Troubleshooting Yourself in the Foot

The article says that successful debuggers made better use of a mental model. I've seen people do lazy debugging many times: they change a line here, build & test, change a line there, build & test, and it doesn't appear as if they're stopping to think at all. I've seen people dive right into this without reading the code at all. I call it "debugging by superstition." Perhaps they were already in the zone and had the code perfectly laid out in their head, but then you'd think their jerky attempts would have worked. :-)

At the other end is what Raymond Chen calls psychic debugging. This is where you have enough experience to be able to guess the root problem from seemingly-unrelated symptoms. This has a superficial resemblance to debugging by superstition, but it's quite different.

So... the takeaway is to stop adding bugs to programs you write? Heh, I've been trying to do that since day 1.

Seriously, now, it appears that experts are more akin to taking the scientific method approach to debugging: isolate and bisect down to the point of the bug. Novices just haphazardly change stuff until stuff appears to work.

I think a lot of us can think of friends and co-workers who fit these profiles.