Slow games

11 views
Skip to first unread message

John E

unread,
May 4, 1998, 3:00:00 AM5/4/98
to

I'm having trouble with the compile
and play speeds of TADS and Inform games.

I'm running a Pentium 233mmx with 96 megs of ram.

Is anyone else having this problem?

A 2 meg game is taking me OVER 2 seconds to compile!

(ha ha just kidding of course)

Just wondering if there where any high-end users out there
who are still amazed by the games you can make in text adventures.

John

BROUGHT TO YOU BY THE Pamela.gam TEXT ADVENTURE.
"You draw the pictures!"

Eric O'Dell

unread,
May 4, 1998, 3:00:00 AM5/4/98
to

On Mon, 4 May 1998 13:03:43 -0500, "John E"
<johnDEL...@earthlink.net> wrote:

>I'm having trouble with the compile
>and play speeds of TADS and Inform games.
>I'm running a Pentium 233mmx with 96 megs of ram.
>Is anyone else having this problem?
>A 2 meg game is taking me OVER 2 seconds to compile!
>(ha ha just kidding of course)
>Just wondering if there where any high-end users out there
>who are still amazed by the games you can make in text adventures.

You're not alone. I've got a P-166 with 96 megs of RAM, and even
substantially slower and less memory-laden machines like my P-90
laptop with 16MB of RAM give one pause to consider how much vaster and
complex text adventures could be on modern hardware. Quake bears
precious little resemblance to its ancient ancestor, the Apple II
version of Castle Wolfenstein, so why have text adventures progressed
so little?


--E.


Dan Shiovitz

unread,
May 5, 1998, 3:00:00 AM5/5/98
to

In article <354e2bc1....@enews.newsguy.com>,
Eric O'Dell <er...@gadgetguru.com> wrote:
[..]

>You're not alone. I've got a P-166 with 96 megs of RAM, and even
>substantially slower and less memory-laden machines like my P-90
>laptop with 16MB of RAM give one pause to consider how much vaster and
>complex text adventures could be on modern hardware. Quake bears
>precious little resemblance to its ancient ancestor, the Apple II
>version of Castle Wolfenstein, so why have text adventures progressed
>so little?

Because when games like Legend come out with more processor-intensive
code, people complain they're unplayably slow on their 8086's.

>--E.
--
(Dan Shiovitz) (d...@cs.wisc.edu) (look, I have a new e-mail address)
(http://www.cs.wisc.edu/~dbs) (and a new web page also)
(the content, of course, is the same)

Eric O'Dell

unread,
May 5, 1998, 3:00:00 AM5/5/98
to

On 5 May 1998 01:27:24 GMT, d...@mozzarella.cs.wisc.edu (Dan Shiovitz)
wrote:

>>You're not alone. I've got a P-166 with 96 megs of RAM, and even
>>substantially slower and less memory-laden machines like my P-90
>>laptop with 16MB of RAM give one pause to consider how much vaster and
>>complex text adventures could be on modern hardware. Quake bears
>>precious little resemblance to its ancient ancestor, the Apple II
>>version of Castle Wolfenstein, so why have text adventures progressed
>>so little?

>Because when games like Legend come out with more processor-intensive
>code, people complain they're unplayably slow on their 8086's.

This is ultimately self-defeating. While keeping hardware requirements
to a minimum is certainly desirable, anchoring them to the hardware of
fifteen years ago is a bit excessive.


--E.


Matt Kimball

unread,
May 5, 1998, 3:00:00 AM5/5/98
to

Eric O'Dell <er...@gadgetguru.com> wrote:
: On 5 May 1998 01:27:24 GMT, d...@mozzarella.cs.wisc.edu (Dan Shiovitz)
: wrote:

:>Because when games like Legend come out with more processor-intensive

:>code, people complain they're unplayably slow on their 8086's.

: This is ultimately self-defeating. While keeping hardware requirements
: to a minimum is certainly desirable, anchoring them to the hardware of
: fifteen years ago is a bit excessive.

Agreed. If an author has a good reason for increasing the minimum
system requirements, this is fine with me. As far as I am concerned,
the following are all "good reasons":

* More intelligent NPCs
* More intelligent parsers
* More accurate simulation of world physics
* Faster development time for the author
(which means more IF in the long run).

There are probably other good reasons.

--
Matt Kimball
mkim...@xmission.com

Paul Francis Gilbert

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

er...@gadgetguru.com (Eric O'Dell) writes:

>This is ultimately self-defeating. While keeping hardware requirements
>to a minimum is certainly desirable, anchoring them to the hardware of
>fifteen years ago is a bit excessive.

>--E.

Agreed. A good example of a processor intensive IF game is the Legend
Lives, which was written using WorldClass. WC does a lot of intensive behind
the scenes work which makes a game written with it more realistic, but at
the same time raises the minimum system required to play it.

If we were forced to keep the limits low so all games could be played on an
8086 we wouldn't see a lot of the more recent games ever being released.


--
Paul Gilbert | p...@yallara.cs.rmit.edu.au (The DreamMaster)
Bach App Sci, Bach Eng | The opinions expressed are my own, all my own, and
Year 5, RMIT Melbourne | as such will contain no references to small furry
Australia | creatures from Alpha Centauri.

Dan Shiovitz

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

In article <6iner6$37g$1...@news.xmission.com>,

Matt Kimball <mkim...@xmission.com> wrote:
>Eric O'Dell <er...@gadgetguru.com> wrote:
[..]
>: This is ultimately self-defeating. While keeping hardware requirements

>: to a minimum is certainly desirable, anchoring them to the hardware of
>: fifteen years ago is a bit excessive.
>
>Agreed. If an author has a good reason for increasing the minimum
>system requirements, this is fine with me. As far as I am concerned,
>the following are all "good reasons":
>
>* More intelligent NPCs
>* More intelligent parsers
>* More accurate simulation of world physics
>* Faster development time for the author
> (which means more IF in the long run).
>
>There are probably other good reasons.

So someone have ideas about any of these? Much of these seem to have
slowly stagnated in the past few years, as though we've reached the
state of the art in all of them, which is clearly untrue. Anyway. here
goes, last to first:

- Faster development time: Usually this translates to higher-level
languages. I know I've heard at least one person who'd like to see
more Perl-esque features in IF languages. Hmm. The problem, of
course, is sometimes we also need access to some gritty parser
features, so there has to be a way to get at those also. Obviously a
major good thing would be if there was better support for NPC design
in IF. I don't know how much this would be the job of the language,
but perhaps some nice libraries?

- Better physics simulations: here's where it gets a little tricky. I
*know* how to do better stuff than what we do; it's definitely
possible to, say, make a pencil object that burns when put in fire
and leaves some ashes and a melted metal piece behind, and can be
cut if you have a knife, or broken, or whittled into an eel, or
whatever. The problem comes because this means starting to give up
some authorial control of the world, which may make it hard to
place stories in it. There's a web page that suggests a solution,
http://rhodes.www.media.mit.edu/people/rhodes/Papers/aaai95.html,
but it's pretty much all theory at this point.

- More intelligent parsers: again, the feeling seems to be that we've
kind of reached a pinnacle here. Anytime someone suggests using
adverbs they're roundly shouted down, and maybe they aren't a good
idea, but there's lots of other stuff we could be trying too. For
instance, a major assumption in IF is that there's like this little
sub-version of English that everyone will learn to speak to the
parser in. The grammar's very easy: [ACTOR,] VERB [NOUN [PREP NOUN]]
and so it's easy to parse .. but are we making the players sacrifice
too much?

For one thing, there's a small-but-significant learning curve when
you first get into IF where you get it pounded into you that this is
the only way to address the parser and all commands must be phrased in
this manner. Personally, I have no idea what people try when they're
first getting started with IF because I'm so used to doing it The
One True Way. (And, leading into the next point, one reason it's so
difficult to get good NPC interaction is because our parser
interaction is in this funky sub-version of English, but we expect
to be able to communicate with other humans/(N)PCS in real English
(or real German, or Sumerian, depending on how much of an Inform
hacker you are))

- So, we come to NPCs at last. I've really given away most of the
situation in previous parenthetical notes, but there's other stuff
besides conversing with them. Except in mystery games, NPCs hardly
ever have any goals of their own. In Curses, Aunt Jemima just
*stays* there the entire game, gardening or whatever, never leaving
to use the bathroom or eat a snack or anything. No wonder she looks
unrealistic; if the character stays in the same room for more than
three turns in a row it's either a miracle or they're glued to the
floor :)

This is not to pick on Curses in particular; even the best-developed
NPCs that I can think of, the ones in Christminster, are still only
barely acceptable. Is it possible that this is because we need to
radically rethink how we're doing NPCs? (and, if so, the obvious
question is how to do it)


Issues like these almost always get greeted with either silence or
"ok, go implement it, and we'll look at it when it's done." Result?
We're still more or less where we were in the 80's*. Blah.

* And of course this is an oversimplification. We do some very pretty
things now, but -- and this is the point -- we're still using all the
same tools.

>Matt Kimball
>mkim...@xmission.com

Eric O'Dell

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

On 5 May 1998 16:30:30 GMT, Matt Kimball <mkim...@xmission.com>
wrote:

>>>Because when games like Legend come out with more processor-intensive
>>>code, people complain they're unplayably slow on their 8086's.

>> This is ultimately self-defeating. While keeping hardware requirements


>> to a minimum is certainly desirable, anchoring them to the hardware of
>> fifteen years ago is a bit excessive.

>Agreed. If an author has a good reason for increasing the minimum
>system requirements, this is fine with me. As far as I am concerned,
>the following are all "good reasons":
>
>* More intelligent NPCs
>* More intelligent parsers
>* More accurate simulation of world physics
>* Faster development time for the author
> (which means more IF in the long run).
>
>There are probably other good reasons.

That nails a whole bunch of them, to which I'd add:

* A modern IDE for game development
* Substantially less linear game structures
* More dynamic environments and accompanying text

[R.A.I-F WARNING: VAPORWARE AHEAD]

I've been giving a lot of thought to the potential features of
"next-generation" text adventures for awhile now, and I'm hitting the
early design stages on a proof-of-concept game, for which you are
probably well advised not to hold your breath. Most of what I'm
thinking about will substantially raise the bar on hardware
requirements, chiefly with respect to memory. A few of the features
I'm trying to implement include:

1. Directional descriptions: The player will not be restricted to a
simple LOOK to receive a description of a room. Instead, LOOK will
take one of the canonical directions as a direct object, e.g. LOOK UP
(or EXAMINE CEILING/SKY) will produce a description of what is to be
seen in that particular direction. The response may be as bland as
"You see a featureless gray wall," but you should never get a
stereotyped response like "You see nothing unusual."

2. Implicit Objects: This is a general class of objects that (usually)
play a purely atmospheric effect, often nested within other objects.
They may be examined explicitly, or if their presence is assumed or
suspected, searched for with a command like LOOK FOR. If some text
reads, "Beneath a tattered awning, there is a tangled mass of rope
lying on the ground," then you may follow a chain of inquiry something
like this:

>examine awning

The awning is a generic green-and-white striped affair, torn in places
and spotted with mold stains and bird droppings in others.

>examine tears

The tears are small, narrow, and numerous, obviously the result of
small imperfections in the material deteriorating out in the elements.

>examine droppings

It is difficult to tell anything about the droppings, as the awning is
on the other side of the street.

