From: Bullard, Claude L (Len) (clbullar_at_ingr.com)
Date: 23 June 1998
Jeff writes:
I suppose as A.I. technology advances, and we are increasingly able to
use human controlled avatars in interactive fiction, we will find that
quality of these experiences will increase dramatically. I've been reading
Neal Stephensons "Diamond Age" which depicts a system of interactive experiences that are controlled by actors acting remotely. That's an extreme example of a direction interactive fiction could move in.
It isn't that extreme. Take away the script of VRMLDream and let the the chat window be visible to the actors on a monitor. Put humans in the loop and any kind of improvisatory action with the limits of the device can be used. That's near term stuff and given decent audio streaming, stage management etc, you have theatre online. The story == script. If the script enables interaction (controls) then it is interactive even though it might still be linear (non-adaptive).
If not, we are back to bound linear (sit and wait for the fat lady).
NOTE: In IrishSpace, we used the terms cinematic and free-roamer to account for fixed camera tracking and user-navigable worlds. These are devices only and don't account for interactivity. That is the act of being able to move or not move at will does not make a world interactive. Only controls do that. Only controls with feedback make it adaptive. A control that rings a bell is interactive but the bell must induce an action for it to be feedback.
The question is, can Neal do that without *Humans It The Loop*? HITL is a
good
model and worth pursuing, but like most "live" events has a lot
of restrictions in time, space, and resources. Look at the problems
of getting a regularly scheduled chat event to work and you can see
that the productions get expensive.
Step back one level. If instead of actors you have a player and a
single overlord (eg, dungeonMaster) who can interactively
set values for the production bots, you have a reasonably sized
production. Non-linearity (definition 2: adaptive) requires that surprise
or
unpredictability be enabled. Further, it must not simply exhibit
out of range behaviors (Spock kills McCoy for referring to pointy-ears),
it must exhibit adaptation: alter conditions which result in unpredictable
behaviors.
if (starfleet.crew.bravery !< .8 && this.klingon.commander.honor = 1 ) {
Join (this.klingon.value, this.starfleet.crew) }
else if (starfleet.crew.bravery !< .8 && this.klingon.commander.honor != 1 ) {
Kill(this.klingon.value, this.starfleet.crew) }
and so on. Most of the time, the Klingon commander will kill the crew.
Occasionally, a rare Klingon will join. Let's say we want this to
be dependent on the situation, add a condition which evaluates
the environment object and the situation object. That is, if the
space is enclosed (no escape) and the situation is desperate
(only cooperation defeats common enemy), then and only then
will ANY Klingon commander cooperate.
Given the object relationships, it is possible that by choices made invisibly during the course of navigating the world, surprises are produced (sure, more like a sim but that may be the best model). You could even throw in a wildcard. For example, put prayer wheels (cylinder sensors) at the entrance to an outpost (some temple somewhere). If the user spins them going in, they get a wildcard increase in survivability that simply increases a few positive values, eg, the Klingon commander is a little more honorable. That's a nice example of a hidden coupler (hidden from the user because the act of spinning the wheel may not produce any noticeable effect immediately).
Does the storyteller determine the quality of the story? A storyteller determines the quality of the presentation. I posit, some stories have intrinsic qualities and these are worth isolating. Myths for example, can be retold and repackaged in many forms, times, sets, and settings and will usually get the same reactions.
Len
This archive was generated by hypermail pre-2.1.8 : 28 November 2004 CST