[MUD-Dev] narrative

[Joe Andrieu]

Bruce Mitchener wrote:
>Joe Andrieu wrote:

>> The goal here is to get a writer to distill the essence of a good
>> story and use AI to interpret it given the context and actions of
>> the player so that the resulting experience is a coherent
>> dramatic arc built around the natural self-motivated actions of
>> the player.

> Okay, so if we start to look at this, we can see several pieces:

>     * Distilling the essence of a story or plotline
>        * What are the important aspects?
>           * Points of interaction with the world
>           * Points of interaction with the player
>           * Time synchronization points
>           * When and where and how decisions/branches get
>             made.
>           * Plenty of other stuff.

This is a critical design factor, deeply dependant on your
definition of story. The literary community has long realized that a
sequence of actions does not necessarily make a story, but is merely
a history. The game design community often treats action-constrained
quests as stories, which they are not.

We prefer to define story as the experience inside the player's mind
rather than any physical/virtual set of actions. Consider this
conundrum: if the story is truly interactive, then the user can
change its course. If the course can be changed, is each and every
possible course a "story" in its own right?  Certainly, if I changed
the ending of, say, _Star Wars_, so that the Death Star overpowers
the rebels and blows up the planet, it is a DIFFERENT story by
traditional thinking.  So, defining at story based on a particular
set of ordered media (words, images, actions) is fundamentally
limited. So, we model the story as the transition of the player
through the classic story arc--not the character, the PLAYER.

>    * Representing that essence in a way that the computer
>      comprehends.
>       * This was one of the things that has been done in
>         various papers, like the declarative format in one
>         of the papers that originally started this thread.
>       * This is a tough thing because you want it to be really
>         easy as a tool for the end-user (writers) and you don't
>         want them writing code, and so on.

We are looking at Final Draft for inspiration, but don't have any
worthwhile comments (a demo is available for free
download). However, I strongly suggest you separate the
representation from the tool that creates it. Develop a
representation that is easy to use programmatically, then design a
tool that lets creators build it easily.

I believe you need at least four elements defined: structure, theme,
world state, and characters, each of which probably need a different
representation and evaluation system. Keeping these representations
in sync is a non-trivial problem. Heck, just maintaining the
world-state in some knowledge representation and reasoning system is
a hard problem.

>    * Representing the state of the world and how it can be
>      interacted with, in a way that is accessible to the narrative
>      engine. (This is related to some of the previous discussions
>      about the use of ontological models and OpenCyc.)
>    * Keeping track of the events that involve a player or group
>      of players, the stories that they're involved in, the state
>      of the world around them, and then presenting them with
>      things in the gameworld that foster that dramatic arc
>      while still remaining sensible and logical for the player.
>    * The ability to divine the motivation and intent behind the
>      actions of the player.

Very similar to my breakdown above.

> I'm not sure where the industry stands on the last 2 items, and
> I'm not as directly knowledgeable about the first.  But the second
> and third are things that I've been working with.

I would suggest that the last two are at least three: keeping track
of events and presenting players with storybits are two very
different tasks, imo. I think any real system will have them tightly
coupled, but you can see that in today's games events and
world-state are tracked reasonably, while "presenting things that
foster the dramatic arc" is fairly limited.

The final bulletpoint is also extremely important and
challenging. If you can't divine the motivation, how can you provide
a good story? You can't really, not consistently. This is an area
ripe for innovation.

-j

Joe Andrieu  
Realtime Drama
joe at andrieu.net
805.705.8651


_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev