- From: <keshlam@us.ibm.com>
- Date: Fri, 19 May 2000 11:17:16 -0400
- To: "Anthony B. Coates, ESL +61 2 9373 7972 Sydney" <tony.coates@reuters.com>
- cc: xml-uri@w3.org
>When are two URI's the same? When they're the same string. Really. >When do two relative URI's refer to the same thing? There's no such thing as a relative URI. Relative URI _References_ are the same when absolutizing them in terms of the appropriate base URI yields the same string. However, two URI References whcih are not "the same" may refer to the same data after they have been canonicalized. For example, expansion of character escapes, or relative references embedded deeper in the URI, or server-side processing of the URI (directory links and so on) may all cause two URIs which are formally different to point to the same resource. >If they are compared as strings, surely the spec should say >that namespace identifiers are strings, and are not URI's, even if >they often look like URI's. We've been told by folks who participated in the Namespace design that this was the intent. The statement that they're to be compared literally confirms this. The statement that the values "are" URIs was an editing error. The debate now raging is between those who want to correct the conflict by preserving the intended meaning, those who want to correct the conflict by adopting the typo, and those who want to eliminate the problem by forbidding any syntax which exposes differences between the other two positions. That should give you enough background to start evaluating the conflicting demands and proposals...
Received on Friday, 19 May 2000 11:17:43 UTC