So once you figured out what these puzzles and mechanics were supposed to be, what visual cues were necessary to guide the player?

KS: Going back to the barebones design of how we actually picked which puzzles to use is, we would all -- once again, designing by committee -- we would say, "Okay. Our game is separated into three different wings with various difficulties, and the first wing that you go to, you only get normal and two dimensions, so three to play with" -- because we didn't want to inundate the player with too much information at once.

So we say, "Okay, we need to teach the player how to use this dimension, and here's all the things that this dimension can do. So let's pick one of those particular skills that we're trying to teach them, and figure out a level just to teach them that one skill. And then let's figure out how to reinforce it, so we'll come up with a couple puzzles that are going to reuse that particular skill, and then add a new one and keep going."

Once we figure out what puzzle we're going to build, we want to then kind of whitebox it and make sure that it plays right. Then we rearrange the room compositionally so that way... I call it "design compositions."

Instead of it being like a painting where everything is still, you want to use things like motion, lighting cues -- not only lighting but the shadows too. As you can see [playing the demo] our window shadows are actually pointing, literally, at stuff. So figuring out a good lighting composition for the level, as well as detail.

So if you compose a scene where there's a bunch of detail in one area, players tend to just look at that. So just using a lot of psychological cues that you would use composing a painting, and doing a scene like that for a video game.

It seems like you've always done that. It seems similar to the way you must have put Portal together.

KS: Yep! I lit and designed Portal's levels too! [laughs] So yes, that's how I learned how to do things. It works the best. You know you've done a good job composing a scene if a player walks in a room and immediately looks at what you want them to look at first. It's like... "Yes! I won!" And if they walk around and they're just kind of confused and not sure where to go, then that means we did something wrong, so we've got to figure out how to get them to look at what they want to look at first, second, and third, and then compose the scene that way.

How much playtesting does that take?

KS: We have weekly playtests, where we grab somebody and run through from start to finish of what we have, and a couple of us will sit and watch and take notes and then track bugs based on what we've observed. Then Square Enix will also be running testing for us too, as far as bugs go.

You have the shoulder buttons up on the HUD, and you've also got this company logo. Did you realize you needed those?

KS: We're still playing around with the HUD itself, the HUD we have up here right now is not final. But yeah, we definitely like having the ability to have you see on screen which dimensions you have where. Because certain times you will only have a couple dimensions, and sometimes you will have all four. So it's just kind of handy to have them up on the screen.

It doesn't seem to get in the way very much. It feels like you've got a bit of a fisheye camera, where you can really see around it.

KS: Yeah. We definitely test a whole bunch. We're still in the midst of iterating on our HUD. But I think the reticle is going to stay the same.

You've got five dimensions total?

KS: Yep. Five total. When we were first developing this particular project, we were just kind of playing around with logos for Professor Quadwrangle's company. So he has a couple different logos, one for himself, one for his company, Quadwrangle Industries, and that particular one is a take on a biohazard symbol that we thought was kind of clever. We're trying to be clever.