2012 Week 12: inspriation and a redesign occurs

The earliest proof we have on current laptops

We know dates can be manipulated, but this filetime record (see image) is true and correct. The wayback machine on this page will corroborate your findings (if you are interested in investigating such things).

But this image shows when we found our style. The signature white with teal. Since then we haven't looked back, and have only improved things. Oh, and we added a black with teal as well, just for the more moody sections of the game.

This was also the moment we decided to abandon an earlier idea around platforms and move towards thinking in corridors instead. From that point on our engines have all been corridor focused. Little did we know at the time that this was going to make things much harder. People don't design inverse physics engines! But that's what inspiration does half the time, either that or passion... not quite sure which.

Some might even call it mania.

Inverse Physics Engines?

We did try to build one, but quickly found our propensity for geometry not powerful enough for the kinds of equations needed. We needed a mathematical genius, and that wasn't us. So, instead, we set about building on top of known frameworks, Box2d was the one at the time. But, we found optimising things pretty difficult. We would get some parts working, but glitches were-a-plenty.

No, the question was "what is an Inverse Physics Engine?"

Oh, right. Well you know how physics engines model what is hard and leave all the rest of the space being open. Well and inverse physics engine would do the opposite. Everything is hard by default, and the shapes you add are what cut spaces out of the world.

I see. That's a bit mad, isn't it?

Yes, and no. It would definitely have helped with keeping our corridor idea performant, but yes, it would likely have needed to run both. A normal physics engine and an inversed one, because we'd still need solid objects in the game.

POC Multiplicity

Most of our work back at this point was based around Proof of Concepts. Small little projects that showed one element working. It was all that we could fit in amongst working full time jobs. But doing things this way, keeping things modular, meant we could keep the project going for far longer. And it also meant fewer complete rewrites.

True, we rewrote the engine a number of times. But the features came with us. This has been our model for development ever since.