[MUD-Dev] Geometric content generation
Paul Schwanz
paul.schwanz at east.sun.com
Fri Sep 21 11:40:08 CEST 2001
Eric Rhea wrote:
> One problem that I often hear whenever content generation by the
> players is discussed is that of control. I don't think there is
> some generic rule that could be applied here without considering
> how much of the world is setup. I do think that in the majority of
> cases it would be possible to allow the players to add content,
> but there are some cases where it would be exceptionally difficult
> to pull off and not bring over a hoard of issues.
In thinking about the tension between wanting to allow players to
generate content while still maintaining a semblance of control, I
was suddenly reminded of Chris Crawford's "Process Intensive
Designer." I read an essay of his a number of years ago which
addressed a different tension, but one that I think has much in
common. I'll post the entire essay here, because I think it is such
an interesting piece. It seems to me that if process intensive
designers are able to tell interactive stories, they might also be
able to ensure some level of quality in player-generated
(interactive?) content.
Begin Essay by Chris Crawford---->
Interactivity, Plot, Free Will, Determinism, Quantum Mechanics, and
Temporal Irreversibility
Yes, believe it or not, these six things are all tied
together. Moreover, they're tied together in a way that reveals some
useful truths about designing interactive stories. In this essay,
I'll trace those connections.
The starting point of the discussion is the conflict between plot
and interaction. There are theoretical reasons for this
conflict. They are best seen from the point of view of the plot
faction. Many of these people are writers. Plot creation is, from
their point of view, an enormously difficult task demanding great
talent and creative energy. The thought of allowing an audience to
mess up their carefully crafted plots leaves them cold. Knowing how
difficult it is to get a plot to work well, they realize that any
intrusion by the audience into the process will only yield
garbage. If interactivity requires the audience to involve itself in
the direction of the plot, then clearly interactivity and plot are
incompatible.
Adding to this apparent incompatibility is the attitude of the other
side. The protagonists of interaction tend to take a dim view of
plot. The strongest example of this is the possibly apocryphal story
about id software and the creation of Doom. There was, so the story
goes, some dispute within the organization about the proper role of
story in the game. One faction argued that there should be some
story element to tie everything together. The other faction argued
that Doom was to be an action game, pure and simple, and that "we
don't need no steenking story". Eventually, the anti-story faction
won out, the losers left the company, and nowadays story is referred
to within id as "the S-word". So the story goes.
Consider one of the most powerful storytelling products to appear on
a CD: The Madness of Roland. (see "Review: The Madness of Roland")
This was a story with no interaction whatsoever. It would seem that
the author of The Madness of Roland had said to himself, "we don't
need no steenking interaction".
What's particularly interesting is that plot and interaction seem to
contradict each other in the sales figures. The top games of the
last year have been games with all interaction and no plot (Doom II)
or games with all plot and no interaction (Myst, 7th Guest). Could
it be that there is no workable middle ground?
related essay: Plot versus Interactivity
So what we have here is an apparent incompatibility between plot and
interaction. It would seem, from both theoretical considerations
and direct experience, that plot and interaction cannot be
reconciled. This in turn implies that the dream of interactive
storytelling is a chimera.
The central issue that we face here is not new. In slightly
different terms, some of the brightest minds in human history have
struggled with this problem. The results of their efforts might
prove illuminating. Now, you might wonder how a problem in game
design could have attracted the attentions of august thinkers in
times past, but in fact they weren't concerned with games. They were
working with bigger problems. OK, I'll stop being so coy: I'm
talking about the classic theological problem of free will versus
determinism.
How did we get from games to theology? What's the connection between
"plot versus interactivity" and "free will versus determinism"? It
goes like this: God is omniscient and omnipotent. Everything that
happens in the universe happens according to His benevolent
design. There are apparent evils in the universe, but these are all
part of God's greater intentions. But this must include the actions
of people as well as the actions of natural phenomena. Thus, a
terrible disaster is an "act of God", but so is a murder. How then
can human beings have any free will? They are pawns in the hands of
an omnipotent God. If we did have free will, then God would be
neither omnipotent nor omniscient, for then He would neither control
nor know what we would do. But if He is neither omnipotent nor
omniscient, how can he fit any definition of god? Thus, free will
clashes with determinism.
The connection with games should be obvious. Determinism in theology
is analogous to plot in storytelling. Free will corresponds to
interaction, for how else can a player interact without the exercise
of his free will? Indeed, we can make the analogy more explicit by
viewing the creative person as the Creator of a miniature
universe. The storyteller, for example, creates an imaginary
universe populated by his characters. Like some omnipotent god he
decides their actions and predestines their fates. To reverse the
analogy, the history of the universe is nothing more than a huge
story written by God that we act out.
But wait! The games creator is also a god of sorts. He too creates a
tiny universe and exercises godlike control over that universe. Yet,
free will seems to exist in the game universe. What is the
difference?
