Lambda the Ultimate

inactiveTopic A Declarative Model for Simple Narratives
started 8/6/2002; 2:42:34 AM - last post 8/8/2002; 1:21:55 AM
jon fernquest - A Declarative Model for Simple Narratives  blueArrow
8/6/2002; 2:42:34 AM (reads: 1699, responses: 4)
A Declarative Model for Simple Narratives

A DSL embedded in Prolog for programming narratives. A story consists of a setting and a sequence of episodes. Programs are written by declaring a world model in Prolog clauses. The AI Agent Oriented Programming paradigm provides the logical foundation for modelling human intentions and goal directed behavior as well as the temporal sequencing of events. This DSL might be useful for generating software design Use-Cases automatically given a model of user behavior in a given domain. Others have worked on programming narrative [Survey, Visual Programming with Multi-Perspective Story Agents, Narrative Intelligence], but this is the first open sourced work that I know of.


Posted to Logic/Declerative by jon fernquest on 8/6/02; 2:59:44 AM

jon fernquest - Re: A Declarative Model for Simple Narratives  blueArrow
8/6/2002; 3:23:55 AM (reads: 925, responses: 1)
I know I'm pushing the definition of a programming language by calling this story generation system a DSL. It doesn't have any special syntax, just plain Prolog clauses, but it is made so you can declaratively create new templates for stories, so it is programmable, and it is a little different, combining technical computer programming with the art of writing.

I'm trying to understand the open sourced Prolog code and program new stories with it in my weblog.

Ehud Lamm - Re: A Declarative Model for Simple Narratives  blueArrow
8/6/2002; 4:21:19 AM (reads: 968, responses: 0)
I don't really have a problem with seeing this as related to DSLs. But what I am unsure about is what's the point? Why would you want such a system? (I am not trying to be funny. I really want to know)

jon fernquest - Re: A Declarative Model for Simple Narratives  blueArrow
8/7/2002; 3:51:20 AM (reads: 868, responses: 0)
> Why would you want such a system?

If you want to simulate the behavior of humans.

For instance, in a more sophisticated SIMS game the characters would have to have life stories and experiences if they were actually going to talk about anything.

If authoring was possible in the SIMS, it could be used for language teaching, but the characters would have to be made more complex than ants. One of the main problems language teachers face is bringing reality into the classroom and providing their students with something to talk about. They are forever generating what-if types of situations and stories.

Teachers already use simulation games to get people to start thinking and talking about a subject. The Stanford Business school uses them in their banking course.

Another application would be generating use-case stories for the users of a software system. I've done consulting work integrating payroll systems into general ledgers and consolidating the general ledgers of multi-nationals and in every department managers have different sets of goals (issue pay checks, month end closing, produce reports, etc...) that all have different deadlines, all of which is hashed out in innumerable meetings (ugh) There are some quite complex managerial coordination stories to be told. If you could create a couple of different possible stories and just have people check the applicable one or modify it a little or use them as the basis of discussion......people can relate to stories easily and use them to reason about a problem.

Economic simulations where expectations play a crucial role also come to mind. .

I was very happy to see that Lang actually had made some headway into writing computer programs that generate reasonable stories, obviously there's a lot left to be done.

jon fernquest - Re: A Declarative Model for Simple Narratives  blueArrow
8/8/2002; 1:21:55 AM (reads: 860, responses: 0)
It would be cool to see more innovation and experiment in the open source world. Bruce Sterling's recent hilarious speech at the 2002 O'Reilly Open Source convention [Speech, Slashdot] criticizes the lack of creativity in open source:

There's nothing newly created. Even free software guys, who like to spend a lot of time talking about grand community-building schemes, spend most of their working time aping commercial products. That's what they do. "We've built something that can interoperate with Microsoft!" That's like sticking banderillas in a bull, when the world really needs at this point is something like... a piping-hot catfish dinner.....

There's a noticeable lack of basic creativity in the free software world, that is alarming and not very flattering. People in free software still have a basically piratical state of mind....

"GNU's Not Unix." Okay, you're "not Unix" but what are you really? Why do you have to live in that shadow? The shadow of this other enterprise. There's something basically juvenile about that. Something that is unworthy, creatively feeble, childish.

"Aping" new innovative ideas in PhD theses that don't have any commercial existence yet might be one way to overcome this impasse.