After nearly oversleeping before Day 2, I went and watched
Soren Johnson talk about his experiences developing the AI in Civilization 4. He discussed the differences between AI designed to replace a human (which he called "good" AI) and AI designed to entertain a human (which he called "fun" AI). The differences between these two are pretty obvious, though he enumerated more subtleties than I would have. He said that a good AI doesn't cheat and a fun AI can't cheat, so what does that say about Civilization 4's, which is somewhere between good and fun (the snarky-comment-inclined in the audience may be tempted to say that it is neither good nor fun)? It was a fairly interesting exploration. However, I was more interested in the technical aspect of the AI. It instantly and automagically works with any mod that adds or removes tokens to and from the game, due to some good structure that I can only imagine must have been a pain in some intern's ass to bring to code. The AI itself is open-source, and compiles to a DLL which can just be dropped into the program, rather than being a part of the executable. Yeah, none of this is new for people who've ever played Civ 4 before, but this is a group of people that doesn't include me. Hearing him explain it made it into a very potent lesson in how to make a game modder-friendly. It also related greatly to this conference's annual buzzword,
Democratization. This is not just a theme any more; it's a cliché. But as clichés go there are many worse.
After that I saw the keynote by Ray Kurzweil. Right off the bat this had me in a good mood. I
like Ray Kurzweil. And the music was nowhere near as obnoxiously loud as it was during the previous day's keynote. I just wish the guy would write a different speech every now and then, rather than just giving a tangential introduction and then using the same one he always does and cutting material out to fit it into the time that he's given (in this case, an hour). Still, though, hearing from that guy is always a blast, and I got to show off how much I know about it, and quickly made notes about how more specifically to apply it to the various game-making professions.
After that I saw the (interestingly named) Rod Humble, and his colleague whose name I've forgotten, from Maxis, show off the latest spinoff of The Sims brand,
The Sims Carnival, which is even more of a YouTube for games than XNA is. It's a flash game portal with an interface for making (apparently pretty broad) parametric variations on a handful of arcadey game standards, with user-supplied graphics or graphics taken from one of the
Sims games (which, he says, represents about fifty million dollars in graphics in total, so that's neat). Also he showed off the more advanced interface, which is a separate program that implements a very simplified graphical programming language. I didn't get a good look at that; however, he says that after stripping away the presets it's a rather fully-featured programming language, which compiles to Flash games. Additionally your own Flash games can be uploaded there. It's looking like it will have a bright green but otherwise very YouTubey interface, complete with categories, tags, popularity rankings, five-star ratings, comments, and the rest. Democratization, my nemesis, we meet again!
Next I attended a roundtable about preserving old games and peripheral materials (design docs, prototypes, concepts arts, correspondence, contracts, etc.) related to them, in a manner that might become appropriate for use by Serious Academic Research. I think maybe it was a mistake for me to go there, rather than one of the other sessions, but they can't all be winners. I stuck around and listened to people who are obviously quite serious about mixing games and academics, and it was nice to see Warren Spector and Steve Meretzky hanging out in there. (Speaking of Steve Meretzky, I think I forgot to mention that yesterday he and Marc Laidlaw, Ken Rolston, and Richard Rouse III analyzed some of their favorite game stories. It made me want to play
Loom)
I hung out in the expo for a little while, too, honing my schmoozing skills. Prospects: favorable! But it would be inappropriate to go into detail.
Finally I went to unwind at the Game Design Competition, where returning champion Alexey Pajitnov and his Remote-Control Dolphin Team Underwater Paintball were
trounced by Brenda Braithwaite's Dog-Having Facebook ARG and Steve Meretzky's Real-Time Bacteria-Exploding Strategy. Meretzky won by a hair. Frankly I think Braithwaite's design was the most likely to
succeed, and she seemed to think so too because she registered a URL for it, but when the extremely subjective judging came along, I had to vote for Meretzky's. Meretzky's combined lasers and
real-life cellular automata, making it the sixth coolest thing ever, whereas Braithwaite's consisted entirely of three of my five least favorite things: ARGs, Facebook, and dog ownership.
Everybody I know in town is out drinking right now, but I'm an underage teetotaler so instead I get to talk about it on the Internet.