ACTION-304
Write up issue around normative references to particular versions of specs
- State:
- pending review
- Person:
- Larry Masinter
- Due on:
- 2009-09-30
- Created on:
- 2009-09-23
- Associated Product:
- HTML 5 review
- Related emails:
- agenda for TAG meeting 8-10 Dec (from connolly@w3.org on 2009-11-25)
- Minutes of Nov 19 TAG Telcon (from ashok.malhotra@oracle.com on 2009-11-22)
- suggestion for 19 Nov agenda (from connolly@w3.org on 2009-11-16)
- please update overdue actions as input to Dec f2f agenda (from connolly@w3.org on 2009-11-09)
- Minutes of 23-25 Sept 2009 TAG F2F now available (from noah_mendelsohn@us.ibm.com on 2009-10-16)
Related notes:
2009-09-25 18:09:42:
[Noah Mendelsohn]
2009-10-21 14:22:12: I think this action item was related to a comment about elaborating the
"orthogonal specifications" finding. If spec A points to spec B, does it say
"B version N" or "B version N or later" or "any version of B"?
The question around HTML5 in particular was the choice of pointing to specific versions of MathML, RDFa, etc.
I think this is quite overlapping (but not a subset) of the distributed extensibility issue.
My recommendation for HTML5 is that the base HyperText Markup Language document should *not* reference specific versions of URI, URL, RDFa, SVG, MathML or any other specifications. Where absolutely necessary (because of unavodiable dependencies) it can reference "version X or any later compatible version". If there's a desire to nail down what the right combination is for "HTML user agents in 2010", then a separate "profile" document could call out the specific combination of versions which should be uniformly implemented and tested. [Larry Masinter]
2009-11-19 21:09:59: [DanC]: kinda withdrawn in favor of HT's action
Display change log.