For anyone interested in or associated with the project/mentorship side of this, watch this presentation from the most recent linux.conf.au: https://www.youtube.com/watch?v=ydS4vXNzN0I
Summary: if you think of this purely in terms of getting contributions to your FOSS project, and in particular if you think about the amount of time spent mentoring versus the amount of time needed for a developer to write the code themselves, you're doing it wrong. It exists primarily to provide mentorship so that we'll have more FOSS developers in the future, and from that perspective it works awesomely, whether or not any individual project gets lucky and gets a pile of useful code.
This is a great organization. I've been a gsoc student 2 years ago and learned a lot from my project and earned a decent wage. Google pays hundreds of thousands of dollars for Open Source contributions done by these students.
It's worth noting that GSOC has become really expansive in the last couple of years, approving 988 projects last year alone. While that kind of expansiveness is great, I don't think it should be held as a prerequisite -- even if companies only did 10 or 20 sponsorships like this every summer it would have a really good impact.
Those sponsorships exist already -- they're called internships.
What would be great would be a FOSS-only list of internships and push for a change where a remote internship was more common, which I think is more what you're getting at with how GSOC differs from existing options.
The biggest difference is that GSOC does not support Google directly. Plenty of the projects Google supports this way are not even useful to them at all: for example, a silly 3D game I used to play was part of the program for three years. Internships are great, and there are some open source companies that have really neat ones (Mozilla comes to mind), but most small projects run by volunteers can't afford to pay for an intern themselves.
I love GSOC and I really, really wish that more companies would do something similar. It is a GREAT way to get much-needed improvements in open-source software, teach a student some real, in-the-trenches skills that will be used throughout his career (and hopefully kicking off a meaningful career in OSS and diverting from the dark path of corporate .NET/Java jockey), and get an amazing amount of development value per dollar.
The real question is, "why is Google the only company that does something like this?" Though it doesn't fit exactly the same way I could even see a similar mentorship program as part of Y Combinator or other startup incubators. There's really no excuse for companies like Red Hat, Yahoo, Canonical and others that are heavily dependent on OSS not to do this.
I completely agree with you on the merits of the program. Why don't more companies do it on the scale of GSOC? Organizing it must take multiple full-time people. It'd be great to see more programs where students get paid for their work but I want to point out that it's not the only way companies and students are both benefiting from collaboration.
I'm at Portland State and we've got a solid six-month capstone program in computer science where local companies, including several in open source, are working with student teams. I know electrical/computer has a similar program with a more hardware focus. It's not that GSOC is the only program where companies work with students, only the most visible.
Except for GSOC, the students aren't working for Google at all: they're just working on random open source projects and get a stipend from Google for it. This is very different from a company--even an open source one--working with students directly.
> "why is Google the only company that does something like this?"
Not a company, but last summer the European Space Agency organised the Summer of Code in Space. The homepage (http://sophia.estec.esa.int/socis2011/) seems to be down, but the mailing-list (https://groups.google.com/group/esa-socis) is still up. One important difference with the GSoC is that it's limited to students of European schools.
Last year I got to spend the entire summer rewriting 10K+ LOC of C into 2k loc of Perl(with increased performance!) and adding support for sqlite backends.
It was awesome.
What was I working on? Printing under linux.
Perhaps the very definition of not-sexy. Which is just fine by me since that leaves tons of unsolved challenges.
Before I would never have had the courage to contribute to a linux sub-system. Who am I to bug true unix hackers with inane questions like: "How do I build our package?" or "I'm sorry I made a mess of our changelog."!
GSoC solved this barrier. How? Because I knew I would be able to put in the time to learn the domain and be useful. And in turn they knew I would be around long enough to justify the inane questions.
I am very thankful for GSoC, and I think my new co-contributer, and former mentor, is as well.
This is an awesome program. I did it last year, and having it on my otherwise empty resume really helped in the internship hunt this year (I don't know if they used my GSoC evaluations or anything during the process, but I ended up landing a Google internship!).
I am considering applying as a mentor for an open-source project I maintain. The project - an Android app - has 30,000+ users, but the development team is basically me and a handful of folks who have submitted patches or are willing to mentor. Is this too "small-potatoes?"
I wasn't able to find a link for what the actual application will entail. Where can I find that?
Any tips, thoughts, or advice for people who have previously administered or mentored a project - particularly for small groups?
I have no idea about how GSoC process or mentoring works, but I just wanted to say as a potential GSoC student it sounds like a cool project, and I hope that things like that are out there, whether yours or someone else's.
Summary: if you think of this purely in terms of getting contributions to your FOSS project, and in particular if you think about the amount of time spent mentoring versus the amount of time needed for a developer to write the code themselves, you're doing it wrong. It exists primarily to provide mentorship so that we'll have more FOSS developers in the future, and from that perspective it works awesomely, whether or not any individual project gets lucky and gets a pile of useful code.