Minutes of SoC meeting

[Vincent Untz <vuntz gnome org>]

Hi,

We had a meeting about Summer of Code yesterday in #soc. Here are the
minutes. We'll have another meeting next week, probably same day + same
time (Tuesday, 20:00 UTC) since there were still some topics we didn't
have time to discuss.

Note that we'll soon create the wiki page to post ideas about potential
SoC projects. We'll announce it widely soonish.

Vincent

=================

Present:
 SoC/WSOP students from previous years: Ryan Lortie, Max Wiehle,
   Clare So
 SoC/WSOP mentors from previous years: Behdad Esfahbod, Shaun McCance,
   Danilo Segan, Joe Shaw, Vincent Untz
 Others: Lucas Rocha, Boyd Timothy, Olav Vitters, Mariano Suarez-Alvarez
   Dan Winship, Larry Ewing, Aaron Bockover, Kjartan Maraas, Paolo Maggi,
   Sven Herzberg (and others I missed, but not all were active)

ACTION: Lucas to keep track of community involvement of SoC
        partipicants

ACTION: Vincent to talk with sysadmins about getting a sandbox
        repository or main repository access for students

ACTION: Behdad to ask mizmo about a poster (black and white, A4 and
        letter)

ACTION: Lucas to ping web people about putting SoC on the front page

ACTION: Vincent to send a mail to know who would like to be in the
        selection committee

ACTION: Vincent to create the mentors mailing list

Good points from previous years
===============================
 + the successful projects
 + students who stayed in the community
 + students came on IRC, and people were helpful to them
 + other social aspects (meeting at GUADEC, eg)
  - note that face-to-face meetings really help
 + SoC helps people start contributing because students have a clear
   mission and a mandate
 + GNOME SoC Planet
 + "from an outside point of view WSOP seemed more successful than SoC"

Negative points/What could be improved?
=======================================
 + we didn't keep track of community involvement of SoC students
 + disconnect between mentors and admins: which projects have been useful
   in the end?
 + not enough communication between mentors and students
  - we'll ask for weekly reports
  - they'll be sent to one list (gnome-soc-list)
  - SoC admins are responsible for pinging students who don't send
    reports
  - have a rule like "we will seriously consider your case if you miss 2
    reports"
  - we should stress that having written nearly no code in one week is
    not always an issue, since sometimes a lot of thinking is needed
 + no infrastructure for students at the beginning
  - having a sandbox repository before SoC begins would really help
  - questions whether this should be separate from the main svn
    repository or not
   . probably separate, for security reasons
  - bzr/git/etc. could help too
   . some students might not feel comfortable with such tools
   . we're using svn for GNOME, so using the same tool is probably
     better
 + selection process was a bit chaotic, since all potential mentors were
   able to vote. See 4th item in the agenda.
 + we should send mails to rejected students, pointing them at useful
   resources on how to contribute

What kind of projects do we want for this year?
===============================================
 + maybe we've accepted projects that are not urgently useful
 + performance analysis projects are interesting but might not produce
   concrete results in the end
 + projects with a high learning curve might be too difficult
 + try to avoid projects that are essential to one of our modules,
   because it can block progress of the whole module, especially if the
   student has difficulties
 + we want a mix of immediately useful projects and pie-in-the-sky
   projects
  - 20-30% of experiemental/weird/cool stuff and 70-80% of needed features
    (for example)
  - "something in roadmap" and "something unexpected" can also define
    the mix
 + would be great to have sexy bling projects

How do we select student/projects?
==================================
 + we want to make our selection process well-known so that all students
   know about it
 + general agreement that we should avoid picking someone who already
   participated in a GNOME-related SoC/WSOP project or who is already
   a GNOME contributor
  - we want to favour new contributors
  - this is not a general rule, since a really good project proposed by
    a past participant/contributor will be considered too
 + we're fine with students having participated in SoC, with a totally
   unrelated to GNOME project
 + we don't require prior open source involvement
 + we'll let all potential mentors comment on all applications
 + in the end, a selection committee (~10 people) will read the
   comments, and decide which applications we select
 + all foundation members can be mentors
  - people who could be foundation members but don't want to can also
    be mentors

Other discussion
================
 + we'll create a mailing list for mentors
 + we need to advertise better SoC and our participation
  - put some SoC flyers in universities
  - we'll get a GNOME-specific poster
  - talk about SoC on our webpage

-- 
Les gens heureux ne sont pas pressés.