Re http://lists.w3.org/Archives/Public/www-archive/2011Nov/0009.html
I have a TAG action http://www.w3.org/2001/tag/group/track/actions/531 to draft a document on "architectural good practices relating to registries" which is overdue. We also have a suggestion of reviewing the IAB document on extensibility http://tools.ietf.org/html/draft-iab-extension-recs and various discussion about IANA and registries. I'm hoping that experience with "vendor prefixes" in CSS might give us some insights on things that might go wrong, guidelines for establishing registries and extensibility points in languages. Nominally, there's been some favor around using URIs for naming extensibility points, presuming longevity of the ability to use a single URI for a stable meaning as documenting or identifying the extensibility point. The most prominent use of URIs as naming extensibility points is, of course, XML namespaces, but there are others (I don't have a list). Using vendor prefixes with vendor-supplied tag-names might be an alternative? But are they any more successful than using URIs? Sounds like less, actually. Larry -- http://larry.masinter.net
