One goal of the compilerMonitor is to help understand
how to make a developer more productive with the
compiler. Crappy compiler error messages can really
waste a lot of developer time, especially when averaged
out over months/years.
Currently, the compilerMonitor archives a record of the
developer's javac usage/behavior over time. I think the
tool could really use a facility that boils down all the raw
recorded data.
Just glancing at the summarized data using this facility
should help the user quickly understand more about
these questions:
1) Which compiler errors cause the biggest drain in
developer productivity?
2) Has a developer become more efficient in using the
compiler over time? This might be helpful in an
educational setting.
a) Perhaps a shrinking average amount of time between
a compiler error and a corresponding fix suggests
increasing productivity.
b) Perhaps few errors for a given time period suggests
increasing productivity.