It struck me. Obviously we had the postmortem in the magazine. I was wondering -- why don't other people do it this way? I mean, I'm sure other people can't afford to take the time that you're afforded by being with Valve, but that kind of collaborative, iterative process that intertwines every element seems like a good way to go, but a lot of people seem to do every part of the game separately and then merging.

KS: To be perfectly honest, I think it would save people time. Like I said, you know if the player's unhappy, so if you think you've got a really great idea for some piece of exposition or gameplay, the faster you get it in the game and start exposing it to people, the faster you realize whether it was a good idea or a bad idea, and whether you should abandon it as soon as possible or keep improving.

EW: I'm just speculating, but there's a psychological component to it. Valve's, for lack of a better term, corporate culture is "Playtest, playtest, playtest." Portal was being playtested in the first week. But when you're working on some sort of creative project, your natural inclination is to not want to show anyone until it's in a state that you think you can be proud of.

But at Valve, we're putting you out there, and you're going to fall, fail, fail. You'll have little successes and little failures, and until you get used to the process, it's a little bit scary and painful. But it's really worth it.

KS: Yeah, we definitely start by showing players very unpolished stuff. For instance, in the levels, once we all sit down together and decide what a level's going to look like, we try to build it as fast as possible, and in two to five days we'll have it up and running.

We'll get someone to go through it for the first time, and you know, they're not pretty looking. They'll definitely need a lot of polish and care before we're ready to ship, but the sooner you start getting player feedback, the better off you're going to be.

How difficult was it adjusting to that process?

KS: Not hard at all. It was actually something... when we were working on Narbacular Drop, we wish we had known that that was actually a good idea. When we were working on Narbacular Drop, we playtested the game a few times, maybe a month before it was due.

We came up with a lot of things we didn't fix, and we just didn't have time to fix them. The sooner you start getting people to really know your games and start playing around with it, the sooner you'll find all those problems that because you're exposed to the game every day, you're not going to see them.

Do you think it was easier to adjust to it because you came straight out of school? I was just wondering that, because it seems like someone who was entrenched in, "All right, I'm going to wait until it's nice to show it to people," might have a hard time.

KS: I don't know if it necessarily made it easier for us to adjust to it. From our standpoint, we want to make the best player experience we possibly can. This seemed to work really, really well for us to figure out what we've done right, and what we've done wrong. It keeps you really objective, too, which is great.

When the writing and design was happening, did you intend to do something new?

EW: No, it wasn't like, "want to do something new." We knew the mechanic was going to be pushed. Even though the portal concept was sort of out there, even in -- I hate to say it -- Prey or whatever, there were games that had it, but we knew that we were going to push it as far as we'd think we could push it, and really focus on that particular mechanic.

In terms of the writing, the only thing was that -- and I wasn't sure if we were going to pull it off -- I kind of wanted the player to not have this nemesis that you eventually meet and then destroy.

I wanted you to actually see you progressively breaking the evil villain character down, and actually give you what I thought would be the satisfying experience of seeing your actions really cause the villain some emotional pain. And we wanted it to be funny, too, so that was a goal.

KS: There's something inherently kind of humorous in playing around with portals and bouncing through these objects. The fact that Erik is really great at writing humor... I think it just played really well together.

I don't know if it was intentional, but obviously GLaDOS is also a sympathetic character. It's like, you're breaking her heart, but she was fun, you know?

EW: Yeah.

KS: She still loves you! (laughs)