Re: [ietf-types] Registration of media typeimage/svg+xml

+1. =20

Let's try to make things clear for people who are reading just the type r=
egistration.  Someone who looks at the registration separately from the R=
FC or other specification shouldn't have to guess as to what "this specif=
ication" is.

Keith

On Nov 25, 2010, at 12:46 AM, Martin J. D=FCrst wrote:

> Hello Chris, Larry,
>=20
> On 2010/11/25 7:16, Chris Lilley wrote:
>>=20
>> On Wednesday, November 24, 2010, 9:14:12 PM, Larry wrote:
>>=20
>> LM>  Martin, in a couple of places you complained that the SVG registr=
ation
>> LM>  template, contained in a W3C document, referred to "this specific=
ation",
>> LM>  and said:
>>=20
>> LM>  # "this specification" doesn't work when the registration templat=
e is
>> LM>  # taken out of the SVG spec. Either say "the SVG specification" o=
r
>> LM>  # explicitly reference a specific version of the specification.
>>=20
>> LM>  However, I think it is common practice both in W3C and IETF, that
>> LM>  a registration template embedded within another document can
>> LM>  reasonably say "this specification" to mean the document in which
>> LM>  it is embedded,
>=20
> In that context, this looks reasonable, but out of context, it does no =
longer make sense. Registration are often taken out of context, and shoul=
d be able to stand alone, without readers having to guess what "the speci=
fication" might be.
>=20
>> LM>  with the registry itself pointing to an
>> LM>  explicit version of the spec as well as the template within.
>=20
> IANA often, if not always, lists the relevant RFC. But I haven't seen t=
hem listing other specifications. I may have missed it. Anyway, that's us=
ually outside the registration, so the registration is still not standalo=
ne.
>=20
>> Larry, your interjection is timely, as I was just about to edit the re=
gistration to address Martin's comment.
>>=20
>> Given that we already have
>>=20
>> Published specification:
>>=20
>>     This media type registration is extracted from Appendix P of the
>>     SVG 1.1 specification. http://www.w3.org/TR/SVG/
>>=20
>> I'm tempted to just s/this specification/the published specification/
>=20
> Great! That makes the registration standalone, while at the same time d=
oesn't look weird inside the specification.
>=20
>> LM>  There are plenty of examples... perhaps those comments don't appl=
y?
>=20
> The fact that there are plenty of examples doesn't mean that we can't f=
ix it when we have a chance. I don't think it's worth doing back and fixi=
ng one by one, but if we are working on a registration anyway, we can mak=
e it better form the start.
>=20
>> LM>  (This comment applies to all registries, not just of media types.=
)
>=20
> Agreed.
>=20
>=20
> Regards,   Martin.
>=20
> --=20
> #-# Martin J. D=FCrst, Professor, Aoyama Gakuin University
> #-# http://www.sw.it.aoyama.ac.jp   mailto:[email protected]
> _______________________________________________
> ietf-types mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/ietf-types
>=20