specific media-types for MathML3 ?

[Ned Freed]

> Le 09-avr.-09 à 17:35, Ned Freed a écrit :

> >> The Expert Reviewer for mime type registrations is Ned Freed, so he
> >> has
> >> the last say.
> >
> > That's correct for types registered through the IANA process. Since
> > these types
> > appear to be the standards tree (no vnd. or prs. prefix), that
> > requires IESG
> > processing, which is not my baliwick.

> I understand the W3C liaison will do this part.

> >
> > FWIW, in the case of IANA registered types, my job is to apply the
> > criteria
> > spelled out in the relevant registration procedures. AFAIK there
> > isn't a lot of
> > guidance in there as to when it's appropriate to use multiple types
> > versus,
> > say, a single parameterized type. So, absent a clear indication that
> > what's
> > happening are multiple types being defined for the same thing, I
> > would allow
> > all three types to be registered. But as I said, these are standards
> > tree
> > registrations so it isn't my call in any case.

> some indicated that lack of implement (or "significant"
> implementations?) could be a critical factor. is it the case?

For vnd. and prs. media types, no, it is not a factor. Indeed, since we want
to encourage consistent naming it often makes sense to get the names nailed
down very early, so we would not want this to be a requirement.

While there are a some exceptions, we generally don't look for extensive
implementation before approving documents as proposed standards in the IETF,
including documents that register media types. (Implementation experience does
come into play later in the process though.) However, this is W3C work and I
don't know what criteria the W3C use in this case.

> >> My personal opinion is that either one or three types should be fine,
> >> and that probably the WG has the most expertise to decide this, but
> >> that you should really consider what others on this list
> >> (in particular Mark) said.
> >
> > Mark's point was, as I recall, the presence of +mathml in the type
> > name. That
> > is indeed somewhat problematic and I would ask it to be changed if I
> > were the
> > reviwer. A -xmathml+xml suffix would be fine though.

> I have now drafted a first version of the forms for three types:

> W3C members can access at:
>    http://www.w3.org/Math/Group/wiki/MimeForms
> the public at:
>    http://www.activemath.org/~paul/tmp/media-type-regs-draft.html

> I have changed the subtypes to mathml-presentation+xml and mathml-
> content+xml.

Sounds good to me.

				Ned