IF References

4 views
Skip to first unread message

David Graves

unread,
Jan 2, 1991, 5:59:26 PM1/2/91
to
Breck Baldwin (br...@linc.cis.upenn.edu) posted, saying:

> I'm somewhat of a neophyte in the world of interactive fiction, but
> pretty much have my act together in computational linguistics--my
> research is in discourse processing.
>
> So, I could use some intro material on the subject--particularly how
> the programs are coded. Also, some names of good examples of the
> genre would be useful. Machines available to run stuff are Mac's and
> Sun-4's.
>
> References to books, software, e-mail, etc are all appreciated.


Well, here are the two references that had the most influence on my work:

Brenda Laurel: "Towards the Design of a Computer-based Interative Fantasy
System" (1986). Defines a vision and identifies the hardward and software
technologies required to implement it. Brenda's bibliography will point
you towards many other works of note.

James Meehan, "The Metanovel: Writing Stories by Computer" (1976).
An AI classic, with implementation details (in LISP). This work is pretty
old now, but it was a landmark and laid the foundation for others to build
upon. The stories generated were pretty bland, but that's not the point.
(I was lucky enough to see the system myself; Jim was my AI prof at UCI).

Both dissertations may be obtained from University Microfilms,
(800) 521-0600, at about $50 each in soft cover. Half price for students.

While I'm standing on my soap box for the "references most likely to influence
the advancement of Interactive Fiction", I might as well restate my beliefs
about Interactive Fiction, as I have done here so many times before.

IF offers great potential, but since its appearance about a decade ago,
its growth has been plagued by two problems: how to develop the computer
technology required to support a work of Interactive Fiction, and how to
develop stories that exploit this new genre.

Interactive Fiction differs from traditional fiction in that the author
give up some of the control of the flow of the story. This is because the
reader (or player) is allowed to participate to some degree in the shaping
of the plot through his role as a character in the action. Since the
player-as-character will be making decisions about what he will do next,
the author must allow for multiple paths through a set of plot potentials.
The most primitive way of doing this is through plot branching:
presenting the reader with a small set of fixed choices, each set
corresponding to a branch in a fixed set of potential plot paths.
Unfortunately, this technique is intrinsicly limited and has historically
resulted in relatively uninteresting games.

A more interesting approach is to create a rich set of plot fragments and
character behaviors which may be assembled by the computer to allow the
creation of new stories each time the program is used. In the finished
product, the individual elements of the story can combine in new and
wonderful ways not anticipated by the author or programmer.

David Daniel

unread,
Jan 9, 1991, 8:17:22 AM1/9/91
to

[]A more interesting approach is to create a rich set of plot fragments and


[]character behaviors which may be assembled by the computer to allow the
[]creation of new stories each time the program is used. In the finished
[]product, the individual elements of the story can combine in new and
[]wonderful ways not anticipated by the author or programmer.

Whoa! I don't see the above as having a snoball's chance in hell of being
workable to any significant degree. Let's see if I can present myself
coherently enough to present a workable alternative.

First off let's say we have a story something like this:

A secret agent has to stop Iraqi terrorists from sabotaging a US oil
refinery (topical, ain't I? :] )

For charactors we have:

Abduhl - The head terrorist

Omar - Abduhl's assistant

Rex Reamer - The CIA agent

Celia - Rex's lover

Mr. X - Head of Mobius Oil

Big Guy - Rex's superior

A cab driver

A janitor

Rex's secretary

A security guard


You get the idea...

Now if we had all the text related to this story/plot sitting within some
yet-to-be-written IF program what could we expect to see?

Let's say the reader/player could choose to be any of the main charactors:
Rex, Abduhl, Omar, Big Guy or Mr. X.

This means that 5 (count 'em) stories would need to exist, one from each
charactor's point of view. The next factor is one of narration - would the
narrator be the same persona for each point of view? If so, you would have
a sorry state of affairs. Here's why:

If you're reading/playing the story from Rex's point of view, the narrator
should have tough, perhaps street-wise persona. Celia's persona would entail
a femanine viewpoint, etc. You
would NOT want the computer placing Celia's persona onto Rex.
One way around this is to have each of the 5 stories in first person, but you
then sacrifce the advantage of the narrator as a meta-charactor/interface.

Now for sub-plots:

Lets say that we want to involve a few of the minor charactors into the main
plot. We must decide in advance how much influence they will/can have.
If the janitor can only realisically have a minor or null effect then we don't
want that factor being assigned to Rex or vice versa.

My point is this: Now matter what you want the computer to do as a mechanism
for IF, you MUST maintain the proper balance of what comprises a good game -
freedoms, barriers and purposes.

Add to the above the need to maintain correct literary aspects and you have
created on hell of a programming nightmare.

I don't see ANY good IF application coming into existence without the merger
of two disciplines: programming and writing.
These would best be filled by different people, especially in an initial
venture.

Summary

Here are the basic factors that must be in place before an IF application
can work to its fullest potential.

The main plot.

The sub-plot(s)

All charactors

Narration

The story tself:

a) How many beginning points?

b) How many ending points?

c) How many resolutions?

Number of viewpoints.

What 'role' does the user play in the IF?
a) Participant

b) Observer

I can't imagine an IF application being written in either the vacuum of
programming or of fiction writer. At the point that a talented team forms that
contains both disciplines we'll start seeing real progress.
--
David Daniel (The man with no disclaimer) tro...@polari.UUCP
"Beware the Truth. If you find a Truth it can demand that you make painful
changes." - Frank Herbert

David Graves

unread,
Jan 14, 1991, 8:08:43 PM1/14/91
to
David Daniel writes:

> Whoa! I don't see the above as having a snoball's chance in hell of being
> workable to any significant degree.

