This is fast becoming my favorite mailing list -- low volume, very high
signal-to-noise ratio, plenty of cool musings and fascinating digressions.
Thanks, folks!
My comments and thoughts on the current topic are themselves pretty darn
nonlinear, but I'll try to present them in a quasi-organised fashion:
1. Plotting. Most of you have heard me babble enough about roleplaying
games (RPGs) as nonlinear storytelling, so I'll try and keep this item
fairly short. I've recently been preparing to run a game of _The Babylon
Project_, an RPG set in the _Babylon 5_ universe. The most interesting
thing to me about the _Babylon Project_ rulebook is that it includes lots
of suggestions on planning a story (beginning, middle, and end), including
the idea of a "story chart" which, flowchartlike, maps out possible routes
the characters can take through the episodes. Some episodes are in a fixed
order -- to make any sense, they must be encountered sequentially, and
certain things must occur during them, such as encounters with specific
NPCs (Non-Player Characters). Others can be encountered in a variety of
orders, but the characters must visit each of them before proceeding to the
next stage. (As Bob put it, "serial in which the end conditions of each
episode are the pre-conditions of the following sequence.") Others are set
up to provide potentially useful information or sidelights but aren't
strictly necessary (and can be inserted at any time). I immediately
thought, hey, this sort of chart could be extremely useful in developing a
VRML story...
2. Linearity. As Dennis indicated, the plot as experienced by an
interactor during a single trip through the story is *always* linear.
(Unless you start having the interactor doing multiple things at once,
which could get really messy.) To me, "nonlinearity" means that the way
the story goes depends on events that occur during experience of the story
-- whether those events derive from interactor action, "NPC"
(computer-controlled character) action, or random number generators. But
nonlinearity doesn't have to imply tree structure, as Bob pointed out.
F'rinstance, you can have storylines that branch and rejoin; you can have
episodes that (even if each is internally linear) can occur in a variety of
orders; you can even have a network of interlocking stories. There's a
wonderful trilogy of plays called _The Norman Conquests_ which take place
at the same time, at the same house: one is set in the garden, one in the
living room, and one in (I think) the kitchen, and (to quote Tom Stoppard)
"Every exit is an entrance someplace else." Each of the plays stands well
on its own, but seeing all three of them gives you the whole interlocking
story... a story that could not be presented linearly because some of the
events occur simultaneously in different places.
3. Autonomy of NPCs. John D. wrote: "The alternative is to simulate the
human brain in characters' reactions[....] The downside: it then becomes
impossible to predict the outcome of the story." (And Miriam added that
some of the other characters could be played by other humans for added
depth.) On the one hand, I think more and more that this is the way to go
if you want a story that doesn't have the arbitrary branching feel of a
Choose Your Own Adventure (CYOA) novel. And I don't think an author needs
to know what the outcome of the story will be. On the other hand, one
aspect of many compelling stories is a sense of closure: a dramatically
satisfying ending. If you as author can't predict the ending, it may be
hard to ensure that it's dramatically satisfying. One avenue is to
restrict the NPCs' autonomy: either use tricks to make them appear
autonomous, or produce stimuli within the game world that you know will
cause them to react in (more or less) a particular way. Another avenue is
to make the story engine smart enough to have a good idea of what's
dramatically satisfying and what's not... Though of course this is a hard
problem. (Many successful authors in linear media are no good at endings;
presumably that means it'd be hard to set up an expert system that's good
at them.)
On a related note, various people are working on motivation engines of
various sorts for bots and NPCs. Oz (www.oz.com) keeps track of emotional
state; Motion Factory (www.motion-factory.com) has a state-machine for
keeping track of NPC states; there was a fascinating early Macintosh (or
Apple II?) game in which NPCs' reactions to you depended on how you (and
other NPCs?) had treated them in the past. The pieces exist; it's a
question of putting them all together.
On another related note, if you put a bunch of player characters
together in an open-ended story world and don't require closure -- well,
then you've got a MUD or a MUSH. A valid form -- like a soap opera, where
characters change and grow over time, and have various sub-arcs of their
own, but there's no overarching arc or goal that the story as a whole is
tending toward.
4. Repeatability. Miriam suggested the cool idea of having to play
multiple times to get the full story (kinda like _Norman Conquests_); my
only suggestion on this point is to make each time through the story an
interesting and fulfilling experience in itself, even if you get a much
richer experience by trying again. There's an extended version of Colossal
Cave Adventure that I played in college; in the closing sequence, you had
to recite a series of five magic words. The problem? You could only get
*one* of those magic words each time through the game. You *could not win*
(without cheating and looking at resource files) without playing through
the game at least five times. If each time through the game had added
depth, this would've been okay; as it was, since you had to solve the
entire rest of the game to reach the closing sequence, repeating even once
was an exercise in boredom.
5. Illusions of free will. There are tricks for getting people to make
the choice you want them to make. The "magician's force," for instance, as
I understand it, involves pushing one card (of a fanned-out deck) almost
imperceptibly forward, causing the mark to pay attention to it without
consciously knowing why, and therefore to choose it from the deck. In
RPGs, GMs (Game Masters) often change reality behind the scenes: "Pick door
A or door B. You picked door A? Too bad, that's where the dragon was
hiding." The outcome would've been the same if they'd picked door B -- but
the players never know it. (It's good to have a backup plan in case the
players get a chance to open the other door later on...) Plays havoc with
repeatability -- but nonlinearity doesn't necessarily imply the story will
be worth experiencing more than once. (Text adventures are definitely
nonlinear, but once you've solved all the puzzles it's not all that
interesting to play again.)
6. Quality. Dennis wrote: "Has anyone ever seen a nonlinear story (if you
choose to kill the villan go to page 167) that they thought was any good? I
haven't." I haven't seen a CYOA novel that I thought was any good -- but
I've seen *plenty* of nonlinear stories that I thought were good, in
various ways. (And to varying degrees of nonlinearity.) For yet another
approach to nonlinear storytelling, see the HyperCard "novel" _Uncle
Buddy's Funhouse_ (http://www.eastgate.com/catalog/Funhouse.html; for info
on why I was so impressed with it (though this account contains information
that could be considered spoilers, sorta), see
http://www.kith.org/logos/wander/10.west/Buddy.html) -- no whizzy graphics,
minimal interactivity as such, not even a real *plot* as such, but a fine
example of nonlinear storytelling nonetheless.
7. Media. An essentially linear story can be told in 3D, just as a
nonlinear story can be told in 2D (or in no-D -- how many dimensions do the
human voice and imagination have?). Film allows you to do jump-cuts;
doesn't mean that _Rope_ (shot in ten 8-minute continuous takes) was a
waste of the film medium. I'm nitpicking here; I do agree that VRML is
well-suited to nonlinearity, and that a *totally* linear 3D story (like
_Toy Story_) might best be pre-recorded. But I think it's worth noting
that nonlinearity comes in varying degrees as well as different forms.
F'rinstance, even the relatively mild interactivity provided by IrishSpace
makes, imo, a huge difference to the story.
8. Modularity. Very cool idea from Len about being able to drop in new
behaviors (and possibly new characters, or new motivations) each time
through the story. Might make it even harder to write, but would be a
great effect.
--jed, long-windedly
