Core Developer application for Michael Casadevall (NCommander/sonicmctails)
sonicmctails at gmail.com
Mon Jan 12 21:18:02 GMT 2009
Steve Kowalik wrote:
> Michael Casadevall wrote:
>> Although I'm saddened by the -1, I would ask you to please clarify
>> both points. Can you please tell me that if now is too soon from MOTU
>> (which is a little less than two months ago at time of written), then
>> how long is an acceptable timeframe?
> A recent example is kde4bindings on ARM -- yes, you completed the work,
> and I'll certainly agree it's impressive -- but the journey you took to
> get a working upload was a long one, which involved multiple people,
> multiple times, due to silly mistakes and you not being 100% careful,
> which shows me that you aren't ready yet.
One of the greatest difficulties with doing this was that it was at the
time impossible to build these packages on the ARM hardware I had
available. With a 233MHz/32MB IXP422, it took almost 45 minutes to run
CMake alone, the build itself would have taken weeks. The best I could
do is build the module of the build that was failing, and then hope for
I tried to obtain faster build hardware, or access to porting boxes,
both without success; I eventually found a way to offload the builds to
get build times down to a day, but in a difficult situation, I was doing
the best I could. My sponsors were fully aware that some of my patches
were not fully test built on arm, and I even cleared it by the KDE devs
that this would be the case, and I would be fully testing each patch on
x86/amd64 to ensure nothing broke with my fixes.
> You stated in your MOTU application that one of your
> focuses has been on the PowerPC port -- it seems to me, at least in the
> past two months, that PowerPC has taken a backseat to your new love of
> the armel port -- is that trend likely to continue, and what work have
> you done on the PowerPC port specifically over the past two months?
Not a lot, since from a general perspective, the PowerPC port is
actually in fairly decent shape. I'm working with Luke on updating the
kernel, and to provide a more permament solution to the ports kernel issue.
> As a core-dev, this means that the newer projects are likely to directly
> affect daily builds, and the quality of the release for a large segment
> of the users. Universe is somewhat shielded from this, of course, as
> many projects don't pull directly from it, and most users don't install
> all the packages in universe.
What does having upload rights to main have to do with this? Plenty of
people contribute to projects in main who are not core devs, and we have
core devs who currently don't do anything in main.
> It's all very well and good to say that you can step down before burning
> out, but stepping down from things that you have started as a core-dev
> often has a considerably larger effect on the wider Ubuntu Community, as
> so many more people are exposed to your work in the first place. This is
> why sustained contribution in the first place is so important -- because
> Ubuntu, like every other software project, tries to minimise people
> walking away from the project, and parts of the project suffering, or
> even being removed completely, as a result.
I don't disagree with you here, and your right, having someone who was
working on a project die on that project is a bad thing, something I am
always working to try and avoid on my projects, I still fail to see what
having upload rights to main has to do with this. The only difference
between a MOTU and being a core developer is I don't have to pester core
More information about the Motu-council