[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: 24 March conf call minutes
OASIS Regrep TC Conference Call Friday 24 March 2000 Minutes Attending: Terry Allen (chair) Lisa Carnahan, NIST Robin Cover, ISOGEN Yutaka Yoshida, Sun Norbert Mikula, DataChannel Nagwa Abdelghfour, Sun Una Kearns, Documentum Attending as visitors: Ron Daniel, MetaCode Ernie Nishiseki, Dun & Bradstreet Jamie Walker, Boeing NIST is still coding, will have comments on outstanding issues soon. EBXML-Regrep, in its conference call, discussed deliverables for the Brussels meeting and the suitability of UML as a modelling tool. XML.org is going ahead on their implementation; there is continued unclarity about the difference between related and associated data. Lisa promised to write something up on the NIST approach (N.B. this was subsequently clarified in email; association are between data elements and composites; related data is documentation and the like). URNs: Ron suggests that we reuse URN name spaces as much as possible. GUIDs are okay, too; however, it's cleaner all round if GUIDs are expressed as URNs. Microsoft is the author of the GUID spec; Ron will try to dig up an old GUID URN draft. Review of open questions re email of 11 March: Sat Mar 11 15:15 77/3258 "eliminate regrep subm" Terry proposes saving the doctypes and not the packaging method; Lisa proposes leaving it open; we agree on saving the semantics but they need more work. Terry to make a proposal. Extensions: Terry proposes that instead of kitting out the DTDs to support extensions, extended versions of the specified metadata be served as responses to the I2C request, and that (telescoping some discussion) we *do* specify an I2X request that will always return the metadata record specified by OASIS (with contact data suppressed if appropriate, to avoid spam). Ron points out you can add a link to "more info if desired", which, if traversed, can lead to a check of credentials. Sat Mar 11 15:45 71/3230 "individuals as SOs?" We agree that individuals who act as contacts should identify as their SO the largest unit that can coordinate internally. Multiple contacts may be desirable (if they can coordinate), and it should be possible to identify the parent org of the unit (although this can get exceedingly messy, we can handle the simple case). Should we allow multiple contacts per org? NIST's design allows that, but only one would be active at one time. We should add the possibility; it will be up to business rules how this information is employed. Sat Mar 11 15:45 70/2300 "implementors: types " No additional types of related data were identified. Sat Mar 11 17:00 166/7977 "the simple case and t" Much discussion. Terry and others are of the view that if something is registered it must have its own metadata and be a first-class object. Lisa observes that IMS is flat and doesn't have registered related data. The issue of whether to allow related data references to data that isn't registered anywhere remains open. Lisa also remarks that at a recent NIST-hosted conference she encountered AIIM (document imaging group) and NISO, which are both working on registry specs. Laura says AIIM and OASIS have made contact. NISO - National Information Standards Organization; Terry will check it out. regards, Terry
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC