Scott,
You are right to point out that accepting reference to outside documents
should be somehow controlled. We have to be aware of IPR issues, have to
check the sanity of the document, etc. My contention, however, is that we
don't need a formal update of the IETF procedures to ensure that. Working
group review, IESG review, IETF last call provide us with ample
opportunities. The point is, it is perfectly OK to ask an editor to remove
reference [X] from a proposed standard because the review showed a problem
with the reference; the specific document status of reference [X] is much
less important. Let's be practical: we don't have to mimic the publication
procedures of other standard bodies, we just have to be rational and
consistent.
Note that we have much more control with references to info RFC than with
any other type of documents. The reason we needed a special clause for
references to external standard is that these standards often don't meet 2
important IETF criteria: they are not always available for free, and they
may evolve independently of the IETF control.
By the way, this debate should move to poisson.
-- Christian Huitema
-----Original Message-----
From: Scott Bradner [mailto:sob(_at_)harvard(_dot_)edu]
Sent: Thursday, August 31, 2000 5:43 PM
To: ietf(_at_)ietf(_dot_)org; rja(_at_)inet(_dot_)org
Subject: Re: Standard Track dependencies on Informational RFCs
We have enough real technical issues to deal with in
the IETF. It isn't constructive to create new non-technical
issues that we don't need to have, IMHO.
I do not much like the idea of establishing precedents that
can come back to
hurt us - saying (as some have in this discussion) that its
OK to make a
normative reference to an informational RFC as a principle is
a very bad
precedent indeed - informational RFCs do not go through
public review and
there is no way that all info RFCs are of the quality to be
made mandatory
to accept in a standards track document.
People seem to be focusing on the specifics of the case at
hand and ignoring
the fact that the words about normative references were put
into rfc 2026 by
poisson for a very good reason.
Now there is a way to deal with the specific case at hand -
if the reference
is to a RSA standard then the rules about referencing
external documents
comes into play. if a copy of that doc happens to be
published as an info
RFC to make it more available then that is still fine.
But I sure do not think it's a good idea to go down the road
that many in
this discussion seem to want to go down
Scott