Posted by Michael on November 14th, 2006, in Books

This is a very fun book 😀 with alternating text and illustration pages, both executed in a witty entertaining style. Probably quite on purpose given the title of the book.

Raph Koster is a game designer (he worked on Ultima Online and other MMOs). So he knows what he’s talking about. The premise of the whole book is an attempt to convince his grandfather that being a game designer is a worthy profession.

The book makes two important statements, in my opinion. First that fun is all about learning. And second that for games to become a mature and respected medium, they need to produce art.

Mr. Koster speaks of learning is pattern recognition: we enjoy finding patterns in the things that we perceive. When we have discovered the pattern, we get bored and desire new challenges. A game that can offer a continuous stream of such challenges, is fun. It is fun because it teaches us things. As a (in my opinion profound) side note, he suggest a definition of beauty as something that is just a little bit off of being completely recognizable. I do think, however, that he underestimates the severity of an aesthetic experience, defining it as something fleeting, something we always ultimately get bored with.

But while he defines the core of games as being their structure of rules and challenges, he does realize that this core alone is not sufficient for an entertaining experience. This is where he sees the role of narrative, comparing games to dance where the choreography is nothing without dancers and a stage, or music that requires musicians who interpret and play the piece.

When it comes to art and expressiveness, he doesn’t really see a problem with the fact that most games are only about violence, claiming that most art, at its core, is about violence and sex. So all games need to do to become art, is to be a bit more subtle about these topics. He does seem to believe that this subtlety can be achieved by creating smart game structures, in other words by manipulating what he sees as the core of games. He makes an interesting comparison between art and a trellis with plants: humans and their experience are the plants that grow on the trellis, while the trellis itself is the artwork.

No need to say that I don’t agree with this. First of all, I don’t like reducing everything to its base components. Sure, all art is about sex and violence. As much is life is about cells and electricity. But at that level, nothing is meaningful! Depth of experience situates itself in much higher regions. I believe that the core of just about everything is nothing. What matters, what truly matters, is the layer around that core, the skin. This is where meaning happens, where humans can become more than organic machines. Truth is completely irrelevant.

And, as you may know, I have my doubts about the potential of game structures as well. I would love to see these expressive games. But if in the thousands of years that games have existed, no one has succeeded in producing a single art piece, why would they suddenly be able to do so now? There is definitely artistic potential in the the new digital technologies but I believe that we need to create much wider and diverse interactive experiences to get there, and drop the requirement of a game structure.

No Comments »

Posted by Michael on November 14th, 2006, in Development

You thought this would become one of the many abandoned AI sites, didn’t you? 😉

Nono. We’re back. We’re done with all the silliness in our lives and can get back to the serious task of building Drama Princesses.

Expect a few posts soon. And some demo software by the end of the month.

1 Comment »

Posted by Michael on September 12th, 2006, in Development

I’m afraid that the recent silence here will probably continue for another month or two. Some urgent projects are demanding immediate attention. But we will be back!

No Comments »

Posted by Michael on July 31st, 2006, in Development

There’s two things that our Drama Princess actors can’t do yet. They can’t be faithfull to a single person and they can’t quarrel for a moment and make up.

They are programmed to love (or hate) everyone with equal passion and loving one person very much wouldn’t stop them from loving another very much. Perhaps it should. This only really applies to romantic relationships that imply monogamy in most narratives. Other types of love don’t necessarily require faithfullness. At least not for the sake of believabilty. Seeing someone betray their best friend can be interesting to witness. But betraying your romantic interest is not something that should be done lightly. Unless the actor is just that kind of person.
So the highest level of intimacy of a relationship should be treated with care. It would make sense if it wouldn’t be easy for actors to achieve such a relationship with each other. So the chance of doing such interactions should be smaller than doing the others. And when you have reached that point with one character, it should become even harder to reach it with another.
For characters with extreme faithfulness, reaching the highest intimacy with one character, may imply a reduction of that intimacy with another, should it exist. Also, if one actor catches another in the act of betrayal, their relationship should drop instantly if it was very high (this would cause a “Shock”).

Another thing that is problematic is quarreling. Drama Princess actors can like or dislike each other. And they can build up relationships of both love and hate. But they can’t temporarily have a falling out and then get back together. When they start arguing, their relationship is shot and they can only get it back slowly and with luck.
Quarelling is a very interesting interaction, though. It’s kind of like rolling the dice on a relationship. If characters start quarreling, the outcome can be one of two things: either they make up and their relationship becomes a lot stronger. Or they don’t and their relationship becomes a lot weaker.
So perhaps “quarrel with me” should be an opportunity offered by a character with a reasonably high level of intimacy. And rather than simply improving the relationship if the Attitude was positive, the system would randomly decide whether the relationship improves or worsens. After which, the characters should express the result by hugging or showing anger.

10 Comments »

Posted by Michael on July 30th, 2006, in Development

I made a mistake. Attitude is the way an Actor feels about a Relationship: does he want to improve it or worsen it, positive or negative. There’s an Attitude value for each relationship in the game. Attention Span is a property of an Actor. It declines as an actor continues to interact with the same object, reflecting the Actor’s boredom. There’s only one Attention Span for each Actor.

While Attention Span and Attitude work together to modify Fondness when making decisions about choosing an Opportunity, they cannot be folded into one variable because there’s only one Attention Span for each Actor and many Attitudes (one for each relationship).

Oops. 😳

1 Comment »

Posted by Michael on July 30th, 2006, in Development

This is another stream of consciousness, an account of a thought process that actually arrives at a solution. I’m just using this blog to think “out loud”. As you can tell, it really helps.

I love it when a plan comes together.

In the first model of Drama Princess, Consumerism, our actors are very selfish. They choose to do the things that give them the most pleasure. In the second model, Romanticism, they prefer to do things with objects that they like and they would choose to do things with these objects that feel comfortable (i.e. that match the relationship that they have with them the best).

The inevitability of having a relatively small set of animations leads to a problem with this, however. When an actor meets a character that he really likes, he will choose to perform the interaction offered by this character that matches their relationship best. The number of interactions offered by this actor will be limited (by the number of animations that we have made and behaviours that we have programmed).

Suppose the character only offers three interactions:

1. one for when you hate him,
2. one for when you’re indifferent, and
3. one for when you like him.

If you like the character, you will choose the third opportunity. But after you’re done, your Fascination with this opportunity will drop to prevent you from repeating yourself. If forced to do another interaction, the next best match would be the opportunity that matches a neutral relationship. And in a third step, you would end up doing the thing that is appropriate when you hate the character. Even though you like him!

This problem can be solved by allowing the characters to do nothing if there is no opportunity that sufficiently matches their relationship.

But what if there is a second object nearby? Say another character that you are fairly indifferent towards. This character also offers three similar opportunities. So you first choose the interaction with the character that you like that feels comfortable. When this opportunity disappears, due to a temporary decline in Fascination, the best match in the area would become the indifferent interaction with the other character. Ergo: rather than hanging around with a character that you like and doing nothing in particular, you would walk over to a character that you don’t care about, to do something with him that is neither here nor there!
If you’re a child you might get away with that kind of behaviour. But for an adult it would look odd.

So it seems like one of Auriea’s and my private one-liners, “It’s not about comfort, baby”, may apply to Drama Princess! 😯

When you like somebody, you want to be around them. Even when there’s nothing in particular that you want to do with them. Just “hanging out” is quite sufficient. Only after some amount of time, you would be inclined to move away to interact with an object that you like a lot less. After which you would return. The latter would already happen in Drama Princess, thanks to Fascination. The only thing we need to add is a way to lock an actor to an object even when they are not interacting with each other.

We already have a value for Attitude. Attitude declines as boredom sets in: each time you interact with the same object, attitude decreases. This allows you to switch to another object when you get bored. Currently, Attitude is maximized, each time you start interacting with a new object. And the inclination to switch to another object -which is what we’re trying to prevent here- increases as you continue to interact with the same object.

Note that as per a later post, the term "attitude" should be replaced by "attention span".

Making the size and speed of the decline of Attitude relative to how good the relationship is, would allow us to differentiate between objects that you like and objects that you don’t: the more you like an object, the longer it would take to get bored with it. Now we need to add a way to actively prevent the actor from switching objects, based on the size of the Attitude, perhaps.

This reminds me of the Urgency attributed to Opportunities offered by objects that are very closeby. If an object is closeby, the urgency of its Opportunities forces them to appear in the final selection of appropriate Opportunities, even if they don’t seem very comfortable. We could expand the Urgency attribute to also take the level of relationship into account, next to distance. And in a similar way: only very close relationships will cause Urgency.

Interesting to see how closeness is used by the same function in both literal and figurative sense. 🙂

This relies on the existence of opportunities. And the limitation in the number of opportunities is exactly what caused our problem. This limitation is caused by the limited amount of animations and behaviours. But if we create opportunities that do not rely on additional animations or behaviours, then the problem is solved!! 😀

So, rather than offering three Opportunities to choose from based on which one matches your relationship best,

1. one of those Opportunities would only be chosen if they match your relationship well enough
2. a fourth Opportunity “hang out with me” would be added. “Hang out with me” feels comfortable for all relationships that are sort of good. So even if “kiss me” is a bit too much for your relationship, or if “kiss me” has been done already and you’ve lost your Fascination with it, you would still be able to choose “hang out with me” rather than be forced by the system to go and interact with another object.

