Given that we have conducted our experiments on four
open source systems and two proprietary systems it is not
clear how generalizable the metrics are. We could reach out
to the maintainers of the proprietary system and verify the
correctness of the analysis. But we could not get similar
feedback from the maintainers of the open source systems.
We recently completed a study on another proprietary system
in the .NET platform (which we have not described here
due to lack of space). Even in this case we received good
feedback. Hence, we believe, the metrics would work well
for other systems as well subject to the availability of good
quality data. We are still in process of obtaining version
histories for few other proprietary business systems in the
banking domain and retail domain. This will help us to base
our results better.