How often should you run your tests?
Often enough that the average number of bugs you add between tests is less than 1.0.
Otherwise, when you run the tests, you've added *multiple* bugs, which makes debugging vastly more difficult.
@peterdrake
How would you measure the frequency of tests run?
By time passed, number of modified lines, number of modified files?
@rastinza It feels like modified lines is closest to "how much have you changed the software", which is the best estimate for "how many bugs did you introduce". That said, it's probably more dangerous if the ten lines you modified are in ten different files.
My main point is to urge people to run their tests more often.
@rastinza I think some have it set up to run the tests every time you compile or commit.
@peterdrake
This could be implemented in an ide, directly advising you to run tests; however you would have to train it by telling him how many bugs you have introduced each time you run your tests