- From: <keshlam@us.ibm.com>
- Date: Mon, 26 Jun 2000 09:17:55 -0400
- To: David Carlisle <david@dcarlisle.demon.co.uk>
- cc: david@dynamicDiagrams.com, xml-uri@w3.org
>Yes but there is a big difference between the general principle that >version n+1 of a standard will need to make some changes from version >n, and deliberately making a change now with the intention of changing >again later. I believe everyone who's in favor of Forbid sees it as the former. We do not have a solid base-principle justifying relative references to Namespaces at this time. There are some conceptual sketches whcih suggest that someone may, possibly, invent such a principle at some point in the future and successfully convince the community that it is needed. *IF* that happens, we will of course have to take this decision back off the table... but since that would constitute a clear and deliberate change in the XML environment, it would be associated with a change in XML version number, which makes it a whole heckuva lot more managable. Meanwhile, the rest of W3C would have a solid baseline to work against... and it would be the responsibility of those proposing the change to work out what the implications for all the other specs would be _before_ we resume the debate. Personally, while I grant that adopting Forbid will not stifle the debate (nor should it), I would be extremely surprised if a fully motivated and analysed proposal for relative references to namespaces is produced in less than a year, possibly much longer. We still haven't seen a good use-case for it; most of the proposals are more than adequately dealt with by other binding mechanisms. The few folks who've said they really do want the NS declaration to be relative have Big Ideas -- and maybe even interesting ideas -- but I've seen nothing even vaguely resembling a design sketch. If they can come back with such a sketch and knock our socks off, that's great and we can consider extending XML at that time. Until they do, I really think we should table the whole issue. Forbid is best in that case, both because it's simplest now and because it clearly reserves the relative syntax for this future application if and when such application is ever defined. (Much like forbidding names that start with "xml" so they're reserved for use by the standards.) I ______________________________________ Joe Kesselman / IBM Research
Received on Monday, 26 June 2000 19:36:12 UTC