Google Summer of Code 2008 improvements
Some thoughts about the programme for next year
Maintaining an Eclipse projects for SOC is a rather poor fit with the Eclipse development process.
For 2008, we should consider (note that these are just ideas for discussion):
- SOC Students must be treated the same way as other committers.
- Specifically, they should be elected into the position and are subject to the same requirements
- If a student's work is related to an existing project, they would follow the same process as anybody else
- Student contribute code through bugzilla like anybody else. That code can be put into CVS by a committer (possibly the mentor) assuming that it meets with standard requirements.
- Projects can set up an incubator if they deem it necessary.
- If, over time, the student demonstrates that they are a worthy committer, an election is held.
- If a student's work is not related to an existing project, then a separate home needs to be found
- For example, a Google Code project (which sort of makes sense)
- Over the term, the student can consider going through the proposal process to make their project into a real Eclipse project.
The short version is that GSoC students should not be treated any different than other committers.
This is good because:
- Students experience the real Eclipse process.
- It doesn't screw up our process, or reinforce dangerous precedence for bending the rules.
- Lowers dependence on Eclipse Foundation infrastructure to get projects going.
This is bad because:
- Will really only work if mentors are already Eclipse committers.
- Mentors become a bottleneck for getting code committed (though this is probably not necessarily a problem)
It may seem a little elitist, but it's probably reasonable for us to require that mentors be Eclipse committers.