Worried about Ray (Universe)
Thursday, May 13th, 2010This post is about Ubuntu development processes. If you don't care about that, skip it now.
I think I'm coming to a sad conclusion: I'm no longer a motivated MOTU.
I joined the team in February 2009, and in my application I said:
I want to continue to work closely with the Debian teams I've been recently involved with. Also, I want to work to convince kaol (Debian's GHC maintainer) that team-maintenance of Haskell related packages is a good idea – maintainer-lock can become a problem here. If successful, pkg-haskell will be another exciting area for me to work in.
Agda (together with its standard library) is crying out for a package.
This all happened. These goals all related to doing stuff in Debian. I also said this:
If bestowed with MOTUhood, I'll definitely join u-u-s and do my share of sponsorship, having helped to fill the queue myself in the past. Backports is also an area that I will consider getting involved with.
This is all in addition to the "bread and butter" QA-style work such as rcbugs and NBS, which I will obviously do too.
Ah, ahem. This may have happened a bit. But not any more; take a look at my +uploaded-packages page. It's obvious that I didn't care about the wider Universe in Lucid, opting to focus on my personal areas of interest which are Haskell packages (pkg-haskell) and CLI/Mono packages (pkg-mono pkg-cli-apps pkg-cli-libs).
Does this make me a bad MOTU? Debian is our upstream, and we are always talking about sending work there. It's only natural (and zack mentioned this happening in his talk today) that Ubuntu developers over time should migrate to do most of their work in Debian directly. But when people specialise and/or go over to Debian, they obviously work less — or not at all — on Universe QA.
I've been talking to some people at UDS — stefanlsd and persia mainly — about this problem: I'm not motivated to do general Universe work any more. When I was learning packaging this stuff was a great way to become familiar with the tools and different ways that things can be done. But now I know that, and have found my niche, I don't really feel that doing random merges/syncs/NBS/FTBFS/… is a good use of my time.
But we had a session this morning on Universe QA, and it seemed like the agreement there is that we do need to do better at this. There was also consensus that MOTU as a whole isn't doing very well these days, both in terms of keeping up with the tasks that we should be doing, and in recruiting new developers to the project. I have just described why I'm not pulling my weight on this front. On a motivational front, it seems as if the archive reorganisation effort and other developments (the 'cool apps' spec) are going to diminish the role of MOTU even further.
I really like being an Ubuntu developer, and I'd love to be better at it. I'm concerned that just concentrating on my packages isn't enough. Is it? Is it inevitable that contributors will specialise? If so, who fulfils the QA function of MOTU? Maybe there are superstars who love managing random transitions.
In summary: I suck (at being a good MOTU), and I don't know if I even can get better. But if I don't step up, who else will? And how can I expect others to do it if I don't set a good example?
he;lp





