ietf
[Top] [All Lists]

Re: Pointing to IANA registries

2010-04-22 04:50:39
On 22.04.2010 11:32, Simon Josefsson wrote:
Julian Reschke<julian(_dot_)reschke(_at_)gmx(_dot_)de>  writes:
...
It is easy to fail to catch a change like that, even if you are an
author.
...

I have learned the hard way to repeat all RFC Editor edits in my copy of the draft, and only to approve publication when the only difference that is left is formatting. I also learned recently that it may be necessary to request *another* review step after NROFFing (generation of the "final" NROFF format involves manual intervention, so there's a risk new errors are introduced).

The RFC Editor modified SASL GS2 (still in RFC Editor's queue) in a
similar way: they replaced a reference to an old I-D with the final RFC,
but the reference to that particular old I-D was intentional.

How about a "IETF48" review period?  The to-be-published RFC is
published again as a new I-D and the entire community can do final
review of the changes introduced during the RFC editing process.

+1.

Right now, the I-D approved by the community and the IESG can be
considerably different from what is published as an RFC.  There is a
risk that errors are introduced.  For example, compare SCRAM and GS2
changes:

http://tools.ietf.org/rfcdiff?url2=http://www.rfc-editor.org/authors/rfc5802.txt&url1=draft-ietf-sasl-scram-11
http://tools.ietf.org/rfcdiff?url2=http://www.rfc-editor.org/authors/rfc5801.txt&url1=draft-ietf-sasl-gs2-20

There is a bunch of quite technical changes in there that we really want
to get right, and the review of these modifications have been limited.

In some cases there's also a considerable amount of fine-tuning post-IESG approval (*), both with individual ADs to get them clear their DISCUSSes, and with the RFC Editor team. For transparency it would be good if this was more visible.

In particular, the idea of attaching "RFC Editor instructions" instead of just providing a clean new ID is ... odd.

Best regards, Julian

(*) It's probably always well-intended, but there's a risk of making changes after the community has stopped paying attention.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf