game-development

poor software architecture in games?

there seems to be this theme going around lately of people saying that game developers have no clue about proper software architecture. here's an example going around..

http://www.jeffplummer.com/Writings/Writings.htm

i've seen a lot of gamedever codebases and most of them do suck, but suck compared to what? suck compared to any other apps out there? not really. it is a myth cause by gamedevers themselves saying their own stuff sucks and that they think they can do better.

that thesis above points to the example of doom3 being rewritten from scratch. geez.. there must be something wrong with the architecture then. no. carmack rewrites his stuff from scratch each time.. because he can. it might take up front time for the rewrite, but in the end he will end up with a better codebase. why? because he understands the problem domain better.

the doom3 codebase is significantly better designed than the previous engines. this might have a lot to do with certain other team members pushing the architecture, but that too is based upon experience. the result is a better engine.

you can masterbate all day with uml, but in the end all that matters is experience in the domain. experience of course is something that usually grad students don't have. YET!

i'm sure if he rewrote his thesis in 6 months he'd have completely differing opinions on the matter.. ie. a rewrite or rethinking is in order.