Killing old/slow groups - transition thinking

[John C Klensin]

Dave,

While I, too, would prefer to see meaningful milestones and
regular and more serious interactions with WGs who miss them, I
think you are rather significantly overstating the case for
them.  See below...

--On Tuesday, 10 December, 2002 15:47 -0800 Dave Crocker
<dhc@dcrocker.net> wrote:

> Randy> e.g., what would happen if we removed all dates from
> the milestones?
> 
> It is difficult to imagine the need for answering this
> question, since a venture into such a basic question leads to
> such other basics as "why bother to do standards."

A bit of hyperbole, don't you think?  "Why not remove dates from
milestones" doesn't seem to me to lead any more obviously to
that question than it does to the question "why don't more pigs
fly".

In particular, undated milestones may still provide considerable
guidance about what is expected to be done, and even the order
in which it is expected to be done.

> However, since you ask: The practical reality of current IETF
> operation is that that is pretty much what we have done
> already. We ignore milestones.
> 
> As a result, we have extremely protracted efforts. And the
> result often are specifications that are bloated, complex,
> and/or buggy and have lost market relevance.

As a mostly-rhetorical question, how do you get to "as a
result"?  I can assure you, from experience with both other
standards bodies and various commercial operations, that having
carefully-specified milestones, with dates and elaborate
procedures and rituals when those dates are exceeded, does not
provide immunization against bloat, complexity, bugs, or market
irrelevance.  They are a management tool that, like most other
tools, is more or less effective in different environments and
with different management strategies.

ADs who find specific dates and benchmarks useful enough will
presumably figure out ways to keep them current and enforced.
To conclude that imposing a "current benchmark" discipline on
ADs or WG chairs whom we don't feel are managing effectively
(assuming such people exist) will make them into effective
managers is pretty fanciful, don't you think?  Or should we be
telling the Nomcom that they shouldn't put anyone on the IESG
who doesn't believe that benchmarks are the most critical
management tool around because we are adopting and requiring a
new IETF orthodoxy about management tools?

> Projects that have reasonable milestones -- that is,
> milestones that balance market need, technical quality and
> market timeliness requirements -- and show intelligent
> attention to those milestones are usually called "well
> managed" because they turn out good work.

And projects that turn out good work on a timely basis are often
described that way whether they have reasonable milestones or
not.

Sigh.
     john