Re: [RAI] RAI processes for handling work effectively

[Martin Thomson <[e-mail]>]

On 28 July 2013 03:31, Harald Alvestrand <[e-mail]> wrote:
> That's where my thinking about review teams, directorates and other
> possible constructs comes in. I think one core point here is that for
> any question asked, there needs to be some identified "someone" who's
> charged with coming up with "the answer" (or "the information needed to
> make sense of the question") - and is able to do so on a relatively
> short timeframe.
>
> Quite hard to achieve consistently in a volunteer organization, but I
> don't think it's impossible to try something that might work better than
> the WG format.

I don't think that you have the human-factors aspects of your argument
quite right here.

The desire to have some mechanism for aggregating expertise is
understandable, but I just can't get around the volunteer organization
problems.  Personally, I am quite reluctant to volunteer for review
teams or directorate for a number of reasons.  Those reasons include
the continuing, unwelcome imposition on my time (I review for gen-art
and appsdir, and on balance am happy doing those occasional reviews,
but I still dread the emails that include assignments).  I am also
unlikely to volunteer for something like an SDP directorate, simply
because I know that my knowledge is lacking.  However, if that were
the criterion used to gate entry into any directorate, then no
directorate would ever have members.

On the other hand, working groups accrue members as people want to do
work.  That produces groups of people who might only know a small
amount about the given topic, but often those people are still capable
of detecting bugs.  That said, you aren't thinking much like an
engineer if that is your goal.  Surely you can think of better ways to
test hypotheses than using an imperfect system like the IETF?

--Martin

p.s. I have little faith in the value of reorganization or other such
nonsense in light of the discussion thus far.  Is there actually a
clear, unambiguous problem here to solve?