The Rut

Document Sample
The Rut Powered By Docstoc
					The Rut
If we were to compare the state of computer entertainment with an
analogous stage of development in film, it would undoubtedly be most
appropriate to compare the vast majority of video games, even the best,
to the sci-fi and western serials of the 1930s. Occasionally a memorable
character will emerge (and no, this does not mean a graphically-
recognizeable chracter, but a relatable one). Once in a while a non-
linear plot will be written. But by and large, characters, such as they
are, exist to drive the action, and the action is blowing stuff the hell
From the elaborate streets of Grand Theft Auto, to the sprawling fantasy
world of Oblivion, to the claustrophobic hallways of Splinter Cell, each
experience ultimately maintains a porn-like fixation on a unified goal:
to take out the enemy, save the world, and score the loot. This, needless
to say, is the same essential structure of the vast majority of games
since Space Invaders. Thopugh graphics and physics keep getting better,
the narrative structure of games is in a gigantic rut.
This month, The Atlantic magazine ran an article describing the
development of a game called Façade (or Facade for those of you with
different character encoding). The game was written this year by Michael
Mateas and Andrew Stern, formerly an AI scientist and a virtual pet
developer, respectively. They set to work on a verbal scripting language,
built on a more powerful series of algorithms, for developing a game in
which it would be possible to subtly communicate with human-like
The result was a twenty-minute non-linear experience in which the player
is invited to the house of a feuding couple in the midst of some sort of
argument (what sort, exactly, varies each time). The player then types
responses to various questions, insults, and goadings thrown around, make
comments on a statement made by the husband, for example, and receiving a
snide comeback in return. There is no dialogue tree; the player can say
whatever he or she chooses, and expect a cogent reply.
So far, the response has been that it works...most of the time. Outside
of certain limits, the couple will not discuss everything the player
might bring up; a query concerning the epistemic certainty of the
existence of God will likely produce only unsure laughs from the couple.
But the fact that they can laugh in an unsure manner, or "feel" awkward
at all, is a fantastic leap forward.
The next project for Mateas and Stern will be a $2 million commercial
game called The Party. Essentially Façade on a much larger scale, it will
allow the player to maneuver through a high-society dinner party,
flirting, betraying, or coercing the other guests into various forms of
(hopefully) dramatic behavior.
But then there are the ambitious ideas: after a few more iterations of
this basic concept, Mateas and Stern aim to write software that will
generate an entirely functional dramatic world based on a few criteria
input by the player. Want to be a Victorian detective and solve a murder
mystery? If their concept gets off the ground, the software will write an
original mystery, replete with deep, complex characters, and drop you in
medias res.
Their timeline claims that this will be a reality in twenty years, but we
all know that in the world of technology, the future catches up to us
rapidly. In articles from the early 1980s, programmers were speculating
that by the year 2000, we might be able to render still paintings on the
screen to accompany their text adventures. That happened by 1985; by 2000
we had Ocarina of Time. In short: with the application of the right
minds, this sort of software will arrive sooner than we think.
Now it ought to be time for me to play the Devil's Advocate, and complain
about the likelihood that such algorithms may someday replace writers
altogether. Actually, it seems rather unlikely, at least within my
lifetime. Twenty years brought us from Asteroids to Wing Commander, but
better graphics and sound do not a Star Wars experience make. It's
intimidating to think of engineered narratives, but if anything, I would
hope it'll push human writers to greater levels of skill.
If you're competing with a cheaper, faster mechanical author, you'll be
much more pressured to exceed expectations. And both Hollywood and
Silicon Valley could use more of that these days.