NOT "inciting to riot" (was: Last Call: IP over MIME toProposed Standard)

[Kurt D. Zeilenga]

At 09:29 PM 5/25/2003, Keith Moore wrote:
>> What we need to do is come
>> up with clear rules how things progress (and perhaps what is & is not
>> appropriate for IETF standardization); a simple and straightforward 
>> way to resolve disputes; transparent leadership selection process and
>> transparent decision making.  If we had all of these, we probably would
>> be about 80% done (in my opinion).
>
>I think it's more like 30%.  The thing that we really need to do most is
>to fix how working groups operate.

I would have to agree 100% with this.

I see two major problems with working group operations.

Well, the first, I guess is the mis-conception held by some that
you need a WG to develop a standard track document.  Some of
the most successful standard track documents we have were
developed on an individual basis.  SASL is a good example here
(we're now revising it within a WG, but the original work was
done individually).

The second is the tendency of WG not to thoroughly consider
alternatives offered to them, especially alternatives which are
fundamentally different in design.  This, I think, is due to
WG not addressing discussing fundamentals (what's the problem,
what are possible solutions, what applications do we serve,
what are their requirements, etc.) early on.  We often like to
jump right into solutions.  That's fine if there is implicit
agreement on fundamentals but its bad if there isn't.

Note that I am not a fan of requirement documents... but
I do think it is wise for working groups to have some list
discussions on fundamentals early on.  (Guess that's a third).

Kurt