>go north

Dashing across the busy street, you find yourself on a sidewalk in
front of a store, identified by an amateurishly painted sign as
Schwartz's Delicatessen. A sign declaring the store to be open hangs
at a tilt on the door, which is ajar. The glare of the noontime sun is
mercifully blocked by a green-and-white striped awning.

>examine awning

The awning presents an almost artistic contrast between bold green and
white vertical stripes, bird droppings, mold stains, and the brilliant
light of the sun streaming through innumerable small tears.

>examine droppings

The droppings, as near as you can tell from the underside of the
awning, must have come from small berry-eating birds if the purple
stains bleeding through the underside of the awning are any
indication.

>get droppings

Despite a valiant attempt to jump to the necessary height, you
conclude that your daily requirement of phosphates will have to be
found elsewhere.

>examine birds

There are no birds here.

>look for birds

Down the street to the south, you see several small birds cavorting in
a mulberry bush.

[and so on]

Needless to say, if an absolute programming nightmare is to be
avoided, most of these objects will have to be carefully kept out of
the player's hands.

3. Explicit barriers. The player should never get a bland "You can't
go that way," but should instead always be told why, e.g., "There is a
solid wall in that direction," or "Excuse me, but there is a yawning
abyss to the north." I know this has been done to some extent before,
but it should be thorough.

4. Environmental variables. Weather and atmospheric feature should
progress dynamically, perhaps in a semi-random fashion depending on
the feature. This may be driven in any number of ways, but the text
should reflect the time of day, whether it is raining or not, how
visibility is limited by fog, etc.

5. Multiplicity of response. As much as possible, descriptions of
objects should include a variety of alternatives, perhaps selected at
random or drawn from a circular queue. Perhaps the player notices
things he has not seen before if revisiting a place, perhaps his
perspective changes based on experiences elsewhere, perhaps the
weather has changed or the order of features described in a room
changes. This will, IMHO, go a long way towards dispelling the static
(and boring) quality oft revisted places and things can generate.
Additionally, standard responses such as, "I see no [thing] here,"
should be drawn from a set sufficiently large that no one occasion
will cause them to repeat except in the case of a player who just
wants to find out how big the set is.

6. Atmospheric spicer events. In any remotely dynamic environment,
many little things are going on. Birds are cavorting on mulberry
bushes, an ice cream vendor is hawking his wares, cars are going by,
pipes in the steam tunnels are making odd noises. Some of these events
should be subject to extended observation. Watching the same birds
cavorting in the aforementioned mulberry bush over several turns might
result in extended descriptions of their fights over berries, their
courtship rituals, etc. Some of these events might have some bearing
on the actual plot of the game, but the majority will not.

7. NPC-to-NPC interactions. While it is obvious that extended,
flexible interactions with players are necessarily limited, that does
not mean that NPC's cannot have fairly elaborate interactions with
each other. These might be scripted exchanges that they have with each
other whenever they are in the same location, among other things,
perhaps depending upon the overall game state. These may range from
pure entertainment value to genuinely useful information.

8. NPC/Player interactions. The complexity of these might increase in
any number of ways, many of which have been discussed at length
elsewhere, including running responses to the player's actions, the
ability to answer more extended queries, and directed dialog.

9. Unbounded but finite plot. The basic plot structure branches until
the midgame, at which point, the branches begin to draw back together
towards a single (or at least a very few) possible conclusions. That
is, up until the middle of the game, every major choice leads two (at
least) two resulting states. After that point, every two states lead
to a single resulting state. Used skilfully, this can contribute to
rising tension as the game draws to a close, and it increases the
replayability of the game.

10. Irrelevant side plots. Many smaller but unnecessary sub-adventures
are scattered throughout the game. The player may be drawn into these
for the sake of adventure and/or the quest for the last lousy point.
It also gives players something to do if they are, fr'example, waiting
for the sun to set to accomplish some more important task. This is not
a new idea, but an old idea projected on a much larger scale.

11. Improved real-world physics. The sudden profusion of objects
demands that the simulation of physics be built upon a well-designed
class hierarchy that can handle totally unforeseen combinations. A
hammer should, for example, be able to smash all easily breakable
objects unless there is some very good reason not to. Obviously, there
are limits to how far this can be carried, but it can go a good deal
farther than has previously been the case.

12. Dynamic synthetic events. The player encounters puzzles, NPC's,
buildings, etc., that are constructed on the fly from a relatively
large set of randomly selected components. To use a simple example,
while wandering lost through the trackless wastes, the player finds a
small village, probably with some puzzle or other to solve. If, lost
again in the same waste, he may never run across the same village, but
he will encounter another similar village (same culture, obviously),
or perhaps a lonely house with a strange old man, or maybe a cave.
IMHO, this would require careful planning of the basic set of elements
from which these events were constructed, but this is not without
precedent in other fields of computer science.

There's more, but I don't think I can sum it up point by point...

Right after college, one of my favorite games was Sid Meier's
Pirates!, which is a sort of archetype of the kind of adventure game
I'd like to write. There is a definite long-term objective ---
liberate your family members from enslavement on Spanish plantations
--- and the day-to-day activities of being a pirate (plundering,
combat, trade, and fulfilling missions for colonial governors) which
lead, through a scoring system, to a well-placed marriage, wealth, and
high social standing or, as was often the case for me, abject poverty
as a castaway with broken health. (Pirates! is also worth study for
the way it creates an impression of complexity in what is really a
very simple program.)

I envision a text adventure with a similar structure. The player has
some long term goal, which is implemented as a more or less tradition
IF quest, though with a more complex and dynamic plot structure and a
graduated scale of total success, partial success, and total failure.
In addition, he pursues the day-to-day business of whatever role he
plays in the world --- merchant, detective, explorer, whatever. His
success or failure in this arena is determined by his standing when he
completes the core game. Conceivably, the player could, if he so
desired, spend a very long time playing this secondary game, wandering
around in a very rich and detailed world, without pursuing the primary
plot. (Alternatively, the primary plot could force itself on him at
intervals not of his choosing.)

