« Gunslingers and Samurai | Main | Golfin' with the Boys »

August 25, 2005

Discussion: Two Plus One

Jules and Jim Façade

Jules et Jim is a great little movie about the complications of friendship and romance, of the duties we owe one another and to our own happiness. The titular pair are great friends who meet in Paris, Jules an Austrian and Jim a Frenchman. They grow to be great friends, thoroughly understanding one another.

Soon, a woman enters the picture, Catherine, and her amazing resemblance to a statue they both admired in Greece strikes them both, and makes them realize that she is somehow different than their other girlfriends, which they have sometimes shared. Catherine is a free spirit, and as portrayed by Jeanne Moreau, she crackles with a frantic, radiant energy that must be seen to be understood. Both men fall for her, Jim perhaps the hardest, but Catherine chooses Jules, and Jim respects her choice and does not try to interfere.

It is at this point that the film gets a little strange, or perhaps, a little stranger; war intervenes, and Jules returns to Austria with Catherine now his wife. Time and the war pass quickly; each man worries for his friend and hopes that he will not meet him on the battlefield¹. They do not, both survive, and time marches on until one day Jim pays them a visit.

Jim encounters the couple, very unhappily married, with Jules still caring only for Catherine's happiness so long as she can be made to stay near to him, to still share in his life if she will not share his bed. She has had lovers, and soon takes Jim as a lover, with Jules' blessing, as they will live in the house, and no harm will come to Sabine, Jules' and Catherine's daughter, thereby.

There are other wrinkles, but by this point in the story I've told you enough to impart the film's strange flavor. At the time that I watched it, I was a little stymied, but like most great films, it sticks with you and grows a little bit in your mind; you recall its images and its subject, and it ends up making you think a bit about what two people who love each other owe to each other, and what concessions they should make for the other's happiness. In the case of Jules and Jim, there are three pairings -- the two men clearly love each other, and each of them loves Catherine. Where it gets interesting is when that third wheel is added to the mix².

It's funny, but I kind of feel the same way about Façade, the research project everyone's been talking about³. What's interesting about Façade, at least in theory, is that it does exactly what Jules and Jim does, but it puts the player in charge of exploring his relationships with these other characters.

It's not properly a game, unless you'd call it a role-playing game, with a heavy emphasis on the role. Players4 arrive at the home of Trip and Grace, a married couple who are clearly having a domestic dispute which is interrupted by the ringing of the bell (at a point of your choosing).

The simulation takes input through typing, as anything you type is something you say. This is, unfortunately, a rather clumsy interface, and even with my very high word-per-minute rate, I continually find myself just a beat behind in conversation, often cutting off one of the participants mid-sentence as I furiously pound out my words. That aspect is quite frustrating.

What really thrills me about it, though, is that I can approach it with my own role in mind. Am I to be the cad, who has always had the hots for Grace and now can make my move? Am I supportive of one character or the other? Am I uncomfortable? Interface issues aside, it aims to let me make these choices, and despite those flaws, it's still really interesting. Jules and Jim explored the interactions of two people and one other, and so does Façade.

I wondered over in Jamie's blog whether the implementors had done any filmed tests to see what worked for the experience and what didn't. Often in academics, you look to see what explanatory power your model has, and sometimes you take what you have and measure it against what people actually do5. I think it'd be really interesting to take some blind subjects and run this scenario with real people -- actually walk these blind subjects up to the door with exactly the information they get when playing the game, and let it unfold with a couple of actors.

I'd love to know how they might have changed their simulation in response. Would people ask "Well, wait a minute -- did I meet Trip first, or Grace first? How long have I known them?" How much more information would people want before they felt comfortable? In what ways would they connect with the characters that weren't reflected in their simulation -- longer hugs? Back rubbing? Peering attentively? Making faces? There's so much richness there, and I'm curious about how much of it you'd have to add in to feel like you had enough interface to emote properly?

Would speaking directly to it be enough, through a microphone? I don't know. But it'd be a start. I don't want to push more hard problems on them, but given the simulation, that kind of real-time interface is pretty desirable. Had they simply presented it as a text "adventure", the typing interface might have worked much better, though the results wouldn't have had the immediacy. Maybe playing it out as a text adventure and then playing it back as film might work.

So, I guess I think Façade is pretty important too. Like Jules and Jim, the more I think about it the more questions it makes me ask, the more it makes me think. That's a significant contribution.

¹Talk about your pronomial binding problems. Anyway, Jim worries for Jules, Jules for Jim, and neither wishes to meet the other in battle.
²Note: Wildly diverging metaphors!
³See site for links and quotes. I learned of it through Ernest Adams' write-up on GamaSutra.
4Interactors? Consumers? Experiencers? Participants? Since it's not really a game, it's not really proper to call us players. An experience which drives me to seek new terminology is often a good thing.
5Years ago when I worked in graphics research, I was co-author of a paper about generating speech and gesture for animated conversations, which still shows up in searches on my name. Anyway, one of the things I found interesting about the project was some of the errors we would get, and the ways it would fall into the uncanny valley (behaviorally speaking; visually speaking, the poly models were far below what we have today). The chief researcher, Justine Cassell, had done her graduate work in gesture, and behind her theories were some good indications of why we actually can mis-gesture. Fun stuff. While academia is really not for me, some of the intellectual questions it poses are still really interesting to me.

Posted by Brett Douville at August 25, 2005 09:33 PM