At this point, some people step forward with the observation that
free will in the real world could be an illusion. After all, God
would want us to think that we have free will, but in fact has
already determined our actions for us. We think that we are making
our own choices, but in fact our choices were predestined. Even if
we try to assert our free will by deliberately making apparently
arbitrary decisions, that too could be explained as God's plan for
us.
Now, the debate over free will versus determinism took a new turn
about 70 years ago with the Heisenberg Uncertainty Principle and the
introduction of quantum mechanics. The Uncertainty Principle
established that the basic behavior of the universe was
fundamentally random. That is, the most basic processes that
underlie the functioning of the universe are unpredictable. This
blows determinism right out of the water. If you can't even be sure
where an electron is or where it's going, then you certainly can't
be sure what a complex system like a human being will
do. Predestination just went down the tubes.
But this was not an unalloyed triumph for free will. Quantum
mechanics replaces determinism with randomness. We aren't
predestined to go to hell; it's all in a flip of the coin. That
doesn't make you feel any better, does it?
Quantum mechanics also had another consequence: not only did it
shatter determinism, it also shattered temporal reversibility. This
is the notion that the laws of physics can work backwords in time
just as well as they work forwards. Before quantum mechanics,
physicists were embarassed to admit that they could not explain why
time always moves forward. In the entire structure of physics, there
wasn't a single fundamental reason for time to be unidirectional.
The fact that time is unidirectional was a baffling reality for
physics. But quantum mechanics changed all that. (Warning: at this
point I am expounding personal opinions rather than generally
acknowledged truths.) For example, the final destruction of
Maxwell's Demon (an imaginary creature who violated the Second Law
of Thermodynamics, thereby challenging the unidirectionality of
time) was not accomplished until Leon Brilloun used
quantum-mechanical arguements to finish him off.
But it's easier to see the relationship between quantum mechanics
and temporal irreversibility if you think in terms of the
Uncertainty Principle. This principle establishes that information
knowable about the universe is finite. Now combine this fact with
the knowledge that information "draws interest". That is to say,
information gained about a physical system at one time can be
combined with information obtained about that system at a later time
to gain even greater knowledge of the system, in a manner that
exceeds the simple sum of the measured information. The longer you
wait between measurements, the more "interest" (additional
information) you can earn from a second measurement. This of course
would permit you to gain gigantic amounts of information about the
universe, thereby violating the Uncertainty Principle. The
resolution to this apparent quandary lies in the fact that
information "degrades" with time. If you gather information about a
physical system, and then gather more information at a later time,
you won't be able to meaningfully combine the data from the two
measurements because the system will have randomly changed in ways
that render the combination useless.
The upshot of this is that the Uncertainty Principle establishes
temporal irreversibility. Time has an unambiguous arrow defined by
the necessary degradation of information arising from the
Uncertainty Principle.
So what does any of this have to do with game design? Well, let me
talk about temporal reversibility. Computer games permit temporal
reversibility within their universes. How? Simple: you play the game
for a while and reach a critical juncture. You save the game, then
choose Door A. Woops, you got eaten by an orc! No problem &emdash;
just reload the game and avoid Door A the second time through. In
effect, you went back in time and changed your decision.
Note that this action proves your possession of free will. If you
choose path A, then one could argue that you were predestined to
choose path A, but if you go back and choose path B, then there can
be no arguement about predestination. Temporal reversibility allows
us to prove free will.
Interestingly enough, I have heard some designers &emdash; mostly
people with noninteractive backgrounds &emdash; complain about this
reloading option and suggest that the games should be designed to
obviate the possibility. It would seem that some people have it in
for free will.
Note further that we don't get temporal reversibility in the real
world, which means that we cannot use it to prove our free
will. Falling back on the theological discussion, this suggests that
temporal irreversibility is God's kluge to cover up His decision to
deny us free will, but allow us the belief that we possess it. If we
could go back in time and change our decisions, then we could prove
that we have free will. The fact that we can't suggests that maybe
we don't... right?
Thus, we see theology, physics, and game design all brushing elbows
on the issue of free will and determinism. Indeed, the intellectual
possibilities here suggest that a merging of game design with
theology could yield an exciting new field: experimental theology. I
wonder if I could get an NSF grant...
As it happens, however, there is another resolution to the problem
of free will versus determinism, one that embraces physics and
rationalizes faith. It says that God is omnipotent with respect to
process, not data. That is, God controls the universe through His
laws, but not through the details. God does not dictate the position
and velocity of every electron and proton in the universe; instead,
He merely declares, "Let there be physics" and then allows the
clockwork of the universe to run according to His laws. In an
indirect way, we could say that He does control everything that
happens in the universe, but it is only indirect control.