At the moment, I have no firm idea on what would replace the combat
that is the mainstay of Pirates. At least, every IF implementation I
have seen so far of combat is pretty lame. (If you know of one that
isn't, let me know.) Whatever that something is, I presume that it
could be implemented through the dynamic synthetic events in 12,
above.

[VAPORWARE MODE OFF]

Constructive criticism is welcome. So is unconstructive abuse: if
there's anything you'd really like to see, the best way to get it out
of me is to tell me that it's not possible, and that even if it was, I
don't have the ability to produce it. <g>


--Eric


Neil K.

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

er...@gadgetguru.com wrote:

[ much stuff deleted ]

> Constructive criticism is welcome. So is unconstructive abuse: if
> there's anything you'd really like to see, the best way to get it out
> of me is to tell me that it's not possible, and that even if it was, I
> don't have the ability to produce it. <g>

Well. A lot of those ideas are good in and of themselves. My WIP (I
noticed my acronym for work in progress appears to have migrated from the
MUD to raif, so... :) contains quite a lot of them. But the thing is - it
takes fnarking forever to write something huge that incorporates this
level of detail. So I think it's a great idea in principle, but very few
people are crazy enough to dedicate the amount of time required to create
something with this kind of resolution.

- Neil K.

--
t e l a computer consulting + design * Vancouver, BC, Canada
web: http://www.tela.bc.ca/tela/ * email: tela @ tela.bc.ca

GLYPH

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

Dan Shiovitz wrote:
>
> - Better physics simulations: here's where it gets a little tricky. I
> *know* how to do better stuff than what we do; it's definitely
> possible to, say, make a pencil object that burns when put in fire
> and leaves some ashes and a melted metal piece behind

Hehe. The Inform game I'm working on right now lets you do just that.
You can light some papers on fire, put them on your desk and go outside,
only to come back later and find that your entire house is on fire.
It's pretty neat and fairly realistic, keeping track of heat and stuff,
but it can screw the player over if he burns something vital.

> - More intelligent parsers: again, the feeling seems to be that we've
> kind of reached a pinnacle here. Anytime someone suggests using
> adverbs they're roundly shouted down, and maybe they aren't a good
> idea, but there's lots of other stuff we could be trying too.

I agree.

> In Curses, Aunt Jemima just
> *stays* there the entire game, gardening or whatever, never leaving
> to use the bathroom or eat a snack or anything.

I remember two games, Ultima V and Times of Lore, that had NPCs who
walked around and did stuff on their own. That was pretty cool, but
they weren't very intelligent otherwise. I think we're stepping into
comp.ai territory here, which isn't necessarily a bad thing. Some sort
of AI include file for Inform would be pretty neat, with an NPC class or
something. Maybe we should all throw in some ideas for this class, like
how do they communicate, make decisions, etc.

- GLYPH

Matt Kimball

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

Eric O'Dell <er...@gadgetguru.com> wrote:
: Right after college, one of my favorite games was Sid Meier's

: Pirates!, which is a sort of archetype of the kind of adventure game
: I'd like to write. There is a definite long-term objective ---
: liberate your family members from enslavement on Spanish plantations
: --- and the day-to-day activities of being a pirate (plundering,
: combat, trade, and fulfilling missions for colonial governors) which
: lead, through a scoring system, to a well-placed marriage, wealth, and
: high social standing or, as was often the case for me, abject poverty
: as a castaway with broken health. (Pirates! is also worth study for
: the way it creates an impression of complexity in what is really a
: very simple program.)

Yes, I have very fond memories of Pirates! too. (Although I was in
grade school at the time. :) It is just as good as the best of Sid
Meier's other games. And the man has had more than one moment of
genius in his career.

But, as you mention below, combat was a large filler in Pirates! I
find this to be the case with many of the classics. I recently went
back and played some of the early Ultima games on an Apple II
emulator, and I found that combat was a large part of the game, but it
wasn't the part that I remembered with nostalgia. I haven't played
Pirates! recently, but the parts that I remember aren't combat, but
the sense of adventure and unlimited possibility in a large world. If
you can capture this same feeling in a work of IF, well, my hat is off
to you. (And just because it is IF doesn't mean you can't use
graphics. I am having trouble imagining a text-only version of Ultima
or Pirates!).

I won't comment individually on the goals you listed in your message,
but they all sound intriguing. A new IF work that implemented only
one or two of those things would be enough to get the attention of the
IF community. I hope you succeed in even just some of those goals.

--
Matt Kimball
mkim...@xmission.com

Eric O'Dell

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

On Wed, 06 May 1998 06:12:25 GMT, fake...@anti-spam.address (Neil
K.) wrote:

> Well. A lot of those ideas are good in and of themselves. My WIP (I
>noticed my acronym for work in progress appears to have migrated from the
>MUD to raif, so... :) contains quite a lot of them. But the thing is - it
>takes fnarking forever to write something huge that incorporates this
>level of detail. So I think it's a great idea in principle, but very few
>people are crazy enough to dedicate the amount of time required to create
>something with this kind of resolution.

<wiping drool from my mouth as I giggle at things that aren't there>

I've got crazy covered. Time may be another matter. But it does seem
to be a natural progression for IF in comparison to other game genres,
where the size and complexity of games has increased over time. The
great advantage we have, as IF authors, over other kinds of game
programmers, is that an enormously large and complicated text
adventure is --- at the level of coding --- much simpler than a
mediocre flight simulator or Doom-ish game. And we don't need thirty
artists and a bank of SGI's to produce scenery --- individual authors
routinely describe more than big-budget movies depict.

