11-16-04 - 3


We were originally scheduled to ship about a year before we actually are. There's certainly no good reason why we didn't make that date, but various things came up, like having to make demos and sign a new publisher, and the fact that we had no game design or designers for the whole first year of dev. Oh well, minor issues. Anyway, shipping at the wrong date really screws up development. If I had known our actual ship date, we would have done things very differently - better renderer (with dynamic lighting), better physics (maybe a real licensed physics engine), better tools, better animation system & pipeline, etc. - all those things were cut to make the original ship date. For the whole last half of development (18 months) we were in the mode of "we have to ship in two months", and that deadline was continually moved back. Well, if you only have two months to go, you have to be in feature freeze, so we had lots of problems with pipeline, etc. that we couldn't fix. The result was that everyone was working less efficiently than they needed to. When you know your real schedule you can do proper planning - heavy tools work in the first 25%, finish the engine by the next 25%, finish the features in the next 25%, and polish in the last 25%.

No comments:

old rants