Bug Generation

One of my pet peeves is when people try to objectify something subjective. One good example is employee ranking.

I know about a large software company that hires both “developers” and “testers”. The developers are graded primarily on how fast they can code. The testers are graded primarily on how many bugs they find. It sounds pretty innocuous, but there are some major problems.

As far as I can tell, if Joe Developer writes X lines of code per hour with Y defects per line, he’s not penalized for Y in any real sense. Maybe he’ll get a reputation for producing sloppy code, but that’s nothing compared to the wrath he’ll face if he’s late on a deliverable.

If Bill Tester finds 10 different ways to exhibit the same bug, he gets credit for them all. He has no incentive to understand the product such that he could identify these as duplicate bugs. In fact, he has no real incentive to open good bugs, because management just uses the raw number.

The end result is a system organized around bug generation.

Instead of fostering practices that create defect-free code (pair programming, test-driven-development, etc), the company is actually encouraging the creation of defects.

And that’s exactly what they get.


%d bloggers like this: