[rfc-i] obsoletes/updates

John C Klensin john-ietf at jck.com
Thu Jun 18 12:53:29 PDT 2015

--On Thursday, June 18, 2015 08:30:53 +0200 Julian Reschke
<julian.reschke at gmx.de> wrote:

> On 2015-06-17 22:51, Paul Hoffman wrote:
>> The most significant change in
>> this version is to give examples of the "updates" and
>> "obsoletes" attributes of <rfc>. In the same attributes,
>> removed the ability to say that an Internet-Draft updates or
>> obsoletes another Internet-Draft, based on RFC 2026. ...
> I think that's a regression.
> There are cases where a *draft* wants to say that it
> updates/obsoletes  another one, and there's no reason to
> forbid it. The situation for RFCs  is of course different.
I'm usually quiet on this list (and rarely read it these days),
but I have to agree with both Julian and Paul (even though they
seem to be disagreeing with each other).

An I-D really serves two purposes, as a working document for
development and discussion and, at least for a subset of I-Ds,
as a proto-RFC.    For the latter, including the tracking
information ("Obsoletes", "Updates") that would be associated
with the document if published as an RFC is very important for
both information and checking.   However, for the first
function, tracking for I-D relationships may, in some
circumstances, be equally important.  I hope we never get
pedantic and petty enough to want to see that
draft-foo-bar-baz-03 obsoleted draft-foo-bar-baz-02.  However,
as documents are consolidated and move in and out of WGs, having
a standard way to note, in an I-D (not just the tracker), that
draft-ietf-bar-baz-00 has replaced draft-foo-bar-baz-05 seems to
me to be very helpful, independent of how it is marked up and

> As a matter of fact, this very draft shouldn't say "obsoletes:
> 2629" but  "obsoletes: draft-iab-xml2rfcv2".

It is very unusual to have a document and its successor under
active development at the same time, but my vague memory is that
it has even happened with standards track protocol specs.
Certainly "obsoletes: draft-iab-xml2rfcv2" (or "obsoletes:
2629bis" or even "obsoletes: 2629, draft-iab-xml2rfcv2") convey
a lot more information to those trying to understand what is
going on and how things should be reviewed than "obsoletes:
2629". The latter, if draft-iab-xml2rfcv2 is published, will
simply be false.

As a general principle for I-Ds, I believe we should be striving
for maximum communication and information content, even if
getting there requires temporarily doing things that the RFC
Editor would not allow in permanent/archival RFCs.


(back to intermittent lurking)

More information about the rfc-interest mailing list