From 1.1 defect density to 0.08: How did you do it?

I’m writing about open source testing issues and just read the Coverity press release. How the heck did you people manage to do it. How many people, how did you divide up the files, etc?
thanks,
Jeff Thomas
jeff.thomas@se-eng.com
303-604-1020

I have seen a lot of Coverity-related changes for several months now, but I too would like a clearer understand of the mechanics of how this was achieved. Will likely require a developer to comment. It may be better to contact the developers directly.

Answered by email on the 16th. The presentation at http://www.skynet.ie/~caolan/libcon2014/LiboCon2014-coverity.odp or http://www.skynet.ie/~caolan/libcon2014/LiboCon2014-coverity.pdf basically covers it.

Thank you and well done. Very interesting reading.

Further blog entry indicating the Coverity defect density is now 0.02.

Subsequent blog entry indicating the Coverity defect density is now ~0.00.