Re: [apps-discuss] New Version Notification for draft-wilde-xml-patch-00.txt

[Ned Freed <[e-mail]>]

> hello ned.

> On 2013-01-14 20:33 , Ned Freed wrote:
> > Since in fact we do have such a registry now, you can take that as given.
> > And the use of the suffix is now a SHOULD; you can only not do it if you
> > have a
> > very good reason for that choice. I personally don't think that name
> > consistency with other types is even close to rising to that level; YMMV.

> http://tools.ietf.org/html/draft-ietf-appsawg-media-type-suffix-regs-08
> is what you're referring to and this one is going to get published soon,
> right?

When it gets published is irrelevant. It's already approved as a standard
and specifies the current rules we operate under. The relevant IANA web
pages already cite it.

I'll note in passing that this use by IANA and others of documents that haven't
completed the publication process is not up to me; it's just how things work.

> in that case you'd recommend to go the xml-patch+xml way for the
> subtype, i guess?

Yes, that's what I would recommend, but really, it's not a question of what I
recommend. It's instead a question of what the standards require. I already
stated the requirements; if you believe you can make a compelling case for
omitting the suffix for this type you are free to do so.

> i am really wondering how this is going to play out for RDF, have their
> been discussions on how to handle this case (multiple serializations of
> the same data model) for the suffix registry? the reason why i am asking
> is that RDF needs a patch type as well, and there it will become an
> issue how to handle the fact that usually, a patch type's semantics are
> targeted at a model, while the type's syntax is based on some specific
> serialization.

That a question to be answered if and when someone proposes a suffix for rdf.
AFAIK there isn't one at present. I personally don't know enough about RDF and
how it is used to comment, although I'll note that the XML format has the
application/rdf+xml assigned. I will also note that EXI was considered and
rejected as a suffix.

				Ned