- From: Tim Berners-Lee <timbl@w3.org>
- Date: Wed, 14 Jun 2000 16:55:57 -0400
- To: <xml-uri@w3.org>, "Simon St.Laurent" <simonstl@simonstl.com>
-----Original Message----- From: Simon St.Laurent <simonstl@simonstl.com> To: xml-uri@w3.org <xml-uri@w3.org> Date: Tuesday, June 13, 2000 7:36 PM Subject: Re: 1343 messages later >At 04:12 PM 6/13/00 -0400, Tim Berners-Lee wrote: >>I would summarize [1] as clarified as the abolutization in principle >>but with it pointed out that in certain cases systems >>can be demonstrably correct without actually knowing the base URI. > >That's unfortunate, because I summarize it as no absolutization in >principle, but absolutization as something that can be layered on top of a >parser by an application. I think the proposal by refering to the URI spec assumes the properties of URIs, and uses string-comparison of absolute URIs as URI identity. The authors probably thought that was obvious, but I agree it should be explained explicitly in a non-normative comment. The proposal makes it clear that the application [reading a document] has a duty to ensure that if it does not absolutize that it does not behave incorrectly. It says, "An application is also responsible for ensuring that relative identifiers are not treated as unique identifiers across contexts as ignorance of context can make distinct identifiers appear undifferentiated" For example, an application [reading a document] could chose to track the base URI or just object when it find relative URIs which it needs to compare, or a system could be set up in which litteral string comparison is used and relative URIs are used but never between documents (ie each local to the document). To the authors of the proposal realize that there are many cases in which problems don't arise with string comparison, even when string equality of absolute URIs is the definition of namesapce identity. >>I would find it very acceptable if there were a consensus >>with this proposal (with clarifications in the wording) on this >>list. I would then suggest we take it back to DOM and >>all the groups which are looking for a way to go forward. > >I hope there is consensus, but as I find myself reading [1] with the >clarifications made in [2] and [4], I'm still not sure we're on the same page. I think the proposed solution could be rewritten to be more explicit the wording of (undefined term) "context" should be changed for wording which talks about "document URI". >>I would second Al's call for a lull in issues which do not >>pertain to this until we can declare victory on it. > >I'm finding those other issues far more interesting than the absolutization >debate, and hope they find a welcome home somewhere, whether or not victory >is declared. I think everyone wants to get this over with, but no one wants to just plaster over. I hope that we can first agree on a way forward here for namesapce URIs, and then I agree they are interesting. (There is an RDF interest group list and there will shortly be a semantic web public list. There is also a URI@w3.org list of course) You can bet I am more interested in the other issues too - but they are not all holding us up. Tim >>[1] - http://lists.w3.org/Archives/Public/xml-uri/2000Jun/0406.html >>[2] - http://lists.w3.org/Archives/Public/xml-uri/2000Jun/0460.html >>[4] - http://lists.w3.org/Archives/Public/xml-uri/2000Jun/0481.html > >Simon St.Laurent
Received on Wednesday, 14 June 2000 16:54:32 UTC