I am wondering what would be a good GSoC task (or tasks) this year? I don't see 
a good fit on my list of tasks:

"Simple tasks" a 3.1 bugs that won't take 3 months to fix and anyways we want 
them fixed (and will fix) before the summer. 

"Hard tasks" are hard because they involve serious design decisions that can 
only be made by a person who both (a) gets the Cayenne way of doing persistence 
and (b) can communicate with the community. An example would be a task to align 
EJBQL/SelectQuery into a single superquery and implement support for generics 
(this is a pretty vast task that I didn't even want to schedule for 3.1). 
Another example is support for inheritance. Past history shows that students 
who are not already Cayenne hackers will have trouble doing it right.

"Maintenance tasks". We need to switch all unit tests to be based on injection. 
Skill-wise this is an ok GSoC project. I am seeing two problems though - it may 
be deemed too boring to potential candidates (??) and it will generate a flood 
of patches, so an in-depth review may become problematic.

"Modeler tasks". Modeler tasks that require conceptual changes (such as an 
Eclipse plugin or graph representation of the Model) have been our GSoC 
favorites, still none of that made it to production. So if say I suggest a task 
of switching Modeler to OSGi plugin architecture, is there a chance this will 
get done to a point that we can release it?

Any comments or better ideas?

Andrus

Reply via email to