Developer Application Criteria - Was Re: New Application processes
laserjock at ubuntu.com
Thu Jan 8 17:15:56 GMT 2009
On Thu, Jan 8, 2009 at 7:16 AM, Dustin Kirkland <kirkland at ubuntu.com> wrote:
> On Wed, Jan 7, 2009 at 6:22 PM, Robert Collins
> <robertc at robertcollins.net> wrote:
>> I completely agree. MOTU and core-dev membership is a combination of
>> * technical knowledge [for which two key points apply: arbitrary
>> have-done-X metrics don't assess any more reliably than peer assessment
>> of the work done, and the knowledge ages rapidly as technologies change.
>> Packaging of python today is not the same as it was 5 years ago].
>> * trust - which is entirely subjective
>> * fitting in the team - which can be assessed by who objects :)
> Okay, so I'll restate my point in another way ...
> If there is *no* objective component to MOTU/CoreDev application
> assessment, then I don't think it's fair to make arbitrary,
> *quantitative* criticisms on an application.
I don't think that's necessarily a logical conclusion. You're saying
that if the +1/-1 of a MOTU Council member is based on a subjective
decision that they can't use objective data in making that decision.
> Criticisms of the form:
> * You haven't done enough merges
> * You haven't touched enough Universe packages
> * You haven't been a MOTU or Developer long enough
> * You haven't ... enough ...
> should be invalid.
Why? There's no a priori reason to reject such reasons just because
they are objective data. What you're really arguing against is that
objective data is being used to make subjective decisions and I don't
know why that would be a problem pre se.
> These things are actually measurable, and we could very well set the
> value of "enough". If we are consciously choosing *not* to set these
> values, I think it's totally unfair to criticize someone for not
> achieving these arbitrary, dynamic, mystery thresholds.
I very much doubt that "enough" could be easy to set. The fact of the
matter is that each MOTU will see that differently, and the point of
having a MOTU Council is to take in all the data and determine if
"enough" has been met. If it was easy to do we'd just have a secretary
go down a checklist to make MOTUs.
My overall feeling with this thread is that perhaps we're attacking
the wrong problem. People are wanting to reject the criteria and the
subjectivity of a mostly subjective decision. I think rather the
problem is that the MOTU team has lost its collective understanding of
what being a MOTU is. The main argument has been "my sponsors told me
I was ready and I got rejected". I think there are some reasons why
this could be happening:
1) as MOTU has scaled we've lost a lot of the connection to the "whole".
2) contributors are perhaps not seeing as much of "what does a MOTU
actually do" as it has gotten into a lot of "paperwork" type tasks
rather than tackling tough packaging problems.
3) there is quite a bit of variability in the skill and experience
level of MOTUs. We have "old school" MOTU who may have a fairly
different view of what a MOTU should be from what a newly minted MOTU
might think, for instance.
4) people seem to take the MOTU applications much more personal
these days. "rejected" "criticized" seem to come up a lot when it's
always been "not yet, but keep trying". This perhaps is due to losing
some of that Ubuntu Ethos Jono has been talking about.
I think perhaps effort would be better spent trying to figure out how
to make MOTU a *real* team again rather than trying to figure out how
to objectify an essential subjective decision. People should be
tapping into collective knowledge and understanding rather than
reading page after page of docs and running through checklists. We
should be a community rather than a MOTU mill. The self-assessment
checklist might help people figure out what we're looking for but I'm
afraid it would quickly turn into "but I can do the checklist, you
have to let me in".
More information about the ubuntu-devel