Brian E Carpenter wrote:
On 2007-06-15 18:04, Michael Thomas wrote:
Thomas Narten wrote:
If a respected security expert (one who has reviewed many documents,
contributed significantly to WG efforts, etc.) comes to a WG and says
"there is a problem here", but 5 WG members stand up and say "I
disagree and don't see a problem", do you really expect the security
expert's opinion to be given strictly equal weight and to just be
overruled since 5 voices are greater than 1?
Context matters here. I've seen situations where the so-called "expert"
is unable to articulate the problem, is rocking a personal hobby
horse of theirs, is expressing a general philosophic point without much
attention to the actual problem (see "unable to articulate"). Etc. Also:
it may just be me, but this gradual creep toward "experts" having a
named, and different status is bothersome ("expert review"). It sweeps
under the rug that this is really a continuum of cluefulness,
experience,
as well as frankly political considerations, some of which are more
odious than others.
Mike, just to try clarify the "expert review" issue. That term comes
up specifically for IANA assignments that are controlled more than
First Come First Served and less than RFC Required. There's a full
discussion in draft-narten-iana-considerations-rfc2434bis but my
short version is: we can't expect IANA to have in-house expertise
for every registry, and we don't want an IETF debate about every
relatively routine assignment. Having designated experts seems to
be the only reasonable and scaleable solution.
Yet, this not the only place where I've seen "expert review"...
I prefer personally to use a term like "specialist" for review
of drafts, or "generalist" in the case of Gen-ART reviews. You
can be a specialist without being an expert :-)
which I take to be what you're referring to here. Let me say that
I'm happy to get cross functional review whenever I can get it. But
the way that Thomas posed his question "should 5 wg ignorants
overrule one `expert'?" shows that there can be a downside when,
in fact, the `expert' is off in the weeds for the reasons (and more)
I mentioned above. Since "expertry" is really a continuum of
competence and that becoming a recognized "expert" is one part
clue, one part clique, and probably one part availability, we shouldn't
give their expertise _too_ much power -- especially on well worn
ratholes. Also: as this "expert review" thingy becomes more popular --
and powerful -- parceling it out to a single person has a tendency to
pave over what happens when subject area experts meet: they often
disagree. Therein lies danger for the unhappy working group on
the receiving end: you're just postponing that fight for later in the
process. Worse is when the "expert" demands satiation of a favorite
hobby horse, typically of the kind that other experts find either
uninteresting, or not worth the fight since there's no skin off their
nose.
So getting back to Thomas' question: as always with engineering,
"it depends.", and I hope that those sending out the "experts" are
cognizant of that as well.
Mike
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf