Killing old/slow groups - transition thinking

[Margaret Wasserman]

>>Part of this is the in the process by which milestones are set.  ADs
>>strongly encourage chairs to get work done rapidly -- to the extent that
>>(in some cases) they are encouraged to set milestones they don't believe
>>in, or not get chartered.  The seriousness of the milestones is diluted
>>from the first.
>
>this should not happen; it's just too stupid.

Agreed.

It makes sense to use push-goals to motivate work (i.e. goals that are
intentionally aggressive), assuming that the engineers involved are
willing to "sign-up" for the push-goals.  But, it doesn't make sense to
set arbitrary goals to please management.

One of the weaknesses of our chartering process is that we view the
charter as a contract between the AD and the WG chairs, when it
really ought to be a multi-party contract between the AD, the WG
chairs and the WG participants.  Participants should be encouraged
to comment on the charter and milestones, and the document editors
and key technical contributors should be explicitly "signed-up" for
the milestones -- placing some of their personal credibility on the
line.

>Lots of WGs-to-be have had negative feedback on milestones placed two 
>years or more out - in this industry, planning that far ahead is extremely 
>chancy.

There are many benefits to having a longer term plan, even if it
changes over time...  I would like to see us set goals for a
two-year planning horizon and update them periodically.

If I ran the world, all WG chairs would be expected to update
milestones within 4 weeks after each IETF meeting and to send me
a plan for their accomplishments before the next meeting, and we'd
do a full charter/milestone review of every group once a year.

I'd also appoint WG chairs for a time period (perhaps two
years, like IESG members) and try to make WG chair changes less
unusual and less emotional than they are now.

But, of course, no one has asked me to run the world -- and
there may be some good reasons why... :-).

Margaret