There are a lot of things we could borrow from ordinary applications
programming that could speed up the ugly, tedious parts of
development. Excluding for the moment more sophisticated languages,
better underlying object models, better class libraries, and improved
parsers, we could probably make better headway with a full-on IDE
designed around TADS, Inform, or their successors. Something like what
VB and Delphi provide --- a graphical interface where you can draw
maps, place objects, etc., and have the IDE provide the skeleton code
to fill in. Clickable lists of objects could bring up an object
browser where each property can be viewed and modified. A sufficiently
sophisticated tool might help with consistency checking, pointing out
cases where a default behavior for a class might be inappropriate, or
where an object is described somewhere but never defined, and so on.

Which is a medium-to-large development project that would require more
nuts with lots of time to kill, I know.


--Eric


Eric O'Dell

unread,
May 6, 1998, 3:00:00 AM5/6/98
to

On 6 May 1998 15:41:54 GMT, Matt Kimball <mkim...@xmission.com>
wrote:

>(And just because it is IF doesn't mean you can't use


>graphics. I am having trouble imagining a text-only version of Ultima
>or Pirates!).

As am I, which is why I posted the query to the group. I guess I
should be ashamed of the lapse of imagination that leaves me unable to
think of a filler other than combat. But I suppose something will come
to mind if I let it simmer long enough.

>I won't comment individually on the goals you listed in your message,
>but they all sound intriguing. A new IF work that implemented only
>one or two of those things would be enough to get the attention of the
>IF community. I hope you succeed in even just some of those goals.

Thank you. I hope I can count on your support as a beta tester. :)


--Eric


Eric O'Dell

unread,
May 7, 1998, 3:00:00 AM5/7/98
to

On 6 May 1998 05:20:09 GMT, d...@mozzarella.cs.wisc.edu (Dan Shiovitz)
wrote:

>- Better physics simulations: here's where it gets a little tricky. I
> *know* how to do better stuff than what we do; it's definitely
> possible to, say, make a pencil object that burns when put in fire

> and leaves some ashes and a melted metal piece behind, and can be
> cut if you have a knife, or broken, or whittled into an eel, or
> whatever. The problem comes because this means starting to give up
> some authorial control of the world, which may make it hard to
> place stories in it. There's a web page that suggests a solution,
> http://rhodes.www.media.mit.edu/people/rhodes/Papers/aaai95.html,
> but it's pretty much all theory at this point.

Where more accurate physics are concerned, there are some important
issues to consider besides the effect on story. Take, for example, the
many variations on the theme of container objects (or, more
accurately, dependency relationships between objects). We use
containers as a convenient shorthand for more complex relationships,
e.g. a book hidden behind a dresser, which we access with a command
like, "look behind dresser" or "get book from behind dresser". This is
fairly easy to code, and libraries already exist for that purpose.

In the real world, however, the book doesn't have a dependency
relationship on the dresser except in our minds. The book is
(momentarily) hidden or otherwise inaccessible because it, the
dresser, and the wall are arranged in three-dimensional space in such
a way that all routes to the book are blocked by the wall and the
dresser.

We *can* model this and other spatial relationships accurately; many
venerate mathematical subspecialties are devoted to this sort of
thing. The question is, do we want to? At what point do we draw the
line and say, hey, this is a work of fiction, and the player must be
willing to suspend judgment at least a little.

I'd like to draw that line a lot further down the scale than has been
the case previously, but there is obviously a point beyond which the
only people left who will care are not "playing along" anyway.


--Eric

Paul Francis Gilbert

unread,
May 8, 1998, 3:00:00 AM5/8/98
to

GLYPH <graham...@hotmail.com> writes:

>I remember two games, Ultima V and Times of Lore, that had NPCs who
>walked around and did stuff on their own. That was pretty cool, but
>they weren't very intelligent otherwise. I think we're stepping into
>comp.ai territory here, which isn't necessarily a bad thing. Some sort
>of AI include file for Inform would be pretty neat, with an NPC class or
>something. Maybe we should all throw in some ideas for this class, like
>how do they communicate, make decisions, etc.

> - GLYPH

I agree, that would be good. Back when I first found out about the Oz
Project, I was fascinated by one of the discussion papers where they had
a virtual cat (I forget his name), that went around a house and you could
interact with. His reactions to you were based upon a database of your
past reactions to him, his hunger, etc.

Now if we could use that sort of technology, but with an NPC on top it'd
be nifty. Of couse, it'd probably require a great deal of programming
power, especially in a full game where they may be many actors [and the
computer would need to be maintaining actors, even if they're not "on
screen"].

Maybe that could be a future addition to TADS or Hugo [I'll leave Inform out
for now since it's more tricky to add new features to the engine itself]
where you can have a plug-in "intelligence" module, compiled for the
native code of each computer for maximum performance, and for which NPC
classes act as a wrapper around. That way they could override specific
interaction responses with the user whilst still allowing maximum speed
of execution.

Maxxwell Dugay

unread,
May 8, 1998, 3:00:00 AM5/8/98
to

Paul Francis Gilbert wrote:

> I agree, that would be good. Back when I first found out about the Oz
> Project, I was fascinated by one of the discussion papers where they had
> a virtual cat (I forget his name), that went around a house and you could
> interact with. His reactions to you were based upon a database of your
> past reactions to him, his hunger, etc.
>
> Now if we could use that sort of technology, but with an NPC on top it'd
> be nifty. Of couse, it'd probably require a great deal of programming
> power, especially in a full game where they may be many actors [and the
> computer would need to be maintaining actors, even if they're not "on
> screen"].

Okay, so now we just need ideas for the algorithms. I could code the
Inform version.

First question is this: should it have a call to Do_NPC_Stuff() in the
Game_Pre routine which handles all NPC decision-making, or is it better
to have a daemon running for each NPC?

Next question: can anyone think of any NPC-NPC interactions which cannot
be handled with react_before and react_after?

