I.F. Realities (was Re: More rambling)

5 views
Skip to first unread message

Peter Weyhrauch

unread,
Nov 17, 1993, 11:13:33 AM11/17/93
to
On 16-Nov-93 11:42 Ian Cottee writes:

> What I do want is a totally logical game. Lets take an example where
> you have to open a wooden box with a padlock on. The game wants you to
> unlock the padlock with a key but I have an axe. I want to smash that sucker
> with the axe but the game says 'Sorry, you can't do that'. Why not? This is
> supposed to be emulating reality. Why can't I cut my opponents hand off and
> eat it. Why can't I break the knife in half and then scratch my initals
> onto the wooden bench with my fingernails?

> Yeah - I know - its all too much for the programmer to think of ... but
> its what the game player would like to do. Get teams of programmers going

Let us assume that we can't model all of reality in a simulation, and
let us ignore philisophical questions regarding the what's and
whether's of objective reality.

One approach to address Ian's problem is to simulate as much as
possible of the real world. You playtest for a while, cover most of
the cases, but leave inevitable cracks. This is unsatisfying since
the player ends up asking herself "Why can't I do Z?"

Another approach would be to simulate fully something other than the
real world. I think this might be a good choice.

The question then becomes, what is the simplest possible reality that
can be fully simulated that allows interactive fiction to be created.
In this reality, since it is fully simulated, the player can never
logically ask "Why can't I do Z?" since Z is either doable or
inconsistent with the reality.

So, the question for the creative folks in the audience is,
what is the simplest consistent reality for interactive fiction?

Peter Weyhrauch
Oz Project, CMU

Greg Ewing

unread,
Nov 17, 1993, 7:44:13 PM11/17/93
to
In article <CGn92M...@cs.cmu.edu>, pw...@A.GP.CS.CMU.EDU (Peter

Weyhrauch) writes:
|> The question then becomes, what is the simplest possible reality that
|> can be fully simulated that allows interactive fiction to be created.

The term "interactive fiction" has no definition precise enough to
allow this question to be answered.

Seems to me that if you simulate anything other than the real world,
then your simulator *defines* what it is that you're simulating.
What sets interactive fiction apart from other, more abstract computer
games is that it draws on previous human experience for inspiration
in the same way as fictional literature.

Elements of fantasy are often included, but the simulated reality
needs to have something in common with real reality, otherwise
the game is no more "fiction" than Othello or Space Invaders.

I don't think that there's such a thing as a "simplest possible
IF world". A very simple world could be defined, but it would be
so abstract it wouldn't deserve the title of fiction. How much
reality do you need to include before it does? Where do you
draw the line?

|> Peter Weyhrauch
|> Oz Project, CMU

Greg Ewing, Computer Science Dept, +--------------------------------------+
University of Canterbury, | A citizen of NewZealandCorp, a |
Christchurch, New Zealand | wholly-owned subsidiary of Japan Inc.|
gr...@cosc.canterbury.ac.nz +--------------------------------------+

Gerry Kevin Wilson

unread,
Nov 17, 1993, 9:05:33 PM11/17/93
to
Actually, it is quite simple to define IF if you don't quibble over certain
things. One, IF is interactive, hence, the person interacting with it should
be able to affect it. Two, IF is fiction, this simply means that IF involves
a fictional story set down in a literate fashion. This definition is, of
course, sticking to the basics that form the essense of IF. Why no
non-fiction? Simple, if you allow choices into a non-fiction work, then you
have added an element of fiction. Maybe Washington doesn't become the first
president. Or whatever, the two won't go together. There can be ELEMENTS of
non-fiction, but it would have to have a base in fiction. There, simple
enough for my tastes. (Here come the 'buts', I can hardly wait... :/ )

I personally wouldn't be interested in the so-called Usenet Adventure Game.
You can't create good fiction like that. The whole thing needs a unified
purpose and atmosphere to be worthwhile to me. Some people write GAMES. I
WRITE games. The difference is glaring, believe me. Shrug, oh well, have
fun trying to organize it. <echoing laughter that fades away>

--
*=== If there's one thing I've learned in this silly old thing called ===*
*=== Life, it's....umm....oh Hell, I've forgotten. ===*
*=== whiz...@uclink.berkeley.edu ===*
*=== Disclaimer: I am insane. Deal with it. ===*

Peter Weyhrauch

unread,
Nov 18, 1993, 3:37:03 PM11/18/93
to

Greg, thanks for the comments.

