Hi,
We have been collecting Java class dependency information for Hackystat
project for around 2 week. Basically, for each class in Hackystat, we
know the name of classes it depends on. This is a lot of detailed low
level information, probably too much for anyone to sift through.
I am thinking about how we can use telemetry streams or other types
analysis to offer at least one perspective on complexity of our code.
I still remember that one year ago Christoph used some graphical
application to draw the dependency relationship of Hackystat classes on
the telemetry wall. But no one really used that information, probably
because his dependency graph still contained too much low level information.
So, before I start to write the code, I want to solicit your opinions:
(1) What kind of dependency information might be useful to you (as a
developer, as a maintainer, and as a project manager)?
(2) At what granularity level?
Thanks.
Cedric
