Adventure generators
Morpheus Nosferatu
I'm not talking about an adventure design language like TADS or Alan,
but rather a stand-alone adventure generator that produces complete
adventures, where the user need only give a minimal degree of input,
such as the level of complexity, type of adventure (mystery, treasure
hunt, etc.), size of adventure, and so forth?
I know it would be relatively straight-forward to create a dungeon
generator for some D&D-like monster-bash, since it is done in numerous
games like hack, rogue, moria, etc. (though creating such dungeon as
a text adventure might be more difficult than as an ASCII graphics
display).
But as anyone ever heard of someone trying to come up with a generator
whigh would produce infocom-style text adventures? I can just imagine
what kind of limitations it would have, but I'm curious to know if
anyone has tried this, and if so what degree of success they've had.
Am I trying to create such a generator? No, but it does make for some
interesting contemplation when I've nothing else to do (read, when I'm
sitting through some bloody lecture). The sheer degree of sophistication
alone required to avoid repetitively similar adventures would have to
be the biggest limitation on any such program.
Phil Goetz
>Has anyone ever worked on, or even heard of, an adventure generator?
>
>I'm not talking about an adventure design language like TADS or Alan,
>but rather a stand-alone adventure generator that produces complete
>adventures, where the user need only give a minimal degree of input,
>such as the level of complexity, type of adventure (mystery, treasure
>hunt, etc.), size of adventure, and so forth?
>But as anyone ever heard of someone trying to come up with a generator
>whigh would produce infocom-style text adventures? I can just imagine
>what kind of limitations it would have, but I'm curious to know if
>anyone has tried this, and if so what degree of success they've had.
No. I know some IF guys (David Graves?) are working on programs
that are supposed to be story generators, that can create a new story on
the fly given the user's actions. But A) the stuff they are talking about
is not, in my opinion, as sophisticated as an Infocom adventure, and
B) I doubt they're going to have much success anyway, since the problem
is too hard at present.
The generator you speak of is not written, not being written, and not
anywhere on the horizon. In 50 years, maybe. In 20, definitely not.
The problem of writing interesting stories, which adhere to someone's
definition of a plot (with goal explanations, conflict, resolution,
comlication, climax, etc., all occuring at appropriate intervals)
is very hard, and I don't expect a solution soon.
But the problem of writing clever puzzles involves much greater
creativity, and I have seen NO evidence that ANYBODY has a clue in
these creativity issues; the most you will find in the field are a
few vague theories of creativity.
This problem is what Stuart Shapiro calls "AI-complete": Solving it
would be equivalent to solving all the other problems of AI.
David A Graves
Morpheus Nosferatu asked:
> Has anyone ever worked on, or even heard of, an adventure generator?
Phil Goetz replied:
> The generator you speak of is not written, not being written, and not
> anywhere on the horizon. In 50 years, maybe. In 20, definitely not.
I think Phil forgot to add the phrase "In my opinion". I hate to see anyone
squash enthusiasm for pursuing the goals of automating personality or
automating plot synthesis, stating their opinions as fact.
I agree with Phil that the problems associated with automating character
behavior and plot synthesis are challenging, but in my opinion, they are
certainly not insurmountable. As an "existence proof", I offer the following:
My first adventure game generator was completed in 1981. It was the
initialization module for a multi-user, realtime, text adventure game called
"Quest". The layout of the adventure was different every time; descriptions
varied, computer controlled characters varied, and the goal of the players
(the quest) varied. The generator selected location, character, and object
descriptions from a large library of colorful text, giving it a feeling of
being "authored" rather than "computer generated". The puzzles were not as
complex as those seen in InfoCom games, but the element of multiple
intelligent players added complexity, and therefore increased the challenge
and fun. Up to twenty players would enter the adventure, each of us in our
own office, so that contact between players was through the computer interface
only. It was tremendous fun, and much of my work in IF over the years has
been an attempt to capture that excitment in a single-user IF environment.
Phil summed up with:
> This problem is what Stuart Shapiro calls "AI-complete": Solving it
> would be equivalent to solving all the other problems of AI.
I do not agree. Synthesis of character behavior and plot twists need not
rely on the computer being creative. Rather, we need only create a good
library of plot and character elements, and devise good algorithms for
assembling them. I'm not trying to solve the old problems of computer
intelligence; I'm working on methods for good illusions.
David Graves
Phil Goetz
In article <315...@hpsemc.cup.hp.com> d...@hpsemc.cup.hp.com (David A Graves) writes:
>
>Morpheus Nosferatu asked:
>
>> Has anyone ever worked on, or even heard of, an adventure generator?
>
>Phil Goetz replied:
>
>> The generator you speak of is not written, not being written, and not
>> anywhere on the horizon. In 50 years, maybe. In 20, definitely not.
>
>I think Phil forgot to add the phrase "In my opinion". I hate to see anyone
>squash enthusiasm for pursuing the goals of automating personality or
>automating plot synthesis, stating their opinions as fact.