- GLYPH

okbl...@usa.net

unread,
May 8, 1998, 3:00:00 AM5/8/98
to

In article <6itvok$aqp$1...@goanna.cs.rmit.edu.au>,

p...@cs.rmit.edu.au (Paul Francis Gilbert) wrote:
>
> Maybe that could be a future addition to TADS or Hugo [I'll leave Inform out
> for now since it's more tricky to add new features to the engine itself]
> where you can have a plug-in "intelligence" module, compiled for the
> native code of each computer for maximum performance, and for which NPC
> classes act as a wrapper around. That way they could override specific
> interaction responses with the user whilst still allowing maximum speed
> of execution.
>

I think I'm going to disagree with this. I've designed a number of
(admittedly) crude AI engines for this kind of thing, and the problem I have
is how you keep the player focussed on the story. To create a reasonable
facsimile of an actual (small) village, say, you can populate it with a
hundred or so people, each with their own ideas, history, interactions, and so
on. But only a few are going to actually deal with the story's plot.

As a result, you end up with a lot of potential distractions for the player
which will ultimately detract from the story. The main trick in IF is to keep
the player focussed on the story while creating the illusion that he can do
whatever he wants. In that respect, if it is to be IF as we know it,
everything has to lead him back to the story.

In the above situation, where you have a village, or a city, where everyone
has a story, you're looking at something more like an RPG. The interactions
create stories on their own that the author doesn't really have any control
over.

An RPG author creates a world with a set of rules and inhabitants and throws a
switch. The player becomes just another inhabitant.

An IF author creates a world with a set of rules and inhabitants and directs
the player to his destiny. If he's clever, he can give his player all kinds
of apparent freedom, including freedom to shape the outcome of the story, or
pick from a number of pre-defined stories, but once he enters a great deal of
randomity into it, he's given up the artistic control that I think separates
RPGs from IFs.

Of course, that's just my opinion. I could be wrong.

[ok]

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading

Eric O'Dell

unread,
May 9, 1998, 3:00:00 AM5/9/98
to

On Fri, 08 May 1998 19:29:31 GMT, okbl...@usa.net wrote:

> To create a reasonable
>facsimile of an actual (small) village, say, you can populate it with a
>hundred or so people, each with their own ideas, history, interactions, and so
>on. But only a few are going to actually deal with the story's plot.
>
>As a result, you end up with a lot of potential distractions for the player
>which will ultimately detract from the story. The main trick in IF is to keep
>the player focussed on the story while creating the illusion that he can do
>whatever he wants. In that respect, if it is to be IF as we know it,
>everything has to lead him back to the story.

Up to a point, perhaps, but I think this is partly a side-effect of
the hardware limitations of IF's early days. When you had 64k of RAM
and a couple of double-sided 140k 5.25" diskettes restraining the size
of the games, everything tended to have some relevance to the main
object of the game simply because there wasn't room for much
extraneous detail. Anything extraneous was there because either the
author lacked focus, or because it was a deliberate red herring.

If IF games, or at least some IF games, are to become bigger, grander,
and more detailed, both players and authors will have to break out of
this very artificial and conventional mold. Pretend for a moment that
you aren't an experienced IF player, and imagine a vastly detailed IF
game set in the detective genre. You are standing in a study with
seven suspects and about fifty discernable objects, ranging from a
peach pit in an ashtray to an Edwardian oak armoire. You know -- as
you would in real life, that the vast majority of objects are quite
irrelevant. You would not spend the entire game session looking
underneath, behind, and on top of the ottoman, or counting the number
of cigarette butts that have fallen from the overfull ashtray onto the
floor next to the wastebasket by the corner of the desk. If an object
is significant in the game, the author -- just as the author of a
novel -- will in some way direct your attention to the objects that
are important.

The player will only get off track if he is thinking like a
traditional IF player acting according to traditional IF conventions.
It is important to remember that these conventions, familiar as they
are to us, are very much artificial products of a particular kind of
IF, in exactly the same way as we know to EXAMINE CHAIR instead of
something like GIVE THE CHAIR A CAREFUL EXAMINATION.

>In the above situation, where you have a village, or a city, where everyone
>has a story, you're looking at something more like an RPG. The interactions
>create stories on their own that the author doesn't really have any control
>over.
>
>An RPG author creates a world with a set of rules and inhabitants and throws a
>switch. The player becomes just another inhabitant.
>
>An IF author creates a world with a set of rules and inhabitants and directs
>the player to his destiny. If he's clever, he can give his player all kinds
>of apparent freedom, including freedom to shape the outcome of the story, or
>pick from a number of pre-defined stories, but once he enters a great deal of
>randomity into it, he's given up the artistic control that I think separates
>RPGs from IFs.

I'd say that the dividing line between IF and RPGs isn't that
distinct. There are many degrees between the two, and potentially many
hybrids, as autonomy and fate are no more inseparable in fiction than
in life.

--Eric

Adam J. Thornton

unread,
May 9, 1998, 3:00:00 AM5/9/98
to

In article <3553dbfc.54545142@news>, Eric O'Dell <er...@gadgetguru.com> wrote:
>If IF games, or at least some IF games, are to become bigger, grander,
>and more detailed, both players and authors will have to break out of
>this very artificial and conventional mold. Pretend for a moment that
>you aren't an experienced IF player, and imagine a vastly detailed IF
>game set in the detective genre. You are standing in a study with
>seven suspects and about fifty discernable objects, ranging from a
>peach pit in an ashtray to an Edwardian oak armoire. You know -- as
>you would in real life, that the vast majority of objects are quite
>irrelevant. You would not spend the entire game session looking
>underneath, behind, and on top of the ottoman, or counting the number
>of cigarette butts that have fallen from the overfull ashtray onto the
>floor next to the wastebasket by the corner of the desk. If an object
>is significant in the game, the author -- just as the author of a
>novel -- will in some way direct your attention to the objects that
>are important.

