[NNTP] [Errata Rejected] RFC3977 (2004)

[Russ Allbery]

Barry Leiba <barryleiba at computer.org> writes:

> Well, the response in the rejection reason is taken directly from an
> exchange I had with Clive.  Now, he did also suggest that I might pass
> it by Russ Allbery, which I have not yet done.  Since he's copied on
> this exchange, he can comment.  Clive's point wasn't meant to say that
> your suggestion in the errata report wasn't good, but that it's not
> bringing up an *error* in the original spec.  If what you suggest is an
> appropriate change, it might be considered if there's work done to
> revise the document.

I reluctantly agree.  I believe this is quite obviously an error in RFC
3977, but RFC 3977 is also quite explicit that 423 is not a permissible
response code for ARTICLE without an argument.  My understanding is that
errata can't be used to correct the clear statement of the specification,
even if that clear statement is wrong based on existing deployed code and
consistency of semantics through the specification.

> I'd be happy, if Clive, Russ, and Julien all think it's the right
> approach, to change the erratum to "hold for document update".  Russ
> Allbery, do you have a comment on this?

Yes, please.  We would clearly want to fix this in any revision of RFC
3977.  Personally, I would recommend that any implementor use 423 as
described in Julien's analysis despite the wording of RFC 3977.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>