When I learned of it, I always thought of mutation testing as a great way to test your test suite. Simply stated, if you introduce a common bug into the code, do the tests find it?

The time to run mutation tests is rather large, though. Almost prohibitively so for many shops, I would think.

I'm interested if anyone has run studies of codebases that employ this technique at large.

For Java there is Pitest: http://pitest.org/

I was first introduced to this idea in around 1977, and it wasn't a new idea even then. It went by the cute name bebugging [1].

[1] http://en.m.wikipedia.org/wiki/Bebugging

Hi,

Here at IMVU we found some ways of how to improve the Mutation Testing speed, but I agree that you need some degree of parallelization to achieve its purpose. Anyway, I found Mutation Testing very helpful while creating robust code :)

Llorens