Yesterday several Bay Area Derby folks had lunch together: Bryan
Pendleton, David Vancouvering, Kathey Marsden, Myrna van Lunteren, and
myself. I wanted to bring the Derby-related conversation back to the
community:
1) We talked about our growing bug backlog. It seemed to us that the
bulk of the bugs were recorded by Derby developers--issues which we have
stumbled on ourselves. On the bright side, it seemed to us that the
community was doing a decent job of fixing production problems reported
from real-life usage. We did note that some very hard bugs have
languished several years.
2) We talked about the next feature release (10.5), carefully skirting
the issue of who would manage it. The first quarter of next year looked
like an attractive time to produce a new feature release.
3) We thought it would be useful to compile a multi-year feature roadmap.
4) We talked about how we could improve next year's Google Summer of
Code. Last year's process of selecting students ended up excluding some
strong candidates. We thought that we might be able to improve next
year's process if we started agitating early. Here were some thoughts
for generating more interesting projects next year:
a) Compile bug collections, each of which focuses on a narrow part of
the system, allowing candidates to develop a sense of real mastery over
a part of the code. For extra credit, each collection could be
organized from easy to hard bugs in order to make the learning curve
gentler.
b) Write up specs for interesting features and then provide a workplan
showing a candidate how to chunk the features into bite-sized pieces.
Regards,
-Rick