The OS/2 graphical adventure _Avarice_ did this rather well.

Its interface also presages _Starship Titanic_'s PET. I wonder if the ST
designers had played _Avarice_.

Adam
--
ad...@princeton.edu
"There's a border to somewhere waiting, and a tank full of time." - J. Steinman

Ivan Cockrum

unread,
May 9, 1998, 3:00:00 AM5/9/98
to

Eric,

In article <354fdc82.177068375@news>, er...@gadgetguru.com wrote:

> I've been giving a lot of thought to the potential features of
> "next-generation" text adventures for awhile now, and I'm
> hitting the early design stages on a proof-of-concept game, for
> which you are probably well advised not to hold your breath.
> Most of what I'm thinking about will substantially raise the bar
> on hardware requirements, chiefly with respect to memory.

I applaud your ideas - like you, I enjoy a great amount of detail in my
IF. However, quite a lot of what you describe can be done right now
without raising any hardware bars. I did some of what you describe in my
comp '97 entry, 'Sunset Over Savannah.' None of it impacted the game's
run time (though compiling took a while), and I did it all by taking
advantage of built-in features in TADS.

> 1. Directional descriptions: The player will not be restricted to a
> simple LOOK to receive a description of a room. Instead, LOOK will
> take one of the canonical directions as a direct object, e.g. LOOK UP
> (or EXAMINE CEILING/SKY) will produce a description of what is to be
> seen in that particular direction. The response may be as bland as
> "You see a featureless gray wall," but you should never get a
> stereotyped response like "You see nothing unusual."

This is neither difficult nor memory intensive. I implemented 'look up'
in Savannah on the advice of one of my beta testers, because I had at
least one important object hung overhead (as well as a lot of background
detail in the sky). Players can also look through open doorways and
examine things on the other side of them, as well as examine some far away
objects.

> 2. Implicit Objects: This is a general class of objects that (usually)
> play a purely atmospheric effect, often nested within other objects.

> Needless to say, if an absolute programming nightmare is to be


> avoided, most of these objects will have to be carefully kept out of
> the player's hands.

Again, neither difficult nor memory intensive. These things just take
careful consideration on the part of the game designer. Personally, I
have an obsessive need to describe my environments down to the tiniest
detail. I created literally hundreds of 'unnecessary' objects in
Savannah. All it took to preventing the player from manipulating them was
including a call to dobjGen().

> 3. Explicit barriers. The player should never get a bland "You can't
> go that way," but should instead always be told why, e.g., "There is a
> solid wall in that direction," or "Excuse me, but there is a yawning
> abyss to the north." I know this has been done to some extent before,
> but it should be thorough.

Ditto. Plenty of games do this now. Lost New York comes to mind. Same
with Savannah - you're always told why you can't go in any given
direction. Again, all it requires is careful attention on the designer's
part.

> 4. Environmental variables.
> 5. Multiplicity of response.
> 6. Atmospheric spicer events.

All good things - they just need that same ol' attention to detail.
You've just gotta wade in and write those multiple room descriptions, code
up those tweeting birds, and describe the thunderstorms.

> 11. Improved real-world physics. A


> hammer should, for example, be able to smash all easily breakable
> objects unless there is some very good reason not to.

I at least tried to follow up with this in Savannah. Anything that can be
filled with water can also be filled with sand. Anything can have a rope
tied around it, unless there's a good reason for it. Anything that can be
broken by having a heavy brick dropped on it, will be. It's not dynamic,
but it mostly works. More attention to detail.

If you really want to do all of these things, you can, without reinventing
the wheel. And please do! I love to see games with the kind of detail
you're talking about!

> 9. Unbounded but finite plot.

> 12. Dynamic synthetic events.

Well, there you've got me. Erasmaganza, anyone? ;-)

-- Ivan

----------------------------------------------------------------------
Ivan Cockrum www.cockrumville.com iv...@NOSPAMcockrumville.com
----------------------------------------------------------------------
To reply by email, remove "NOSPAM" from the address above.

okbl...@usa.net

unread,
May 11, 1998, 3:00:00 AM5/11/98
to

In article <3553dbfc.54545142@news>,

er...@gadgetguru.com wrote:
> Up to a point, perhaps, but I think this is partly a side-effect of
> the hardware limitations of IF's early days. When you had 64k of RAM
> and a couple of double-sided 140k 5.25" diskettes restraining the size
> of the games, everything tended to have some relevance to the main
> object of the game simply because there wasn't room for much
> extraneous detail. Anything extraneous was there because either the
> author lacked focus, or because it was a deliberate red herring.

Agreed. But...

> If IF games, or at least some IF games, are to become bigger, grander,
> and more detailed, both players and authors will have to break out of
> this very artificial and conventional mold. Pretend for a moment that
> you aren't an experienced IF player, and imagine a vastly detailed IF
> game set in the detective genre. You are standing in a study with
> seven suspects and about fifty discernable objects, ranging from a
> peach pit in an ashtray to an Edwardian oak armoire. You know -- as
> you would in real life, that the vast majority of objects are quite
> irrelevant. You would not spend the entire game session looking
> underneath, behind, and on top of the ottoman, or counting the number
> of cigarette butts that have fallen from the overfull ashtray onto the
> floor next to the wastebasket by the corner of the desk. If an object
> is significant in the game, the author -- just as the author of a
> novel -- will in some way direct your attention to the objects that
> are important.

Also agreed. But....

> The player will only get off track if he is thinking like a
> traditional IF player acting according to traditional IF conventions.
> It is important to remember that these conventions, familiar as they
> are to us, are very much artificial products of a particular kind of
> IF, in exactly the same way as we know to EXAMINE CHAIR instead of
> something like GIVE THE CHAIR A CAREFUL EXAMINATION.

