Check++
Well that’s milestone 2 about wrapped up, on schedule and no dropped features. I’d be hopelessly naïve if I thought that two successful sprints in a row was down to fantastic management or process, or even down to Pete and I and our amazing skills; rather more likely is that these first two milestones have consisted of things that we knew we could deliver without fuss. Next sprint is a bit more meaty – two sub-games, with all of the design ambiguity and art bottlenecks associated with content generation. We’ve dealt with the engine functionality needed to deliver the content, now we have to deliver the content itself.
Looking over all of the legacy code though, I am having to resist the temptation to write it all again from scratch. Yes, the code is tested in fire, and exceptionally functional, but in many places it is ugly, unmaintainable, and downright hard to use. Combine that with the fact that it came from a team who were constrained to work in straight C in most places, and the code isn’t exactly clean and easy to read. So with that in mind I’m inclined to use it as a reference point rather than a base to work from, and pare it down to the essentials. But then the businessman in me kicks in, and reminds me that we’re not here to make good code, but rather good games. So the decision is really – what will help us make better games, faster and more reliably. Decisions, decisions…