Tuesday, August 05, 2008

training the autopilot

I'm seeing a pattern of things being on autopilot. DNA is like the instructions to automated systems, people respond to things automatically and only get to decide what they do if they change themselves, even organisations react in automatic ways to situations. I'm wondering if this setup and run way of working is more applicable to games development than I'd previously considered. We write graphics rendering as a setup and run, why isn't the rest of the game written like that? Doesn't the brain, an organisation, DNA all benefit from setting up, then multi-dimensionally attacking the problem space? The way forward for games and computing in general then may be large setup phases followed by fast vector processing of the problem domain across more levels.