I recently played an adventure game in which you play character A and have to lock character B in a room, and then in the next chapter you play character B and have to escape. It was a slightly unsettling shift in perspective: I'd been playing character A and identifying with her, and agreed with her that character B needed to be locked up because he was a danger to himself and others. I didn't want to help him escape.

alextfish plays a lot of Starcraft, and he says you get the same phenomenon in that, but even worse. At least the two characters above were ultimately on the same side, but in Starcraft you play the Terrans and build a base, and then you play the aliens and destroy the base you just built.



A good developer needs to be at least a reasonably good tester; and a good tester needs a certain quality which, in normal life, is usually bad. I might even call it malevolence, or at least scepticism. It goes beyond just the destructive desire to try to break things; you have to try to break things which other people have just created, which they've put time and effort into, which they might have invested a part of themselves in. You have to assume those creations are flawed, and make it your mission to expose the flaws.

I am not very good at writing robust code. The testers find even fairly obvious bugs in my software. And I think this is because, on some subconscious level, I'm being precious about the thing I've just built. I don't want to prod it until it falls apart; I don't want to look for the flaws in it.

To be a better developer I need to apply this perspective-shifting, side-switching trick to my work. I need to look for ways out of the room I just locked; I need to bomb the base I just built.

Tags: games, software, work