I've decided to rewrite Deadvale's codebase and reconstruct much of the project, which is a huge undertaking, but should be worth the effort.

There are many reasons for this, the foremost being the original project was based on code I'd written for an earlier, different project, which was built to be multiplayer. Deadvale is meant to be a single player game, and while the current demo does support multiplayer (which does work over Steam), I'm not that well versed in networking and would rather focus on singleplayer rather than trying to balance both. While designing a multiplayer game, every feature has to be tested on a simulated network with two players, and game design itself is ideally balanced around 1-2 players, which is a lot of extra work.

Another reason for the rewrite is the original codebase is poorly architected and doesn't extend well to level editing. As this is a hobby project, I do want to work with artists, and the best way to invite someone to work on this is to totally segment the programming from the level design process. It makes collaboration much easier if others don't need to worry about what's under the hood - and can seamlessly create artwork, level experiences, and design their own levels without needing to edit any code.

The final reason is optimization. The original codebase was very poorly optimized, and I quickly found that scaling to large numbers of enemies, or giving NPCs highly detailed interaction code to make them more lifelike would not work well. The best option given this was a full rewrite. In this new codebase I've focused heavily on composition and inheritance rather than god classes as the original demo uses. This has been a little difficult, but worth the effort for the gains in framerate.

Most recently I've focused on ensuring a test scene with 1000 "pickup" physics objects and 100 hostile NPCs maintains an acceptable framerate with zero garbage collection.

I've also starting working with a fantastic composer, Julius Rimkevičius: