with existing documents? I thought I answered that; you cut it out, so I'll repeat: "The infoset for those document is unspecified. They conform to the namespace spec, but only because of a bug that wasn't caught; the namespace editors never intended to allow relative URI references. Authors are cautioned not to create such documents, since we have observed implementations that are inconsistent with the XPath spec in this regard, and we're not going to specify what DOM 2 does with them." > Declare > them crude and mean? ("For Art stopped short/In the cultivated court/ > Of the Empress Jo-o-sephine....") > > Even people who agree that straight URI interpretation ("absolutize") would > have been the way to go back in 1998 may believe (and some do) that > it's simply too late. > > > "[...] Authors are cautioned not to create such > > documents," > > Too late now. Too late for what? Too late to publish such a caution? Surely not. Too late any such documents from being created? Evidently so. Too late to say that those documents have unspecified results in DOM 2 and the Infoset? I don't think so. Too late for authors to find a way to get by without them, at least until we revise some specs? I don't think so. In case I seem to be inconsistent: my position remains: I much, much prefer that namespace names work the way the XPath spec says they work: expand them to absolute form before doing anything else with them. But I'm not sure I can convince/compel the implementors to do that, so in the interest of interoperability, I would find it acceptable to (essentially) forbid relative URI references in namespace declarations. But I find it unacceptable to use relative URI references as "universal names" without expanding them to absolute form. -- Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 19 May 2000 10:47:21 UTC