This realization provides us with the resolution of our apparent
conflict between free will and determinism. God determines the
principles under which the universe operates, but grants us free
will to choose as we wish within that universe. He even works a
little randomness into the system to insure that we aren't
automatons responding robotlike to our environments. The important
point is this: God is a process-intensive designer!
Expansion on this idea: How to Play God
And the same resolution works with the apparent conflict between
plot and interactivity. If you are a data-intensive designer, then
you are necessarily a deterministic one. Like some Bible-thumping
fundamentalist, you insist that every single word you write must be
obeyed literally by the characters in the story. The fundamentalist
focusses all his beliefs in the data of the Bible rather than the
processes behind it.
But if you are a process-intensive designer like God, then the
characters in your universe can have free will within the confines
of your laws of physics. To accomplish this, however, you must
abandon the self-indulgence of direct control and instead rely on
indirect control. That is, instead of specifying the data of the
plotline, you must specify the processes of the dramatic
conflict. Instead of defining who does what to whom, you must define
how people can do various things to each other.
Perhaps you object that this is too esoteric, too indirect to allow
the richness of tone that a good story requires. If so, consider
what a story really communicates. A story is an instance that
communicates a principle. Moby Dick is not about a whale; it is
about obsession. Luke Skywalker is a lie, but the movie's truths
about growing up and facing the challenges of manhood are its real
message. Stories are literally false but they embody higher
truths. The instances they relate never happened, but the principles
they embody are the truth that we appreciate. They are false in
their data but true in their process.
Given this, consider the nature of the communication between
storyteller and audience. The storyteller seeks to communicate some
truth, some principle of the human condition. Rather than
communicate the truth itself, he creates a particular set of
circumstances that instantiate the truth he seeks to
communicate. This instantiation is what he communicates to his
audience. The audience then interprets the story; it induces the
higher principles from the story's details. Note, however, the
indirection of this process. The storyteller seeks to communicate
some truth of the human condition; the audience seeks to learn the
same. Instead of just telling the principle, the storyteller
translates the principle into an instantiation, then communicates
the instantiation, then the audience translates the instantiation
back into a principle. This is truly a roundabout way to get the job
done.
Interactive storytelling differs from this process in two
fundamental ways. First, the process of translating principle into
instance is delegated to the computer. The storyteller retains full
artistic control, but must now exercise that control at a more
fundamental level. The basic process of translating principle into
instance is retained, but is now performed by the computer. This of
course entails considerable effort in algorithm creation. The second
fundamental difference is that, because the story is generated in
realtime in direct response to the player's actions, the resultant
story is customized to the needs and interests of the audience, and
thereby more than makes up for any loss in polish with its greater
emotional involvement.
Some may object that this is great theory, but in practice, the act
of reducing storytelling to grand principles is beyond human
intellectual ability. Nobody could ever handle so deeply
intellectual a process, the critics cry. Yet this process-intensive
style of storytelling is done all the time, and by amateurs, no
less. Here's Grandpa taking little Annie up to bed:
"Tell me a story, Grandpa!" she asks.
"OK" he replies, "Once upon a time there was a pretty little girl
who had a pony..."
"Was it a white pony?" Annie interrupts.
"Oh, my, yes, it was as white as snow. It was so white that the
sunlight reflected off its coat dazzled the eye. And the little
girl and the pony would go riding along the beach..."
"Did they go riding in the mountains too?"
"Why yes, as a matter of fact, they did. After riding along the
beach, they would ride up the green canyons, jumping over the
brush and ducking under tree branches, until they came to the very
top of the mountains. And there they would play at jumping over
boulders..."
"I don't like to jump."
"Well then, instead of jumping, she would let her pony graze in
the rich deep grass on the mountain's summit while she sat in the
sun..."
And so the story goes on. Note that Grandpa does not respond to
Annie's interruptions with "Shuddup, kid, you're messing up my
carefully prepared plot!" He wants those interruptions, his
storytelling thrives on them. Grandpa does not enter the room with a
carefully planned and polished plot, all set to dazzle Annie. He
comes in with basic principles of storytelling, and then he makes up
the story as he goes along &emdash; in response to Annie's needs and
interests. The story that he creates is his very special story,
just for Annie and himself, and no other story will ever be the
same. Because it is their very special story, it means more and has
more emotional power than any high-tech Hollywood extravaganza. Yes,
it lacks the careful plotting, the intricate development, and the
glorious special effects of the Hollywood product. But its roughness
is more than compensated for by its customization. Sure, Annie likes
The Lion King &emdash; but she treasures Annie and the White Pony.
Now, if some schmuck of an amateur storytelling Grandpa can pull
that off, why can't we bigshot professionals do the same?
<-----End Essay by Chris Crawford
Disclaimer: All theological views expressed in this essay are
Mr. Crawford's and do not necessarily reflect my own personal
beliefs. ;-)
--Phinehas
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev
More information about the mud-dev-archive
mailing list