Query to the group: If we believe we should not hold working groups to
their milestones, why bother to have those milestones?
It's useful for a group to have a sense of direction and an anticipated
timetable even if there is no penalty for changing the direction or failure
to meet that timetable. it's just like having a shared vocabulary - it
helps get people on the same page.
I don't think it's very useful to phrase the question in terms of whether
or not to have milestones or whether to hold groups to those milestones.
My first attempt at a better question would be:
Why do our expectations of a group's progress so often fall short?
Some guesses:
- we often unrealistically expect that the energy that goes into creating
the group will continue through the group's lifetime, so we pick
timeframes based on that level of energy.
the initial energy tends to run out within a year or so. often, putting
the finishing touches on a document takes months or even years because
by the time this happens most of the participants are so burned out
on that topic, and have started to work on other things, so they can
barely devote any attention to it.
- in choosing timeframes, we often overlook the difficulty and time
required to solve fundamental problems in a protocol, especially
security issues.
- there's a reluctance to pick realistic timeframes because of a fear that
the market will fail to wait for a standard solution if it's not expected
for two or three years, even if that's a realistic timeframe.
- once the group's charter has been approved, hardly anyone in the group
ever looks at the charter again. most of the participants' attention is
focused on the day-to-day mailing list traffic, not on the big picture.
- we have unrealistic expectations about the amount of time required
to do external review (last call, IESG, etc), and to revise the
documents based on the results of such review.
- some groups don't limit the number of documents that they take on,
failing to recognize that every additional document takes energy
and meeting time away from the group.
Not meeting milestones is more of a symptom than a problem in itself.
The best way to relieve the symptom might be to focus on some of the
above problems. Of course, groups that aren't making progress should
still be killed or reorganized, but what we really need to do is to find
ways to make the group more effective.
Suggestions:
- limit the goals: charter most groups to only do tasks that can reasonably
be completed within a year, or 18 months at the most. allow a six month
extension when necessary, but expect that the group will *shut down*
(not be rechartered) after this period.
- the charter should specify which documents the group will produce.
every additional document should require a charter change with AD
approval *before* the group starts working on it; until such approval
is obtained the documents are individual submissions and outside of the
scope of the group's work.
- try to identify likely solutions to the fundamental problems *before* the
group is chartered
- have the chair (or the AD) periodically send out the group's charter and a
status report to the mailing list, to remind people of where they are.
- explicitly include last call and IESG review of each document in the
timetables. also include at least one document revision cycle per
document following such review.
Keith