I agree totally that "interactive fiction" is not very precise, but I
will assume that from your comments linking fictional literature and
interactive fiction that your definition cannot be far from mine. You
spoke well when you suggested that IF "draws on previous human


experience for inspiration in the same way as fictional literature."

I would like to see interactive fiction aspire to this level.

And this is where my question of simple realities come in. You
expressed a notion ("No, you can't!") in a previous post discussing
the possibilities of text-based interactive fiction. What this
implies is that, as of now, computers will not be able to express the
real world in text. But now we are in a bit of a pickle. We want
"real" characters, since we want interactive fiction to draw on human
experience. Yet, we can't have them, since no computer-based
characters can express an understanding of the real world.

So, my suggested solution to this problem is to pick a life-like world
that both suggests the real world and also is understandable by a
computer-based agent.

You asked the question yourself. "How much reality do you need to
include before it does [deserve the title of fiction]?"

The second half of the question is, Can this reality be understood by
computers?

If we can find a life-like reality that is both rich enough to support
interactive fiction and simple enough to be understood by the
computer, we will have discovered a world in which we can build
literary quality interactive fiction.

So, I ask you all again, what is this world?

Peter Weyhrauch
Project Oz, CMU

Greg Ewing

unread,
Nov 18, 1993, 5:05:45 PM11/18/93
to
In article <CGpFxt...@cs.cmu.edu>, pw...@A.GP.CS.CMU.EDU (Peter

Weyhrauch) writes:
|> If we can find a life-like reality that is both rich enough to support
|> interactive fiction and simple enough to be understood by the
|> computer, we will have discovered a world in which we can build
|> literary quality interactive fiction.

You seem to be asking for a world with a mathematically precise
description. You need that much precision if you want a computer
simulation of it that is demonstrably complete and consistent.

You also want this world to qualify for "interactive fiction".
What I was trying to point out is that there is no correspondingly
precise definition of "interactive fiction". So while the question
of whether a given world is completely simulable is a crisp
one, whether it constitutes "interactive fiction" is fuzzy.

|> So, I ask you all again, what is this world?

There's no one such world - there's a whole lot of them of
varying degrees of complexity, and different people will have
different opinions about whether they are IF.

Jorn Barger

unread,
Nov 18, 1993, 5:53:37 PM11/18/93
to
Peter Weyhrauch <pw...@A.GP.CS.CMU.EDU> wrote:
>
>You asked the question yourself. "How much reality do you need to
>include before it does [deserve the title of fiction]?"

I'm thinking Samuel Beckett, here...

(ie- not a lot ;^)

Phil Goetz

unread,
Nov 18, 1993, 5:59:21 PM11/18/93
to
In article <CGn92M...@cs.cmu.edu> pw...@A.GP.CS.CMU.EDU (Peter Weyhrauch) writes:
>So, the question for the creative folks in the audience is,
>what is the simplest consistent reality for interactive fiction?
>
> Peter Weyhrauch
> Oz Project, CMU

Any implemented IF world is consistent on its own terms,
since every outcome is algorithmically determined.
It's when you try to map the symbol system in the computer
to the real world that you find inconsistencies. What you're
asking us to do is to stop making that mapping. I refuse.
I'm interested in the types of stories that occur in _my_ world,
not in Flatland or some other mathematical construct.

Phil go...@cs.buffalo.edu

Peter Weyhrauch

unread,
Nov 18, 1993, 9:39:19 PM11/18/93
to

Hi, Phil.

Phil, I agree fully when you write that "any implemented IF world is


consistent on its own terms, since every outcome is algorithmically

determined." The wording of my question was poor. I shouldn't have
used the word consistent, because I didn't mean it any mathematical
sense and perhaps that overshadowed my meaning.

