I've been loosely following the development of Façade for years now, and finally they seem to be finished and have it available for download from their website here (www.interactivestory.net). However, I can't really afford to get a 800 MB download at the moment (due to complex issues regarding my internet connection!). By the way, there's meant to be a curly tail under the c in 'Facade', but I'm having trouble getting it to display in HTML... (Edit: it seems like the ç's are rending now, anyway. Must have been a problem with the preview...) Anyway, I was wondering if anyone here had tried it, and what they thought of it, from a game design perspective?

---

For those of you who want to know a bit about Façade without having to go to the website...

Quote:

Façade is an artificial intelligence-based art/research experiment in electronic narrative – an attempt to move beyond traditional branching or hyper-linked narrative to create a fully-realized, one-act interactive drama. Integrating an interdisciplinary set of artistic practices and artificial intelligence technologies, we have completed a five year collaboration to engineer a novel architecture for supporting emotional, interactive character behavior and drama-managed plot. Within this architecture we have built a dramatically interesting, real-time 3D virtual world inhabited by computer-controlled characters, in which the player experiences a story from a first-person perspective. Façade was publicly released as a freeware download / cd-rom in July 2005. You, the player, using your own name and gender, play the character of a longtime friend of Grace and Trip, an attractive and materially successful couple in their early thirties. During an evening get-together at their apartment that quickly turns ugly, you become entangled in the high-conflict dissolution of Grace and Trip’s marriage. No one is safe as the accusations fly, sides are taken and irreversible decisions are forced to be made. By the end of this intense one-act play you will have changed the course of Grace and Trip’s lives – motivating you to re-play the drama to find out how your interaction could make things turn out differently the next time.

I'm interested in this, but haven't yet experimented with it since I'm not a "Gamespot Complete!" member (I know the bitorrent is available from the interactivestory site). If they want good press for this release, I suggest they offer it for official download at more than 2 sites. :)

Anyway, I'm also a bit confused as to what exactly this project is; I've read too many differing descriptions to fully comprehend it- makes me want to try it even more.

-Razorguts

Whats wrong with bittorrent? I had it down in about a hour.
I only played for 15 minutes and it is interesting.
I will try it more when I have time.

Quote:

Original post by Razorguts Anyway, I'm also a bit confused as to what exactly this project is; I've read too many differing descriptions to fully comprehend it- makes me want to try it even more.

Me too. I've read some of the research of the designers (Andrew Stern and Michael Mateas) when they were at Carnegie Mellon University, but I'm not sure if Facade is similar to their work on interactive characters.

One of the problems with truly interactive stories is the exponenial growth effect; if it takes N hours to design ten minutes of game, it might take 2N hours to design twenty minutes, 4N hours to design thirty, and once you get to a movie length of a hundred minutes you need 1024N hours. It just doesn't scale.

I want to see if they've fixed that problem with the development of Facade. I think it's possible, but it involves some sophisticated character AI. And it's just possible, given the background of these guys, that they might have done it.

One of the things I played with in a recent AI project was trying to get a bot to "learn".

I'm pretty low on the coding knowledge currently (no real titles and I'm not a professional or indy game developer. . . just a student), so I did this by making a file for the AI to write to and recall from. . . simple stuff like remembering my name and using that when asking later questions, as well as asking about interests and things that it learns during the discussion.

Where I got hung up was in trying to get it to learn to program itself as we went along. I could get it to write and pull information, but I couldn't get it to create new answers to to new questions while talking to the player. . . It was ultimately on rails because I could change variables strings in the code but I couldn't initialize new strings and handle initialization, etc. from outside the program.

Anyway, I need to learn to walk before I can run.

Original poster - I read up on this earlier this week when it launched, but I forgot about it. Thanks for bringing it back to my attention. I'm downloading it as I tyep.

I tried this and I'm impressed by some things and not impressed by others.

Subject Matter - The subject is that you are invited to a dinner party held by a married couple having a domestic dispute. . . the problems are inane and both of the characters are pretentious and unlikeable.

