Re: Generic processing of Fragment IDs in RFC 3023bis

Noah Mendelsohn <> writes:
> Norm Walsh wrote:
>> I find that unsatisfactory. It leaves generic XML processors out in
>> the cold once again by expecting them to be aware of all of the media
>> type registrations for all +xml formats.
> That wasn't my intention.  On the contrary, I think we're agreeing that:
>> By creating a +xml format, you're explicitly signing on to a bunch of
>> constraints. The fragment identifier constraints for XML have been
>> informally understood but not standardized for years, that's a bug
>> that 3023bis should resolve.
> Indeed.  All I was suggesting is that we insist that, at least in the
> future, +xml registrations explicitly acknowledge that.  This would
> 1) Help to ensure that the authors of those recommendations noticed
> their responsibility to support the generic syntax and
> 2) Make it somewhat harder for those who first read the +xml
> registration document to fail to notice the inheritance of generic
> fragid (and other) rules from 3023bis.
> So, I think we agree, except perhaps on whether it's worth the trouble
> to require that +xml registrations >explicitly< acknowledge the
> generic rules, and I certainly don't feel strongly about that.  Sorry
> for any confusion.

Ok. Good. To be concrete, here's what I think I'd like 3023 to say:

1. +xml media types SHOULD use application/xml semantics for fragment

2. Media type registrations for +xml media types should explicitly
   acknowledge that they use 3023 fragment identifier semantics

3. Unless a media type registration for a +xml media type explicitly
   says otherwise, generic XML processors are licensed to attempt to
   resolve fragment identifiers using the application/xml

                                        Be seeing you,

Norman Walsh <> | Individuality seems to be Nature's            | whole aim--and she cares nothing for
                              | individuals.-- Goethe

Received on Wednesday, 14 July 2010 15:27:00 UTC