On Fri, 2006-14-04 at 14:01 -0700, Brion Vibber wrote:

Google's doing their Summer of Code program again this year. We missed last years', but I'm going to try signing us up as a mentoring organization this year.

Great! I think SoC is good for MediaWiki for a couple of reasons. First, hey, free code. Second, I think getting the intern process straight for the SoC would make it easier if the Wikimedia Foundation ever wants to stimulate development by giving the same kind of grant packages. (The idea's been floated before, I don't know what the status is right now, just thought I'd point out that SoC might be good practice.)

Feel free to put any relevant-sounding project ideas here: http://meta.wikimedia.org/wiki/Summer_of_Code_2006

I've added the list of future directions that Wikia, Wikitravel, and Wikimedia developed in LA last month. I think there are a lot of hot, juicy projects on there that a smart student might want to take a swipe at.

I want to quickly run through the requirements, from http://code.google.com/soc/mentorfaq.html#2 :

1. A pool of project ideas for students to choose from. Check. Link is above. 2. Someone available to review student "blue-sky" proposals. Uncheck. I'll volunteer here if needed. 3. Someone available to decide which applications should be accepted. This is more of a lead developer job, albeit with input from the community and other developers, so I think it'd have to be brion or Tim. 4. A person (or people) to monitor the progress of the students. Unknown. Again, I would be happy to give some time to this, and I think there are probably other members of the team who would do it, too. 5. A mentor ready to take over for the assigned mentor(s) . Unknown. I can volunteer to be a mentor, but I can't volunteer to back myself up. 6. A written evaluation of each student developer. I can do that.

So I think it will be important to try to engage the folks working on our projects if we want to keep them. :)

I'd strongly suggest that any new development be made as an extension. That will keep it from being problematic for our mainstream codebase.

Brion: will you be sending the application to Google?

~Evan

On 4/15/06, Brion Vibber brion@pobox.com wrote:

Mozilla's experience last year wasn't so good[1]; some of their projects didn't complete, and most of them haven't been touched since the end of the summer. Mono however had a much more positive experience[2], with most of their projects completing and being further developed and adopted after the end.

So I think it will be important to try to engage the folks working on our projects if we want to keep them. :)

The students I mentored were successes, and I think (hope!) it was a great experience for them. Two trends I've seen from looking at the successes and failures of projects, and how it may apply to your selection criteria:

1) If the contribution is an external piece (like an extension as Evan P. sensibly suggested) it may mean the student will develop in isolation. Since part of the SoC is to bring non-free-software types into this world, if the student doesn't have regular discussion with other developers or doesn't have regular exposure of their code they can falter. Additionally, as the Mozilla note mentioned, contributions that aren't part of the mainline will likely bitrot. (I don't have a solution to this; just a cautionary note.)

2) Since writing PHP has a low barrier to entry, I expect you'll get a lot of proposals that sound good but turn out to be poorly written. I urge you to consider "what else has this student done that makes us believe they'll do a good job?" in your selection criteria -- e.g., encourage them to include source code or point at other projects they've done.

I hope it turns out well for you!