Notes on Student Allocations
Notes on Student Allocations
Our first cut at student allocations starts based on an organization's popularity. E.g., if an organization receives 20% of the total applications for the program, then they'll receive ~20% of the student slots for the program. However, we always cap the number of applications an organization will receive at the total of their requested number of projects, since we don't want to assign an organization more students than they can support. Additional students beyond an organization's requested cap are reallocated among other organizations, again based on popularity. Regardless of an organization's popularity, we allocated a minimum of two students per project in 2007 and 2008, though some projects ended up with only one, either by request or due to lack of assigned mentors. Please note that Melange provides a base number for allocations to each organization based on number of proposals received. We take this number into consideration when doing allocations, but this is by no means our only input. The items listed below are weighted far more heavily than any number generated by Melange.
Please note that if your organization has never participated in GSoC before it is unlikely you will receive more than 1-2 slots this year. We are very careful about allocating very few slots to new organizations to ensure you have the experience with students that you can handle this summer. We consistently hear from new organizations that GSoC was *much* more work than they were anticipating and they are thankful they only had a couple students to mentor.
Once we have this base sorting done, we begin manual tweaking. We do manual tweaking based on several factors, including but not limited to:
1) Successful track record with the program, meaning most of your students passed their final evaluations. If we know that you'll deliver, chances are we'll bump up your allocation towards your requested maximum.
2) Organizations that produce committers get more students. It's important to us that students finish their GSoC projects, but it's more important to us that you folks get long-term contributors. If we heard from you that your students are still engaged and, better yet, are now full developers or committers, chances are we bumped up your allocation accordingly.
3) Umbrella organizations get more students. If an organization is acting as an umbrella for other projects, e.g. the Python Software Foundation or the Apache Software Foundation, chances are we'll allocate more students to you simply because there's a great deal of different types of work to be done, and we're hoping our students will end up with something that appeals to them among this wide offering.
4) Mailing list and IRC participation. Are you answering student questions non-stop on IRC or providing helpful advice on the mailing lists? If you are, that shows us that you're committed to the success of the program and that you're already highly engaged. Chances are that we'll bump up your allocations if we can.
5) Failure to complete your organization profile. If you didn't take the time to let us know how many students you're looking for, chances are we gave you the minimum, no matter how many you could reasonably support. Desired number of projects is critical data for us, and if you don't provide us with at least an estimate we'll estimate for you. Usually at low numbers.
6) Issues from previous years. This point refers specifically to larger, well-established organizations that we know we want to keep inviting back due to the volume of their contributions to open source. If you required more support than everyone else - and not in the "hey, this isn't documented, what do I do?" way - or more than one or two your students evaporated, chances are we dialed you back a bit.
Each time we adjust the allocation for one organization, the remaining student pool is reallocated amongst the other organizations, based once again on popularity. This next reranking can adjust an organizations slot allocations by +2 to -2 or so, but doesn't vary much more widely than that.
Once we're satisfied with our initial allocations, we publish them to the organizations so you folks can assign mentors to the best N of your applications. From there, we do more twibbling based on duplicate acceptances (usually, the organization losing the student gets the next one down on the list, but if there's no mentor assigned that a -1 on slot allocations) and on whether or not mentors are assigned (no mentor assigned in the final stages before final announcement means that your organization loses allocated slots).
After our allocations have been decided, we usually set up a waiting list for more slots. Some organizations decide based on their proposals that they can't use all the slots they've been given and they contribute some back to the pool. Some organizations decide they can accommodate more students than they have slots and put themselves on the waiting list. If you don't get the number of slots your org expected for this year, please keep in mind you can always participate in the waiting list as well.