#46 defect statistics

[Richard Cyganiak]

Felix Hubfeld (hubfeld at zib de) on statcvs-devel:

Inspired from an article in the latest CACM, it think it
would be very cool if statcvs would have defect
statistics.

Mainly, number of bug(-fixes) per LOC and timespan and
LOC per bugfix (bug complexity) shouldn't be too hard to
add. From the user perspective, adding a keyword like
BUGFIX to the CVS comment would be sufficient to
trigger it.

[Nobody/Anonymous]

Logged In: NO

This is a great idea. We've used cvswrappers to require
something like 'BUGFIX' to tie CVS to specific problem reports
our problem tracking system, so this would be no impact to our
development team. I suspect that's the case on most
serious/highly managed development projects.

Another stat that could be tracked is fault density (bug[-fixes] per
module). I am sure there are others.

[Nobody/Anonymous]

Logged In: NO

This is a great idea. We've used cvswrappers to require
something like 'BUGFIX' to tie CVS to specific problem reports
our problem tracking system, so this would be no impact to our
development team. I suspect that's the case on most
serious/highly managed development projects.

Another stat that could be tracked is fault density (bug[-fixes] per
module). I am sure there are others.

David.Gathright@lasp.colorado.edu