On Wed, 7 Mar 2012, Jason Taylor wrote:

Ok, this woke up the sleeper cell. Allow me to burst in even though I'm
dormant otherwise. I am however forever attached to this project in heart and
spirit!

> Don't be so hard on yourself. Rockbox is cool. But otoh the fact is you
> guys do not allow just anyone to submit project ideas, so 99% were not even
> allowed to contribute ideas. Perhaps you should allow anyone to be a mentor
> not just an experienced coder.

We already allow "anyone" to become committer and therefore mentor.

We only allow committers as mentors because we take Google's sponsorship and
offer very seriously and if we would guide a student through many weeks of
full-time work, we MUST know that the person who'd be the primary project
contact is a sensible and responsible person with a good sense of the project.
That mark of trust is what the committer badge gives us.

We don't just accept coders as committers. We have several core team members
who never wrote any code at all and yet they are committers since many years
and in no way do we make any distinction between who code and who don't.

Accepting others' projects ideas for the list of gsoc projects could be useful
but do bring the problem with how it should be done/maintained. Perhaps we
could just have a secondary page with "contenders" for the real list open for
anyone to edit.

> Rockbox community is huge, and you have enough support and a big enough user
> base that to not have the energy to even take advantage of this project is
> indeed something to be a little embarrassed about IMO. Google offers to
> improve and you say no thanks? Wow.

You are of course entitled to your own opinion. I have a different one. We've
done the gsoc project for many years now so we have a decent understanding of
what it takes and what we need to make good use of it. I say out of respect
for the offer and out of respect for the potential students, we must only
attempt to do this if we REALLY think we have good projects ideas and enough
mentors volunteering. This year we failed this (so far).

> The projects I saw were indeed a bit non-brilliat, with near zero noticeable
> improvements

I don't see now "noticability" is any factor at all to consider as to wether a
project idea is good or bad for gsoc.

> 1. There is no reason to not have plugins for all the popular machines
> before 1997.

If there really is no reason, then how come we don't already have them all?

-- 
  / daniel.haxx.se