OK, you're right, "in my opinion".
>My first adventure game generator was completed in 1981. It was the
>initialization module for a multi-user, realtime, text adventure game called
>"Quest". The layout of the adventure was different every time; descriptions
>varied, computer controlled characters varied, and the goal of the players
>(the quest) varied. The generator selected location, character, and object
>descriptions from a large library of colorful text, giving it a feeling of
>being "authored" rather than "computer generated". The puzzles were not as
>complex as those seen in InfoCom games, but the element of multiple
>intelligent players added complexity, and therefore increased the challenge
>and fun. Up to twenty players would enter the adventure, each of us in our
>own office, so that contact between players was through the computer interface
>only. It was tremendous fun, and much of my work in IF over the years has
>been an attempt to capture that excitment in a single-user IF environment.
...
>Synthesis of character behavior and plot twists need not
>rely on the computer being creative. Rather, we need only create a good
>library of plot and character elements, and devise good algorithms for
>assembling them. I'm not trying to solve the old problems of computer
>intelligence; I'm working on methods for good illusions.
The original question said something like "adventures comparable to Infocom's."
Infocom's adventures are creative; each one occupies a different world, with a
different atmosphere, different history, etc. The little jokes dispersed
throughout, the scenario, the scenery, AND most importantly (in this case)
the puzzles, all require creativity FAR FAR FAR beyond anything I can
envisage a library of plot and character elements coming up with.
Simple object puzzles involving compositionality and the use of instruments
may be generated automatically. Many Scott Adams adventures could perhaps
someday soon be computer-generated; they typically involve figuring out how
physically to get object A to where it is needed. But the better class of
puzzles (IMHO) (e.g. the Hitchhiker's Guide puzzle where you have to note that
a sensory modality is missing from the description) are not of that nature.
II. PUZZLES & PULP IF
Strangely, I find I remember the general mood of good adventures before I
can recall the specific puzzles. Probably a hint that we really don't play
for the puzzles. God knows I spent lots of time making all the puzzles
in _Inmate_ logical and challenging, but the game on the whole (at least
my playtesters gave me this impression) strikes people as boring.
Perhaps computer fiction has not yet advanced to the point where we can
dispense with exotic locales, otherwordly events and characters, or
high-action scenarios. That is, we're still pulp comics.
Phil Goetz
go...@cs.buffalo.edu
Magnus Olsson
>Strangely, I find I remember the general mood of good adventures before I
>can recall the specific puzzles. Probably a hint that we really don't play
>for the puzzles.
Sure, that's why it's called interactive *fiction*, as opposed to
"puzzle solving games". There are some games (Fool's Errand and Three
of one (?) for the Mac leap to mind) that are basically a collection of
puzzles, with little or nothing of a story to tie them together. Nice,
but quite different from IF.
>Perhaps computer fiction has not yet advanced to the point where we can
>dispense with exotic locales, otherwordly events and characters, or
>high-action scenarios. That is, we're still pulp comics.
You may have a point there, but I'd prefer to compare to SF / Fantasy
rather than to pulp comics. And there's a term for what many (most?)
IF players seem to be after - Sense of Wonder. (I've got a feeling
there must be a connection between being a hacekr and liking SF&F).
I don't think we should speak about IF being primitive because its
devotees are looking for Sense of Wonder. In that case, you should
judge e.g. Tolkein and Zelazny as primitive pulp literature, and wait
for the SF & F genres to "advance" to a state of total drabness.
<Shudder>.
Magnus Olsson | \e+ /_
Dept. of Theoretical Physics | \ Z / q
University of Lund, Sweden | >----<
Internet: mag...@thep.lu.se | / \===== g
Bitnet: THEPMO@SELDC52 | /e- \q
David Baggett
>Sure, that's why it's called interactive *fiction*, as opposed to
>"puzzle solving games". There are some games (Fool's Errand and Three
>of one (?) for the Mac leap to mind) that are basically a collection of
>puzzles, with little or nothing of a story to tie them together. Nice,
>but quite different from IF.
But when it comes down to it, an IF game better have a good dose of
clever, challenging, and logical puzzles; otherwise most people will
find it boring. In fact I think there's a very delicate balance
between exposition (i.e., of plot, setting, and characters) and
the interactive stuff (puzzles). Amusingly enough, this is captured
well in the name of the genre: interactive fiction.
The best we can do now with the interactive part is clever puzzles.
Perhaps introducing new ways in which these games can be interactive is
part of what Phil sees as "advancing computer fiction to the point
where we can dispense with exotic locales, otherwordly events and
characters, or high-action scenarios."
Personally, I think any major improvement will be a tall order. (In
other words, I'm not holding my breath.)
>And there's a term for what many (most?) IF players seem to be after -
>Sense of Wonder.
This does seem to be true, now that you mention it. But conversely,
wondrous places are easy places to create adventures. To put it
simply, it's hard to find a reason why there'd be a Cheez Door or a
blessed +6 ring of Totally Kick-Butt Firebolts just lying around on the
street somewhere. But once you change to a fantasy setting, such
things are acceptable without explanation! In a fantasy setting,
creativity is almost totally unconstrained.
