-----Original Message----- From: Alexandre Prokoudine [mailto:alexandre.prokoudine@...400...]
Not exactly high. The previous year was a 2/5 failure
indeed, but the
year before was 100% success, if my memory serves me well,
and we are
talking about 6 or 7 projects in 2009.
In 2010 we actually had 3/5 fail. :( It was very unfortunate and I hope we can avoid that this year as the "acceptable" (desired) rate is 80% successful.
One of disappeared students last year had been around in
the community
for a rather long time, and the other disappeared student did a successful project for GIMP in 2009, so it's quite possible that we just had our bit of bad luck. We did have disappearing students in even earlier GSoCs though. IIRC, this is where the "two patches" rule comes from.
Actually the two patches rule comes from it's the minimum requirement to get commit access to the repo. It was part of the change in attitude from sodi podi to be more welcoming. It's honestly a pretty low bar. It was decided that it would be good to require this before the students would even be considered as it was "proof" of them checking out the code and showing a form of commitment/desire to participate in GSoC (plus what Johan said about it allowing us to see to some extent if they really can read/write code).
I really think that SVG compliance projects should be a focus this year as well (not exclusive by any means though). The list Tav posted does have some really great ideas, so I think we need to really analyze what those tasks really look like and which would be reasonable GSoC projects.
Cheers, Josh