- From: Williams, Stuart <skw@hplb.hpl.hp.com>
- Date: Tue, 9 Apr 2002 13:38:53 +0100
- To: "'David Orchard'" <dorchard@bea.com>
- Cc: "'TAG'" <www-tag@w3.org>
David, > -----Original Message----- > From: David Orchard [mailto:dorchard@bea.com] > Sent: 08 April 2002 17:54 > To: 'TAG' > Subject: RE: [namespaceDocument-8]: format for machine > readable content > > Stuart, > > Perhaps what the TAG should do on issues that require new or changes to > existing specifications is to proceed to create proof-of-concepts, and then > hand the work over to another WG. In this case, XML Core seems a likely > place to go. > > A Process along the lines of: > 1. Here's the principles of the web as we see them, including > a particular delta. > 2. Here's a POC of a solution to the delta > 3. Can you (XML Core) do the work of making this into a "real" > specification? > 4. We (TAG) volunteer to have at least some of us work with > you on this. I like the spirit of this. > My biggest concern is that I don't think the TAG should be precluded from > writing syntax. IMHO, that would be a terrible mistake. I agree with that too... need to give some thought too... I guess that explicit TAG member participation (because of TAG membership) would be as an invited expert as opposed to participation as normal WG members (because of member company interest). > But the process I suggested about would address the fact that > some group is going to have to do the real work and ongoing maintenance. :-) > From a process perspective, I don't see the difference between an > Architecture Recommendation and a Recommendation. It seems that the > difference is that an Architecture Recommendation has a broader scope. > So Namespaces itself would be considered an Architecture Recommendation - we > seem to believe that Namespaces is a fundamental principle. Not sure.... the Namespaces REC gets to the concrete syntax of how to declare and use XML Namespaces in namespace documents. I think it is less clear architecturally about the properties of a namespace. The main architectural principle in the namespaces rec is that NS's are named by URI. Areas where more architectural clarity would be useful about: - The (intended) relationship (if any) between terms that inhabit the same namespace. - The (intended) relationship (if any) between terms that inhabit different namespaces. - The namespaces inhabited by unqualified element names and unqualified attribute names (the Non-Normative Annex) - The (intended) relationship (if any) between an XML namespace and the vocabulary of an XML application (language) - The (intended) relationship (if any) between an XML language and the XML Namespaces from which its vocabulary is drawn (eg. xlink: is intended to be used in the defn of other languages). There are probably many more like this... > Cheers, > Dave Regards Stuart --
Received on Tuesday, 9 April 2002 08:39:14 UTC