I assumed that it would be in poor form to have every single object in need of collision detection on a frame-by-frame basis to have collision tests performed against every object in the scene aside from itself I figured that now would be the time to add some form of spatial partitioning to my XNA engine/library/framework (Rawr).

With that in mind I looked around for some articles on writing an octree component and stumbled upon a really rad demo by Fabio Policarpo. I then adapted that to my codebase, edited the code as I, in my infinite ignorance, saw fit to do, and then added some rudimentary rendering functionality to the octree to provide what I assume will be a much-needed graphical visualization (in the future).

Anyway, it looks like a squishy death array of lasers.


While I'm at it, I present Cubegasm progress in the form of a failed rim lighting experiment:


As it turns out, six-sided, twelve-polygon entities don't look so hot when lit. I also present bullets:

And incredibly uninteresting class diagrams (only parental relationships; no containment ones):


For the record, working on Asplode! was so much easier. Anyone who says that 2D is harder than 3D is a dirty, blasphemous liar. No one says that, though. Because they are smarter than me. Not that I said that either. I think.