Subject: why bother planning? (was Re: Bugzilla comments)
From: Paul Rohr (paul@abisource.com)
Date: Mon Mar 12 2001 - 11:11:43 CST
At 02:36 PM 3/10/01 +0100, Jesper Skov wrote:
> And while the time estimate could be useful for planning, it
> also only makes sense if you have committed time - I give hours
> to the project when I have some to spare and feel like it. You
> can't make any planning based on contributions like that.
> [i.e., you can't sum the estimated hours for bugs scheduled for
> next milestone, divide by 40h/week per developer and set a
> release date from that - so why bother?]
Huh? Your description of the situation is quite apt, but that's entirely
the wrong conclusion to draw. It's *always* useful to know that feature X
has a minimum of 40 hours of effort remaining before it Just Works.
If a single person had 40 hours of time available to devote to that task,
then you might be able to expect it to Just Work in 40 hours. (Anyone who's
spent much time doing commercial development is drowning in laughter right
now.)
And sure, we're not in that situation, but that means it's even more
important to have the estimates. If we know that a feature can't Just Work
until we get 40 hours of collected effort on it, then we either need to
recruit more folks to help get it done (always my favorite tactic), or else
admit to ourselves that the feature will not Just Work until some later
date.
To conclude that we just don't want to *know* how much effort that would
really take is far too ostrich-like for my tastes. Surely you don't mean
that?
This is just about collecting and communicating reliable status information.
Making the policy decision of actually choosing which work to include in a
particular release is pretty much orthogonal to this. If you're going to
say you don't ship the next release until feature X just works, then that
happens when it happens. If you're going to say you're shipping on or about
the 17th of never, that whatever work makes that cut makes that cut. This
information is useful for predictably implementing *any* such policy.
Paul
motto -- the truth will set you free
This archive was generated by hypermail 2b25 : Mon Mar 12 2001 - 12:52:19 CST