Re: [media-types] image/svg+xml+gzip

[Mark Nottingham <[e-mail]>]

Hi Murray,

Howe we arrive at an answer to that situation is pretty straightforward -- that media type needs to go through the registration process, and if it fails to do so, 9399 needs an errata filed against it. That will take some time, because the discussions here might or might not end up supporting such a registration; if they don't, we might decide to make an exception and 'grandfather' the registration.

We should not presume that just because it failed to be noticed during RFC processing, it can't be questioned; doing so would create an incentive to surreptitiously introduce new registrations into RFCs in the future.

As fodder for the discussion here, +xml+gzip is a good illustration of a pseudo-processing model that can be problematic. Does unzipping it produce something that's +xml? What if that media type isn't registered? How does this interact with Content-Encoding (in HTTP) and Content-Transfer-Encoding (in e-mail)? Do other +xml media types need to also register +xml+gzip? Why is something like image/compressed-svg (not +xml, because it's no longer XML) not sufficient? Etc. 

Cheers,



> On 4 Apr 2024, at 05:36, Murray S. Kucherawy <[e-mail]> wrote:
> 
> Colleagues,
> 
> A situation has been brought to the IESG's attention for assistance, and it falls squarely into the media type suffixes problem space.
> 
> The above-named media type has been referenced in RFC 9399.  It was, however, never registered, and this document neglected to notice this and include the registration.  So now we basically have a proposed standard squatting on a code point.  Even better, the registration includes multiple suffixes, and we don't know exactly what that means, which is one of the things MEDIAMAN was chartered to specify.  We've been asked to sort this out as the type is apparently starting to get used.
> 
> So this adds a dash of urgency to the question we discussed at the mic at 119 and continued discussing when Meetecho came to tear down the A/V setup, resulting in the most recent threads.  The IESG was asked to proceed with the registration on the basis I described above, but I'm inclined not to do so until this WG has published (or at least the IESG has approved) the suffixes BCP; it doesn't make sense to me to potentially derail a piece of chartered work especially when it seems like we're converging on a solution.
> 
> I don't know if that changes the calculus or the momentum, but I wanted to make sure it's part of the equation.
> 
> -MSK, ART AD
> _______________________________________________
> media-types mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/media-types

--
Mark Nottingham   https://www.mnot.net/