I agree that this is not an easy vision to implement, but I think there is
much more than a snowball's chance of being workable. Please note that I
didn't just write a "gee whiz, wouldn't it be great if ...." posting. Many
talented people are working in this area. You can get a PhD in this if you
really want to! Also, I'm not merely reporting on the progress of others.
I've been doing research and development in next generation Interactive
Fiction for about six years, so I speak from experience. But I promise
not to beat you up in this reply. :-)

In David's example, he suggests:

> Let's say the reader/player could choose to be any of the main charactors:
> Rex, Abduhl, Omar, Big Guy or Mr. X.
> This means that 5 (count 'em) stories would need to exist, one from each
> charactor's point of view.

It seems that you are thinking of a story as a static thing. I infer this
from your statement that five points of view would require the creation of
five stories. The writer shouldn't create any stories! The ideas is to
write plot fragments that are selected and assembled at runtime. I admit
that this requires the development of many story pieces, but that's different
from what you are saying above.

Also, who's to say that you would *want* to allow the reader to enter into
more than one character? Given that it is challenging enough to integrate
plot units on the fly, I draw the line at "single character point of view".

> The next factor is one of narration - would the
> narrator be the same persona for each point of view? If so, you would have
> a sorry state of affairs. Here's why:
>
> If you're reading/playing the story from Rex's point of view, the narrator
> should have tough, perhaps street-wise persona. Celia's persona would entail
> a femanine viewpoint, etc. You
> would NOT want the computer placing Celia's persona onto Rex.
> One way around this is to have each of the 5 stories in first person, but you
> then sacrifce the advantage of the narrator as a meta-charactor/interface.

I don't think that the narrator should have any persona at all. The characters
in the story have personality; the narrator is simply part of the interface.
When Celia talks, you hear Celia, not the narrator. When Celia performs some
action, the words appear as narration, but it is Celia's personality that
comes out, not the narrator's. For example, if Celia is mad at you, this
can "leak out" while she is performing some simple act such as giving you
a beer. The narrative text could be "Celia takes a beer from the fridge and
slams it onto the table". This text would be computer generated. You
wouldn't want to have the poor writer developing text for every puny event
that happens!

It's clear that you do want to have a computerized model for personality,
so that each actor does perform actions that are "in character" for him.
You point out that you would not want the computer placing Celia's persona
onto Rex. Yes, indeed! Fortunately, this is not difficult to do.

Maybe you and I are not in major disagreement on this one. If you are saying
that narrative text must reveal the character of each actor, then I agree.
But I strongly disagree that the narrator should be a meta-character. The
reader should not be aware of the presence of the narrator.

> Add to the above the need to maintain correct literary aspects and you have
> created on hell of a programming nightmare.

Challenging, yes. Nightmare? The last several years of work on this haven't
been so bad for me. I think it's alot of fun.



> I don't see ANY good IF application coming into existence without the merger
> of two disciplines: programming and writing.
> These would best be filled by different people, especially in an initial
> venture.

> ....


> I can't imagine an IF application being written in either the vacuum of
> programming or of fiction writer.

I agree. That's why I'm working with "real" writers on each of my IF projects.

David Graves
IF Platform Architect

Michael A. Patton

unread,
Jan 28, 1991, 9:55:04 PM1/28/91
to
In article <1991Jan19....@eng.umd.edu> buz...@eng.umd.edu (Sean T. Barrett) writes:

... it could have pretty solid potential ... But as a literary
device, employing a narrator-with-a-personality could create some
quite interesting twists.

I'm currently looking at writing a 'traditional' adventure game that
allows you to get 'distorted' views of the world ...

Some of this is explored in one of the Infocom games. I don't have my
collection here so I don't remember the title, but the idea was that
you were in a control center operating a set of robots that had
different senses and capabilities. You got different descriptions of
various objects depending on which robot you had to observe through.
The various robots were also operating in parallel which was added a
twist to the standard fare.

G. Maddog Knauss

unread,
Jan 29, 1991, 3:24:17 AM1/29/91
to
In article <MAP.91Ja...@gaak.LCS.MIT.Edu> M...@LCS.MIT.Edu (Michael A. Patton) writes:
>Some of this is explored in one of the Infocom games. I don't have my
>collection here so I don't remember the title, but the idea was that
>you were in a control center operating a set of robots that had
>different senses and capabilities. You got different descriptions of
>various objects depending on which robot you had to observe through.
>The various robots were also operating in parallel which was added a
>twist to the standard fare.

"Suspended."
Just helpful as the dickens,
Greg "Maddog" Knauss

David Graves

unread,
Jan 30, 1991, 4:51:12 PM1/30/91
to
Regarding the narrator-with-a-personality, Lucifer Maleficius writes:

> Some of this is explored in one of the Infocom games. I don't have my
> collection here so I don't remember the title, but the idea was that
> you were in a control center operating a set of robots that had
> different senses and capabilities. You got different descriptions of
> various objects depending on which robot you had to observe through.
> The various robots were also operating in parallel which was added a
> twist to the standard fare.

Those MIT guys! They even name their daemons! I wonder who really posted
that note?

Anyway, in InfoCom's "Suspended", you didn't have a narrator with a
personality -- you had characters. The five robots had differing
capabilities and personalities. I see this as distictly different from
having a narrator with personality.

Reply all
Reply to author
Forward
0 new messages