AI is a mixed bag - The designers decide to give the player enough freedom to hang themselves. You can say whatever you like, but the AI cannot grasp the meaning of phrases and will frequently misinterpret phrases. There is a feeling of being along for the ride and most, if not all, of the interactivity I had with the two AI bots resulted in negative responses even when I was trying to get the "good" ending. The dialogue tree is very much on rails and as the player you must blindly navigate into the "correct" streams of pre-created dialogue without knowledge on how you are doing it. At key moments the AI will ask you "yes" or "no" questions that give you clear routes, but this is rare.

This lack of control is even more evident as you play through the short game several times. In my attempts to steer the couple into destruction, I only pissed them off more and more. They "wanted" to complain and reveal, but when I tried to get them to go in a truly different route, they merely became increasingly belligerent before finally throwing me out (which was nice, but it was an obviously triggered event).

The thing I found most impressive about the game was how the AI moved about the room and used proximity and facial expressions to show their emotions. I think the game's AI also uses YOUR proximity to characters to show your relative concern (which is a mistake as I generally moved so that I could view both of them and this "choice" was made more by the limitations of the interface than an assessment of my involvement with either of the characters). All of this is pre-programmed of course. Go here to say this, or here to say that, go to other AI's or player's location to say this, etc. However, the effect of moving about the space in somewhat different ways was very effective.

High Quality Voice-Acting - While both of the characters are unlikeable. It is not because their dialogue is poorly written or acted. Both the dialogue and the acting feel natural and very real. This is quite immersive and you definitely feel like the conversations could actually take place if two extremely pretentious, melodramatic, and self-involved married people were having an argument in front of you.

Discussion - some minor spoilers

If they had made this an interactive play that put the characters in real danger or even kept the same idea and just wrote them as characters who were less melodramatic and had problems that were less stupid. I'll go into this further, but just stop reading if you don't want spoilers.

Her problem: she's a supressed artist;
His problem: he's a pretentious yuppie who is ashamed of his lower class family.

This is the first layer of conflict between them and both of them have deeper secrets that are actually much better than the two first layer ones I've listed except that they are clique and their motivation for the secret acts are still from the first layer.

This game would have been tons better (even with it's AI limitations) if they were trying to solve a mystery, or were in mortal peril, or if they had more real problems and the first layer was the behavior instead of the motivation (which is more true to life anyway). It would also be great if the setting were a bar and you were trying to seduce them. . . make the player feel like the protaginist instead of the puzzled observer.

Anyway, this is a bad game with interesting AI that fails utterly on subject matter, but is very interesting and definitely worth playing if for no other reason than to study the AI.

Hey, thanks for the detailed reply dink! I'm going to get a copy of this myself eventually, so I might have to resurrect this threat in a few weeks time once I've had a look at it.

Quote:

Original post by dink I tried this and I'm impressed by some things and not impressed by others. Subject Matter - The subject is that you are invited to a dinner party held by a married couple having a domestic dispute. . . the problems are inane and both of the characters are pretentious and unlikeable.

From what I've followed of the project's development, I think the whole purpose of this is as more of an interactive art project, not really a game. That's why the subject matter of a dysfunctional couple's dinner party was chosen; it's something different from what's out there. Hence I'm not surprised that it isn't really a game, nor that it mightn't even be any fun. It's the AI that I'm really interested in.

Quote:

AI is a mixed bag - The designers decide to give the player enough freedom to hang themselves. You can say whatever you like, but the AI cannot grasp the meaning of phrases and will frequently misinterpret phrases. There is a feeling of being along for the ride and most, if not all, of the interactivity I had with the two AI bots resulted in negative responses even when I was trying to get the "good" ending. The dialogue tree is very much on rails and as the player you must blindly navigate into the "correct" streams of pre-created dialogue without knowledge on how you are doing it. At key moments the AI will ask you "yes" or "no" questions that give you clear routes, but this is rare.