And hanging out is nothing more than actors acting out their idle behaviours in each other’s vicinity. Life is good.
Since they are still interacting with each other, the actors’ Attitude will decline. Ultimately boredom will drive them away from each other.

1 Comment »

Posted by Michael on July 30th, 2006, in Development

Now that we’re making the animations for our Drama Princess (with the aid of the very talented Laura Smith), I’m starting to worry more about the amount of animations that is required to make a characters seem believable. We have already concluded previously that that amount is relative to the amount of realism in the appearance of the characters. In a more abstracted style, you can get away with less variation in animation and still retain believability. It’s as if everything becomes symbolic and the viewer translates things to reality in his or her imagination.

But for the games that Tale of Tales wants to make and the stories we want to tell, an extremely cartoony style (as in Animal Crossing) is not suitable. Nevertheless due the limitations of technology, budget and time, we will end up with a relatively small set of animations. According to the logic above, this would require us to stylize the look of our characters somehow. Perhaps we can find a way to do this that is compatible with storytelling aspirations…

I’m thinking of film noir and underacting, where the actors are being “cool” and flegmatic for reasons of style. The actors look realistic, but they behave in a stylized way. This particular form of stylization is based on reductionism and minimalism. And not on the exaggeration and caricature that are common in games (and animated films). I’m not sure, however, whether Humphrey Bogart really has a small set of animations or a very large set of very subtle ones.

Come to think of it, to some extent the appearance of the actors in those old Hollywood movies, is in fact not as realistic as one might assume since they are clearly real humans. The abstraction happens on the level of their clothing. As was the fashion of that day, the clothing that people wore was very formal, very stylized. They tend to wear simple suits and modest dresses. They walk with a straight back and their polite behaviour supports the abstraction of their formal appearance. In fact, a lot of the romantic tension comes from the conflict between this controlled comportment and the passions that we know rage inside of them.

So maybe the solution for the animation problem in Drama Princess is to make our actors very polite. 🙂
( They’re starting to sound like children that we need to educate 😉 )

4 Comments »

Posted by Michael on July 27th, 2006, in Development

Having actors look at things in their environment, greatly improves believability. Perhaps a system to do this could be as simple as to trigger a “look at object” action every time an object enters a given radius around the actor (or semi-circle in front of the actor). This would happen both when objects approach the actor as when the actor moves through the environment.

If this starts looking too nervous, or if there are too many objects, perhaps the actor can choose to look at objects that he likes only. Or at objects that he encounters for the very first time (curiousity). And he should probably differentiate between inanimate items and other characters. Objects that approach him will probably also provoke more response than object he approaches (startled).

No Comments »

Posted by Michael on July 25th, 2006, in Development

I had made a quick test behaviour of one actor walking up to another actor and forcing the other actor the face him. The latter was done by plainly choosing a behaviour for this other actor. This type of logic may be useful for objects (to force a rock to fall on the floor, e.g.) but an actor should have more control over its own behaviour.

A better way to achieve this would be for the initiating actor to broadcast a “If I’m close to you, face me” opportunity. Then the other actor can decide whether or not he wants to do this.

However, in this case (walking up to an actor), the request to “face me” is quite urgent. Perhaps instead of telling the other actor what to do, a “shock” should be sent to him, forcing him to stop his current behaviour and pick a new one.
In the current design, the other actor would only choose to face you, if he was very fond of you or if facing you is something that feels comfortable. The risk that the other actor chooses another behaviour, with you or with another object, is too big.

To solve this, a Shock should not only startle the actors and force them to change their behaviour randomly. A Shock should be accompanied by a suggestion (in the form of a behaviour ID), accompanied by a gliding urgency value between appropriate and imperative.

1 Comment »

Posted by Michael on July 24th, 2006, in Development

There’s an interesting difference in the general social behaviour of men and women. Women tend to pay more attention to each other and try to achieve some sort of consensus about everything. Men tend to be more confrontational and they leave each other alone more. A woman will look at another woman tying her shoelaces, a man will look away, leaving the other to his or her own business. This is probably why women have received the reputation of babbling all the time. They seem more social in their behaviour. While for men it’s a matter of respect for the other, to not seek so much contact. Women look at each other, men look in the same direction.

To translate this to Drama Princess, opportunities would need to be marked as more feminine or more masculine (or perhaps more or less social and contact-seeking) and the actor would need to know whether it is male or female, so that they can tend to choose things that fit their gender better. Also, more “decorative” actions could be added: female characters can look at other characters more quickly, male characters could try and create a bigger distance from other characters.

4 Comments »