I am not making myself clear. I do not want some kind of lifeless
symbolic world that is mathematically sound and consistent. I want
the real world with real characters and real stories.

Greg:

Greg, I think we are speaking the same language. When you say there
isn't one simplest world for interactive fiction I agree. It was just
my way of phrasing the question. I would be satisfied to learn what
you thought a simple world mechanics for your type of interactive
fiction could be.

I understand "interactive fiction" as a topic does have these
ambiguities. I am interested in interactive fiction based on
relationships between the interactor (player) and the characters of
the story, under the influence of some kind of plot, perhaps similar
to plots found in movies. Once again, this description is ambiguous,
but I hope more narrow that the whole field of interactive fiction.
In particular, I am less interested in puzzles and more interested in
experiences with characters.

I would like to hear a discussion about the space of possibilities for
interactive fiction, given current techincal limitations.

Here's two examples of what I mean:

1. For many stories, modelling precisely the behavior of liquids may
be unnecessary. Although this world is inconsistent with the real
world, it may be satisfactory for telling stories. Perhaps it
depends on the exact details of the interactive story, but for most
stories we can safely ignore this detail. I mean, when the player
is getting into some conversation with a character about the plot
point at hand, it hardly seems relevent to dwell on the mechanics
of water. Now, LOVE, on the other hand, is very important to model...

2. I think we can never have satisfying interactive fiction based on
characters until we can really model physics and reality. Unless
our models are very realistic I can't see a player sympathizing with
a character, since they all seem so unrealistic. If the player
can't sympathize with any characters then all is lost. Therefore,
we should be concentrating on puzzle-based i.f. for now, since our
domain will be more limited, and therefore we can do a better job.

Now I happen to agree with opinion one, and disagree with opinion two,
but I want to hear your opinions.

Peter

Erik Max Francis

unread,
Nov 20, 1993, 2:13:41 AM11/20/93
to
go...@cs.buffalo.edu (Phil Goetz) writes:

> I'm interested in the types of stories that occur in _my_ world,
> not in Flatland or some other mathematical construct.

Hmm. The idea of a text adventure in Flatland somehow appeals to me.


Erik Max Francis, &tSftDotIotE ...!apple!uuwest!max m...@west.darkside.com __
USMail: 1070 Oakmont Dr. #1 San Jose, CA 95117 ICBM: 37 20 N 121 53 W / \
-)(- Omnia quia sunt, lumina sunt. All things that are, are lights. -)(- \__/

Jorn Barger

unread,
Nov 20, 1993, 1:53:57 PM11/20/93
to
Erik Max Francis <m...@west.darkside.com> wrote:
>Hmm. The idea of a text adventure in Flatland somehow appeals to me.

No no no no no!!! Gotta have graphics! :^)

------------------ --------- ------- -- ------- --------
^^^^^ you are here
jorn

Greg Ewing

unread,
Nov 21, 1993, 6:02:39 PM11/21/93
to
In article <CGpwpn...@cs.cmu.edu>, pw...@A.GP.CS.CMU.EDU (Peter

Weyhrauch) writes:
|> I would be satisfied to learn what
|> you thought a simple world mechanics for your type of interactive
|> fiction could be.

It probably wouldn't be much different from what we see in IF now.
Perhaps the boundaries could be spelled out more explicitly, so
the player had a better idea of what to expect to be able to do.

On the other hand, fictional literature of all kinds plugs into
the imagination by creating the illusion of a world much bigger
than it really is; making the edges too visible might interfere
with this process and make the game less appealing. I don't
really know if the illusion would be shattered worse by bumping
into the edge expectedly or unexpectedly.

|> I mean, when the player
|> is getting into some conversation with a character about the plot
|> point at hand, it hardly seems relevent to dwell on the mechanics
|> of water. Now, LOVE, on the other hand, is very important to model...

Seems to me that human interactions are *much* more difficult
both to simulate convincingly and to draw clear boundaries
around than physical phenomena.

Maybe if we were modelling dogs instead of humans. Hmmm... Is there
an idea in there somewhere?

|>
|> Peter

Reply all
Reply to author
Forward
0 new messages