On Tue, 3 Jun 2008, Josh Boyer wrote:
1) How do you do it across the Project as a whole without resorting to
a "timecard" that contributors punch.
We make sure that everything is tied as clearly as possible to FAS2
account info, and then we collect measurements where it makes sense.
Right... so if you spend 20 hours working on a bug and a patch for it,
the only thing you're going to get hits for in FAS are (assuming the
tooling is in place) the individual bugzilla entries and the CVS
commit. Those will tally totals of seconds.
You are focusing entirely too much on units of time. I'm thinking more
about units of work. A patch is valuable. If someone puts even one patch
in bugzilla, we should know about it.
2) How do you get around the fact that some people might not want their
contribution time tracked?
We make it so that they don't have to do the work, and whenever we present
"metrics", we do it in such a way that we're never penalizing, only
rewarding. No volunteer wants to hear "you didn't do enough to suit us,"
but many might like to hear "you did a tremendous amount of work that we
saw." And if that drives a bit of competition, so much the better.
That isn't quite what I was thinking. What if people literally don't
want any manhours spent on Fedora to be tracked at all? Good or bad?
Contribution tracking is inevitable given the FAS account requirement.
But time spent on said contributions can be a sensitive subject and not
everyone may want to have that number floating around somewhere.
Again, you are focusing too much on time. I don't think anyone, Jef
included, was looking for strict metrics of time. They're never worth the
3) How do you account for transient contributors? E.g. The overall
contribution time for a particular area may stay the same across two
different time periods. It could be all from the same contributor
base, or it could be spread across a bunch of different contributors
that come and go. The wiki would likely be a decent example of the
It's got to be possible to track different kinds of contributions. How
many times did "gdk" edit the wiki? Check something into a repo? Show up
on various mailing lists?
You are tracking contributions there, not manhours and time. If you
want to track contributions, just have FAS tie into
Sure. Just as soon as we can install our own open source Ohloh instance.
5) How do you quantify "intangibles" like helping people on #fedora?
Make them tangible. Collect IRC screen names in FAS2, and do some
That's iffy at best.
Why? It's data. Data is useful. It's not the pinnacle of civilization,
but it is useful.
6) How do you actively "recruit" people to areas that need help without
driving them away altogether?
By emphasizing them on the Join page, which is increasingly becoming an
effective vector into the project. Of course, the fact is that you can't
make a volunteer do anything they don't really want to do, which is
something we must be cognizant of at all times.
Right, and that is sort of my point. Knowing what areas need help is
one thing, but if said areas are perceived as boring or unrewarding then
pimping them will only be so effective. So we can point people there,
but we don't want to discourage them by not allowing them to work where
their interest lies.
Of course not.
Look, metrics are important. Data is important. It helps you make better
decisions. We shouldn't be ruled by data, but we shouldn't pass up
sensible opportunities to collect it and analyze it, either. Even the
most rudimentary data analysis can frequently tell you things you didn't
fedora-advisory-board mailing list