Ventures of an indie game developer

Premature optimization

I sketched a little, thought some more and started cloning my KillCutie project files to get started. Then I realized that many parts of the code was cut out for that project (from project Push). Especially network coding and some other features that I want. So what I essentially want to do is a Quake III Team Arena using hovercrafts. With worse graphics and sounds, interchangeable parts from a shop, a little different single player campaign and more physics.

The code for project Push was pretty okay, mainly it was pretty straightforward. But it was performance intensive, as each new split screen client instance also required a new physics simulation sandbox, socket instance, game object environment, local time synchronizations and adjustments and then some.

The refactoring required to break those core parts out, then put them back together in one singleton (nope, not the static Singleton pattern, just a single instance) where each client (or server) part isn't violating the vague and implicit contract of use for these core parts are going to be a big undertaking for me. But it has to be done, premature optimization or not.

About the author

Mitt foto
Gothenburg, Sweden