- From: Bullard, Claude L (Len) <clbullar@ingr.com>
- Date: Fri, 13 Sep 2002 13:31:38 -0500
- To: "'Norman Walsh'" <Norman.Walsh@sun.com>, www-tag@w3.org
And that effectively keeps the can of worms closed, or says, if you open it, you have to cut the bait. I don't disagree. That is, as best as I can remember, why it came down to version attributes being the least disruptive way to go. Again, none of this works for aggregates unless the aggregate version number specs the language versions so aggregated. One could resort to a morphological version spec that cites the versions of each included language and concats these into a string as the value of the version att. Ugly, but not an unknown way to go about it and better than a version att per agg'd lang. len From: Norman Walsh [mailto:Norman.Walsh@sun.com] From the point of view of applications that examine the namespace name and local name of an element to determine the vocabulary to which it belongs, it isn't simply a "should never", it's a "can never". To those applications, if you change the namespace name, you haven't changed the namespace of the vocabulary, you've *changed the vocabulary* by definition. That may be a lot more painful than you'd like for a transition from V1.0 to V1.1.
Received on Friday, 13 September 2002 14:32:11 UTC