Re: wg list discipline

[Thomas Narten <narten@us.ibm.com>]

Hi Erik

I've cc'ed the iesg both because I'd be interested in other
thoughts/suggestions/warnings, and because the problem of
out-of-control mailing lists came up today in a completely different
WG.

Erik Guttman <Erik.Guttman@sun.com> writes:

> I had to write Keith 3 separate emails in the past couple days as a
> result of his overheating on the list.  If I have to go further than
> that, what is the procedure for formally warning a WG participant that
> they have to either shape up or get excluded from posting?  If he
> persists, how can I kick him off?

Dealing with Keith (and folk like him) on zeroconf is definitely a
problem. His ability to post faster than normal folk can read is in
fact a DOS attack. As you know, other people have said as much
privately and it makes following the mailing list daunting and is
keeping folk from participating. I count something like 120 messages
in the last two days on the zeroconf list. No one can reasonably be
expected to keep up with this sort of traffic.

I don't believe the IETF has ever censored or tried to limit the
postings of anyone who was actually also considered a real
contributer. So doing so will be tricky and also chart new
ground. However, I suspect there is sympathy for doing so in extreme
cases, especially when not doing so threatens the ability to get
actual work done. IMO, this is something the IETF needs to grapple
with, as it's one of the factors that is leading to an IETF inability
to function. zeroconf may just be a reasonable WG to experiment with
some new approaches.

One thing to keep in mind is that a WG chair can get away with just
about anything -- so long as the WG supports you in the end. So, if
you want to limit postings in some way, make sure you have WG backing
and that the WG will defend what you are doing. 

Some thoughts.

Limiting the number of postings by any one individual to some number
per day might be useful, in that it can force folks to think a bit
more prior to posting. Or, they will just post a single 500 line
message per day. :-( You'd also need to make exceptions for a small
number of folk, like the chair and document editor though.

Moderating the list (as in all postings must be manually approved)
might be useful. Not to moderate, but explicitely to slow down the
interaction rate. If in fact, postings take a few hours to appear, the
result may well be that folk think a bit more before posting.

Advantages of the above type schemes is that they don't single out any
one individual, so they will be viewed as more fair.

If you really feel like you need to single out individuals, I'd
suggest a system with some explicit warnings, plus a way for the WG to
provide confirmation/feedback if they believe you are being too
hard/soft (e.g., by making the warnings public). You'll want the WG's
backing that someone is being too disruptive. And you'll also want a
way of being able to get input if your being to hard, as you don't
also want to have to immediately start dealing with process
appeals. Appeals are much less likely to get traction if the WG is in
support of any moderating activities.

I wonder if a starting point would be a note to the WG saying that
you, as chair, fear that the mailing list is in danger of not
functioning any more due to the number of postings from a small number
of individuals, and that you fear you need to do something about
it. And mention some hypothetical approachs to just see what the
reaction would be. You'd like some feedback (public or private) as to
whether you are right to be concerned, etc. Of course, such a thread
will be a short-term distraction, but in the long run might turn out
to be very useful investment.

Thomas