You were developing the engine simultaneously with the game. That's often looked at as a quite difficult approach to making something, but perhaps it's not as difficult when what you're making is kind of also tools. So, when you're making tools to support a game that's based on tools...

AE: It was a really interesting chicken and egg situation. I was talking to Danny [Leaver], he was one of our level designers, and he was like, "People ask, 'Did you use the in-game tools to create the game?', and how do I answer that?" I say, "Yes, we did." Because what happened was the tools evolved, like I was saying in my [DICE] talk, phenomenally.

They changed without recognition several times. And so, to him, the tool that you get in LBP existed for around a month before we shipped, or maybe two months. So he didn't use that tool to build the game; he used like all the previous fuck-ups.

So, I was like, "Well,you didn't use Pop-It in its current form, but you did use it in a previous form, in fact." Because what we were doing is, we were developing out not just the engine, but we were like, "How do we create?"

The example I've given before is, previously, there was complete free range on depth. We now have three layers with thin layers between them. That was actually added later.

Weactually had to go back and revisit every level and go through and rethink it in terms of those three layers. So, changes like that were really interesting in evolving the toolset and the level design at the same time. Yeah, we were feeling it out.

Pretty much everybody is licensing engines now. How do you feel about that sort of scenario?

AE: To some degree, we're all building on something. You know, I have the PS3 SDK underneath me, and all this stuff. It's basically what level of abstraction you're willing to kind of give up knowledge of how it works.

I think you can license an engine and be really successful and do a fantastic job of it provided that you're willing to stop at their documentation and not have to worry about how the engine works. And to some people, that's a huge bonus, like, "I don't want to have to worry about how their shader component works." But for me, I do.

I don't know BioShock's process, for example, but I know that that team probably knows the Unreal Engine well enough to have written it. It probably was fine that they licensed it, but they had to go through a process of learning that engine all the way down.

Tearing it down.

AE: Tearing it down, yeah. At a certain point, you're like, "Well, you know what? We could have just built something ourselves." There's a fear of that; there shouldn't be.

I think that one of the major benefits that people talk about is the ability to instantly prototype gameplay and stuff. So, if you're going for that like single-player experience, then perhaps it's even more valuable.

AE: Actually, I learned that from [a Maxis talk on prototyping]. They showed all these prototypes they'd done on SimCity years ago.They have this actual structured process for prototyping -- it's four years ago, so I'm a bit hazy on their details -- they actually have this formal thing where if you want to put a feature in one of their games, you had to do a prototype.

You had one week to do it. You had certain constraints. Then you have to prove it out. And if you could get a one-week prototype to really get people really behind the idea, then you got a two-week window to kind of flesh it [out]. Their approach to prototyping was really inspirational to me back then.

I think that might have been related to Spore - I heard they were prototyping hundreds of concepts and ideas, and I can't imagine that even 25 percent of them made it in.

AE: Yeah. I'm going to use [Fallout 3 executive producer] Todd [Howard]'s line of, "We can do anything, but we can't do everything." My other favorite quote in that direction is "ideas are like assholes; everyone has one" sort of thing. The problem is picking which one you go for. We struggled with that, to be honest, internally.

You're right, about one in five ideas probably makes it into the game. And of those one in five, they've probably gone through three iterations, like you've decided on that one, then you've binned it twice. You've actually ended up making eight things, and one of them succeeds. It's a difficult sell.

Sometimes it just comes out the first time. That's epic. That's cool. That happens, too, and I like when that happens.