The stuff being spewed out here by some people indicates an ignorance
of how standards work in the IETF.
Even if we wanted to, history indicates that we don't write perfectly
unamibguous and infintely detailed standards for protocols and we
don't try to enforce or even test "conformance". We develop protocols
by progressive refinement through interoperability of independent
implementations and that interoperability, in effect, defines the
protocol. If, say, there are 10 independent implementations in a
interoperability test and 7 or 8 of them interoperable, the remaining
2 or 3 implementations are irrelevant in most cases no matter
arguments they have that the standard "really" means what they
implemented. The standards documentation is normally clarified to
conform to the interpretation make by the 7 or 8 interoperable
implementations.
Perhaps some will consider this strange or wrong but that is the way
it has been. And people have, in many cases, adopted IETF standards to
get this interoperability even over other organization's standards
that were legally mandated, claimed to be unambiguous, and/or had
massive conformance testing machinery.
If someone wants to start some sort of organization to test
conformance to that organization's interpretation of IETF standards,
that's OK with me, as long as they don't pretend they are testing
anything other than their interpretation of IETF standards. And as
long as they don't claim that successful conformance testing will lead
to successful interoperability, since historic experience disproves
this.
As far as I know, the only official statement by the IETF of its
standards are the standards track RFCs (and approved drafts) that are
in force, not any "confrormance" testing by any organization.
Donald
PS: In my opinion, the oft repeated claim that the IETF "has no legal"
status or existence is hogwash. The unincorporated non-profit
association is a valid form of legal entity, at least in the USA.
From: Ed Gerck <egerck(_at_)NMA(_dot_)COM>
Message-ID: <3C4E0AC7(_dot_)993D52B5(_at_)nma(_dot_)com>
Date: Tue, 22 Jan 2002 16:58:47 -0800
To: "Marshall T. Rose"
<mrose+mtr(_dot_)netnews(_at_)dbc(_dot_)mtview(_dot_)ca(_dot_)us>
CC: ietf(_at_)ietf(_dot_)org, Einar Stefferud <stef(_at_)NMA(_dot_)COM>
References:
<5(_dot_)1(_dot_)0(_dot_)14(_dot_)2(_dot_)20020122085259(_dot_)03a98008(_at_)127(_dot_)0(_dot_)0(_dot_)1>
<v04220813b8735cc508d1(_at_)[192(_dot_)168(_dot_)1(_dot_)14]>
<039201c1a389$eb0ed770$0301000a(_at_)FATORA>
"Marshall T. Rose" wrote:
Stef wrote:
At the minimum, such violations of IETF Standards should be formally
noted in a letter from the IAB to the offending vendor, whoever that
might be, when such information becomes available to the IESG or the
IAB.
...
<snip>
the missing word in the paragraph above is "litigation", and i see no reason
why the voluntary i-* organization should willingly undertake an activity
that invites a lot of it, particularly when the market will sort it all out
(as it eventually does).
Perhaps we all agree that standardization without some sort of conformance
verification is ineffective. We need a two-prong process: standards and
conformance. The IETF takes a "hands off" approach to the conformance
issue and MTR points out why -- fear/cost of litigation.
However, suppose that the "lack of conformance note" mentioned by Stef would
not be sent to the "offending vendor" but would be provided publicly, at a
website, by an entity that has no legal status (IETF) and refers to a consensus
decision published in a RFC by a workgroup (also with no legal status) using
the open process of list consensus. Then, at the very least, the market forces
might receive an incentive to sort it out.
So, what we seem to need is an RFC series dedicated to pointing out
non-conformance
issues in implementations of IETF standards. As an added benefit, we would not
need to talk about it -- just point to the RFC ;-)
Cheers,
Ed Gerck