Dave Baggett
--
d...@ai.mit.edu MIT Artificial Intelligence Laboratory
ADVENTIONS: interactive fiction (text adventures) for the 90's!
d...@ai.mit.edu *** Compu$erve: 76440,2671 *** GEnie: ADVENTIONS
Jamieson Norrish
This does seem to be true, now that you mention it. But conversely,
wondrous places are easy places to create adventures. To put it
simply, it's hard to find a reason why there'd be a Cheez Door or a
blessed +6 ring of Totally Kick-Butt Firebolts just lying around on the
street somewhere. But once you change to a fantasy setting, such
things are acceptable without explanation! In a fantasy setting,
creativity is almost totally unconstrained.
I think that rather than creativity being almost totally
unconstrained, the need to justify what you create is missing, or at
least in many cases. I don't think this is a good thing; I would much
rather find an ordinary object in an ordinary setting, which had a
reason for being there, than finding an amazing object in an amazing
place, with absolutely no justification for it being there.
No matter what sort of setting you are creating, the same amount of
work needs to go in, I think, so that players can actually relate to
the environment, and everything in it. Just because it is a fantasy
setting does not mean that you can just stick in anything anywhere
with no explanation.
IMHO, of course.
Jamie
Mark 'Mark' Sachs
(Jamieson Norrish) says:
>I think that rather than creativity being almost totally
>unconstrained, the need to justify what you create is missing, or at
>least in many cases. I don't think this is a good thing; I would much
>rather find an ordinary object in an ordinary setting, which had a
>reason for being there, than finding an amazing object in an amazing
>place, with absolutely no justification for it being there.
Actually, _my_ preference is for finding an amazing object in an ordinary
setting. This sort of thing is the basis for nearly all of my favorite
adventures, books, movies, et cetera... I might go so far as to state that
(for me) a primary ingredient in the "sense of wonder" is that said wonderful
thing COULD ACTUALLY HAPPEN to the real world.
Thus, in Trinity: you start out in completely ordinary London, 198-, when
what should you happen on in the middle of a nuclear war but a plain white
door leading to a bizarrely fantastic landscape. In Hitchhiker's Guide
your sleep in an ordinary house on ordinary old Earth is rudely disrupted
by the arrival of an alien space fleet intent on destroying the planet. In
Planetfall, the dull life of a Ensign Ninth Class in the Stellar Patrol is
ended as you tumble upon an undiscovered planet full of secrets. And so on.
As I said before -- if a game starts out in the "real world" but jumps into its
totally amazing world from there, it's already stacked up a lot of points with
me.
"But Dr. Radium, was that necessary?" "Rez, _everything_ I do is necessary."
[Your blood pressure just went up.] Mark Sachs IS: mbs...@psuvm.psu.edu
DISCLAIMER: If PSU knew I had opinions they'd probably try to charge me for it.
Phil Goetz
>In article <JAMIE.92N...@kauri.vuw.ac.nz>, ja...@kauri.vuw.ac.nz
>(Jamieson Norrish) says:
>>I think that rather than creativity being almost totally
>>unconstrained, the need to justify what you create is missing, or at
>>least in many cases. I don't think this is a good thing; I would much
>>rather find an ordinary object in an ordinary setting, which had a
>>reason for being there, than finding an amazing object in an amazing
>>place, with absolutely no justification for it being there.
>
>Actually, _my_ preference is for finding an amazing object in an ordinary
>setting. This sort of thing is the basis for nearly all of my favorite
>adventures, books, movies, et cetera... I might go so far as to state that
>(for me) a primary ingredient in the "sense of wonder" is that said wonderful
>thing COULD ACTUALLY HAPPEN to the real world.
>
>Thus, in Trinity: you start out in completely ordinary London, 198-, when
>what should you happen on in the middle of a nuclear war but a plain white
>door leading to a bizarrely fantastic landscape. In Hitchhiker's Guide
>your sleep in an ordinary house on ordinary old Earth is rudely disrupted
>by the arrival of an alien space fleet intent on destroying the planet. In
>Planetfall, the dull life of a Ensign Ninth Class in the Stellar Patrol is
>ended as you tumble upon an undiscovered planet full of secrets. And so on.
I agree with both of you. I think the "door to another world" plot can work
very well (I think of the Narnia Chronicles, which all start this way,
or those Susan Cooper books, which I'm sure must have had some such device,
or the _Wrinkle in Time_ etc. books.) BUT in these books, there are usually
ONE or TWO such anomalies. I think the first poster was objecting to
games where objects are commonly just lying around - for instance, Scott
Adams' _Pyramid of Doom_, where there is no reason for anything in the game,
just a lot of nifty objects thrown together in a spooky setting. (Though I
liked the game.) No strange new world should be so strange that a rational
being would leave a +4 Sword of Sharpness lying on a bench in an empty tavern.