Do you think from your experience that the problem of the AI not grasping the meaning of phrases was a limitation of the text parser, or a more inherent limitation in the dialogue structure itself. You've said there was a feeling that the dialogue was "on rails", which is what I feared; a prescripted dialogue tree fails to be expandable due to that exponential growth problem I mentioned earlier. The presence of "yes or no" story landmarks also seems to confirm my suspicions.

Quote:

This lack of control is even more evident as you play through the short game several times. In my attempts to steer the couple into destruction, I only pissed them off more and more. They "wanted" to complain and reveal, but when I tried to get them to go in a truly different route, they merely became increasingly belligerent before finally throwing me out (which was nice, but it was an obviously triggered event).

That mightn't be a bad thing, from a game design perspective. One of my key models for thinking about interactive game design stems from improvisational theatre, and steering people towards outcomes is a useful tool in that, as long as it isn't too blatant. People just getting angry at you if you steer them in a different route is a neat trick for hiding the limitations of an interactive design. However, it all depends on how cleverly it was done, and I'm not sure how obvious it was to you what was going on. It's all about maintaining the illusion of theatre. [grin]

Quote:

The thing I found most impressive about the game was how the AI moved about the room and used proximity and facial expressions to show their emotions. I think the game's AI also uses YOUR proximity to characters to show your relative concern (which is a mistake as I generally moved so that I could view both of them and this "choice" was made more by the limitations of the interface than an assessment of my involvement with either of the characters). All of this is pre-programmed of course. Go here to say this, or here to say that, go to other AI's or player's location to say this, etc. However, the effect of moving about the space in somewhat different ways was very effective.

Really, I didn't know about the art animation! I've only been seeing screenshots, and I didn't think that artwork was a bit part of the project. I guess I should have guessed from their previous work that they'd include this somehow.

Quote:

High Quality Voice-Acting - While both of the characters are unlikeable. It is not because their dialogue is poorly written or acted. Both the dialogue and the acting feel natural and very real. This is quite immersive and you definitely feel like the conversations could actually take place if two extremely pretentious, melodramatic, and self-involved married people were having an argument in front of you.

That sounds good, although did you think the AI helped increase this effect, or was it down to the skill of the voice actors?

Quote:

Discussion - some minor spoilers If they had made this an interactive play that put the characters in real danger or even kept the same idea and just wrote them as characters who were less melodramatic and had problems that were less stupid. I'll go into this further, but just stop reading if you don't want spoilers. Her problem: she's a supressed artist; His problem: he's a pretentious yuppie who is ashamed of his lower class family. This is the first layer of conflict between them and both of them have deeper secrets that are actually much better than the two first layer ones I've listed except that they are clique and their motivation for the secret acts are still from the first layer. This game would have been tons better (even with it's AI limitations) if they were trying to solve a mystery, or were in mortal peril, or if they had more real problems and the first layer was the behavior instead of the motivation (which is more true to life anyway). It would also be great if the setting were a bar and you were trying to seduce them. . . make the player feel like the protaginist instead of the puzzled observer. Anyway, this is a bad game with interesting AI that fails utterly on subject matter, but is very interesting and definitely worth playing if for no other reason than to study the AI.

I think this is the problem with this approach to interactive stories, assuming they're relying at it's core on a prescripted game tree. There's just not a lot you can do in the space available without it feeling artificial. However, I suspect that their choice of subject matter was done for artistic reasons more that for its entertainment value. Remember that their audience also consists of those studying game theory from the arts side of things, and they'd lap something like this up; it will be part of the Game Studies course in every arts degree across the globe within a year! [grin]

both the text parser and the design structure.

They often didn't understand things properly and additionally it was on rails.

You really should "play" it yourself as it defies in-depth description.

Quote:

Original post by dink You really should "play" it yourself as it defies in-depth description.

Okay, I will. Maybe I'll start up a new thread in a couple of weeks once I've figured out a way to get my hands on a download, or just repost here. There's probably quite a lot to learn and discuss from Facade.

Edit: That doesn't mean that other people who've tried out Facade can't contribute to this thread! Honestly, I'd like to see if there's something here that can be used to spark a revolution in game design.

I downloaded mine via Gamespot.