[rfc-i] ID-Nits obsoles/updates vs mention in Abstract

[barryleiba at computer.org (Barry Leiba)]

We've had that discussion within the IESG, and have mixed results on
it -- so, yes, it's enforced somewhat randomly.

But only somewhat, because the bottom line is that if we (for some
collective value of "we") think that the information is useful to have
in the abstract, we ask for it.  My personal opinion is that it's not
usually important in the abstract, but there are times where it makes
sense (in an RFC that makes obsolete a very well know earlier RFC, for
instance, where the earlier RFC's number is so well known that saying
it explicitly is good).  For example, RFC 2822 says this in its
abstract:

   This standard supersedes the one specified in Request For
   Comments (RFC) 822, "Standard for the Format of ARPA Internet Text
   Messages", updating it to reflect current practice and incorporating
   incremental changes that were specified in other RFCs.

Somewhat random or not, I'd prefer not to have any firm rule on the
matter, but to leave it to judgment.

Barry

On Fri, Jan 15, 2016 at 7:45 AM, Julian Reschke <julian.reschke at gmx.de> wrote:
> Hi there,
>
> triggered by a discussion with the production center about RFC7749-to-be, I
> did some statistics about whether RFCs that obsolete/update something
> mention that in thee Abstract (this is something ID-Nits asks for).
>
> (The statistics were done on AUTH48 versions of RFCs, so it doesn't cover
> all RFCs because not all of them had XML versions to begin with, and some
> are malformed).
>
> I'm attaching the raw data and the XSLT to produce it.
>
> Looking at the results:
>
> 393 RFCs had obsoletes/updates fields.
>
> 281 of these mentioned at least one of the obsoleted/updates RFCs in the
> Abstract.
>
> 130 of these did not mention at least one of these in the Abstract.
>
> (due to the way I count, the two numbers do not exactly add up, because
> there can be a mixture within the same document)
>
> Looking for a trend...:
>
> RFC5*: 92 total, 45 without, 53 with mention
> RFC6*: 135 total, 48 without, 95 with mention
> RFC7*: 153 total, 31 without, 126 with mention
>
> So there is a trend (towards following ID-Nits' advice), but it's obviously
> not enforced.
>
> I have no strong opinion about whether it *should* be enforced, although my
> first reaction was "unnecessary duplication of boilerplate".
>
> However, what's not good is if this policy ends up being enforced somewhat
> randomly (and it appears that's what's happening right now).
>
> Best regards, Julian