[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Google Summer of Code 2008

On Wed, Feb 27, 2008 at 9:42 AM, Lucas Nussbaum
<lucas@lucas-nussbaum.net> wrote:
> On 27/02/08 at 00:42 +0000, Steve McIntyre wrote:
>  > Hey folks,
>  >
>  > Google are running their Summer of Code programme again this year[1],
>  > and if we want to take part again we need to apply between March 3rd
>  > and March 12th. If we're accepted as a mentoring organisation, then
>  > students will be able to apply to work with us up until the 1st
>  > April. [2]
>  Hi,
>  I have had a problem with the way GSOC was handled in Debian in the past
>  years.
>  Many of the students that were selected were already well-known Debian
>  contributors or developers. The first problem with that is that some of
>  those students used their GSOC time to work on their usual Debian tasks
>  instead of their GSOC project, leading to disapointing results,
>  unsuccessful projects, less projects being accepted the next year, etc.
>  The other problem is that some Debian developers who could have applied
>  as well didn't, because they thought that GSOC was only for new
>  contributors.
>  I think that GSOC is a great opportunity to get fresh blood inside
>  Debian, and that we should use it for that, not to get funding for usual
>  Debian work. We should have a policy of not allowing existing Debian
>  developers to apply as students. If DDs want to use GSOC to get some
>  work done inside Debian, they could become mentors instead.
>  However, I'm not sure that many DDs agree with this, so maybe we should
>  just aim for *clarification*. So any of the three following solutions
>  would work for me:
>  (1) Forbid DDs and people in the NM process waiting for FD/DAM to apply
>  as students.
>  (2) Make it crystal clear (through a mail to d-d-a) that DDs that are
>  otherwise eligible can apply as well.
>  (3) Compromise: allow current contributors to apply, but, when
>  classifying applications, do it like that:
>    1. Application from outsider
>    2. Application from current contributor
>    3. Application from outsider
>    4. Application from current contributor
>    [...]
>  What do you think?

I disagree with 1). Both 2) and 3) are fine with me.

If some projects in the past were a failure, it is solely the problem
of the management (=student's mentor:), it doesn't matter if the
student was or wasn't a DD. If the student is working on something
else (doesn't matter it is also related to Debian), his mentor should
fail him in the middle summer evaluation.

Where are the results of the last year? I only found this:


But the information about the results of each project is missing. It
needs to be clear from the beginning, that if the student is not going
to work on his project as written in his application (as a full time
job), he will be failed. And all this information should also be
available on the wiki. That wiki says Debian got over 100 applications
last year - so I am 100% sure there were many students who would
gladly work to meet their applications goals if they were given the

I suggest:

* Each application needs to be a concrete plan.
* Everyone is encouraged to apply.
* You get many applications, both from DDs and non-DDs
* you sort them from best to worst.
* google assigns N slots to Debian.
* You choose N students - you can choose the first N, but you can also
take into account their past contributions in Debian, you can take
into account that we want new blood, etc. You also take into account
if there is a mentor available to mentor the application. Many factors
influence the result.

Disclaimer: I was a mentor last year of 2 students for the SymPy
project (informally actually of 5 students, see [1]). I am in NM. And
I could be a GSoC student too, but I'll be a mentor again this year
for the SymPy project, if any students get accepted of course. :)



[1] http://code.google.com/p/sympy/wiki/GSoC2007

Reply to: