Re: Google Summer of Code 2009

["Justin C. Sherrill" <justin@xxxxxxxxxxxxxxxxxx>]

On Sat, February 7, 2009 8:07 am, Matthias Schmidt wrote:
> Hi all,
>
> Google recently announced that there will be a Google Summer of Code
> this year.  After our IMO successful participation last year, we should
> apply for slots this year as well.  I take the job of being a mentor and
> will help Justin with the organizational stuff (if he agrees to take
> the job of the project organizer again :).
>
> I created a first page on the web site and added some links to last
> years sites.  Stay tuned for more ...
>
> 	http://www.dragonflybsd.org/gsoc2009/
>

I'll organize.  I just got the tax forms for taking the payments for 2008
- :( .

There's slightly less slots this year, but we should be able to get in
given the right projects.  We had a good success rate (5/7), good feedback
from participants, and completed all the required paperwork on time.  That
sounds minor, but I would hazard we did better than at least 50% of the
174 other accepted projects last year.

A few suggestions: Look at last year's work to see what made it in:

http://www.dragonflybsd.org/docs/developer/GoogleSoC2008/

There's some student tips and links too:

http://www.dragonflybsd.org/docs/developer/GoogleSoCStudent/

Also for ideas:
http://www.dragonflybsd.org/docs/developer/ProjectsPage/
http://www.dragonflybsd.org/docs/developer/HackAthonTopics/

My tips based on last year's experience:

- Google likes projects where people are experimenting with something new
that will produce a specific product.  "Build a new I/O scheduler" or
"Create a graphical package management program" are more likely to be
accepted, while projects that are based on porting some existing code or
don't have a clear end product are less likely to get in.

- Plan out your work and dedicate a lot of time.  You MUST have measurable
results within 5 weeks.  You MUST plan to devote a lot of time, because
this is a job you are paid for.  If you fall behind schedule, you will
probably fail.  10 weeks isn't a lot of time.  Time management is always
crucial in software projects.