Absolutely! But....

If you start doing this with NPCs, you're left with an apparent interactivity
that could be distracting for anyone, and the problem of focussing the player
resurfaces.

You're a hard-drinking, hard-on-his-luck hard-boiled detective. You've got a
*lot* of problems, not the least of which is the dame in the $50 hat who looks
like she just stepped out of Vanity Fair. She's not happy because somebody got
a hold of those pictures you took of her philandering husband and sold them to
the National Tattler. But that's what she gets for stiffing you.

Still the money was only part of what you needed to pay the rent, buy a fifth
of vodka, and pay Mary, your secretary, who's all in a lather because some
niece of hers from Iowa ranaway from home to the city of sin and no one's
heard from you.

And your landlord is having trouble with some bookies who....

Oh, sorry. My point, buried in there somewhere, is that the fiction can have
(and often does have) snippets of stories that may not impact on the final,
overall arc. They may be included for thematic reasons. There are 8 million
stories in this city and 8 million ways to die. There has to be some point at
which your NPCs *cease* to have depth.

But I guess it's still possible; just much harder. Much harder.

> I'd say that the dividing line between IF and RPGs isn't that
> distinct. There are many degrees between the two, and potentially many
> hybrids, as autonomy and fate are no more inseparable in fiction than
> in life.

I think you need the clear distinction to separate the two, but obviously
hybrids exist.

Autonomy and fate *are* mutually exclusive. In fiction there is nothing but
fate. In life, there is nothing but autonomy. (There may, however, be the
illusion of autonomy in fiction, and the illusion of fate in life.)

L. Ross Raszewski

unread,
May 11, 1998, 3:00:00 AM5/11/98
to

In article <6j7735$mp2$1...@nnrp1.dejanews.com>,

okbl...@usa.net wrote:
>
> Autonomy and fate *are* mutually exclusive. In fiction there is nothing but
> fate. In life, there is nothing but autonomy. (There may, however, be the
> illusion of autonomy in fiction, and the illusion of fate in life.)
>

Uh... that's realyl an opinion question. there are those who believe that
there is nothign but fate in real life too.

Branko Collin

unread,
May 11, 1998, 3:00:00 AM5/11/98
to

On Tue, 05 May 1998 15:12:39 GMT, er...@gadgetguru.com (Eric O'Dell)
wrote:

>On 5 May 1998 01:27:24 GMT, d...@mozzarella.cs.wisc.edu (Dan Shiovitz)
>wrote:
>
>>>You're not alone. I've got a P-166 with 96 megs of RAM, and even
>>>substantially slower and less memory-laden machines like my P-90
>>>laptop with 16MB of RAM give one pause to consider how much vaster and
>>>complex text adventures could be on modern hardware. Quake bears
>>>precious little resemblance to its ancient ancestor, the Apple II
>>>version of Castle Wolfenstein, so why have text adventures progressed
>>>so little?
>

>>Because when games like Legend come out with more processor-intensive
>>code, people complain they're unplayably slow on their 8086's.

Well, you shouldn't always listen to what other people say. The
ultimate reason why these large games haven't appeared yet, is because
the people who want those large games haven't written them. Blaming it
on people who complain is a bit unfair.

--
branko collin
col...@xs4all.nl
help me pick my glasses: http://www.xs4all.nl/~collin/bril/

Phil Goetz

unread,
May 12, 1998, 3:00:00 AM5/12/98
to

In article <6j7ojg$hhl$1...@nnrp1.dejanews.com>,

L. Ross Raszewski <rras...@hotmail.com> wrote:
>In article <6j7735$mp2$1...@nnrp1.dejanews.com>,
> okbl...@usa.net wrote:
>>
>> Autonomy and fate *are* mutually exclusive. In fiction there is nothing but
>> fate. In life, there is nothing but autonomy. (There may, however, be the
>> illusion of autonomy in fiction, and the illusion of fate in life.)
>
>Uh... that's realyl an opinion question. there are those who believe that
>there is nothign but fate in real life too.

I think generally there are those who believe we have autonomy in real life,
and produce fictions which are written as if the characters had autonomy, vs.
those who believe Fate rules real life, and produce fictions whose characters
are seen (within the fictional world) to have no free will.

I hesitate to recommend Ayn Rand in case there are impressionable young
minds out there, but she has some interesting things to say in
_The Romantic Manifesto_ in which she divides literature into "romantic",
by which she means fiction written assuming people have free will, vs.
"naturalistic", by which she means the Fates view.

Phil

Eric O'Dell

unread,
May 12, 1998, 3:00:00 AM5/12/98
to

On Mon, 11 May 1998 15:56:21 GMT, okbl...@usa.net wrote:

>Oh, sorry. My point, buried in there somewhere, is that the fiction can have
>(and often does have) snippets of stories that may not impact on the final,
>overall arc. They may be included for thematic reasons. There are 8 million
>stories in this city and 8 million ways to die. There has to be some point at
>which your NPCs *cease* to have depth.

I agree. But I don't think we have to restrain ourselves when it comes
to NPCs, insofar as we lack anything remotely approaching artificial
intelligence. To be sure, NPCs could be a lot more elaborate than they
are today, but the ultimate limiting factor there is that we cannot
make them think. This almost ensures that (in those cases where there
is a clearly defined plot) the NPCs will have to play a focusing and
defining role with respect to the plot, since we lack the technology
to do otherwise.

>I think you need the clear distinction to separate the two, but obviously
>hybrids exist.
>

>Autonomy and fate *are* mutually exclusive. In fiction there is nothing but
>fate. In life, there is nothing but autonomy. (There may, however, be the
>illusion of autonomy in fiction, and the illusion of fate in life.)

IMHO, autonomy and fate are not mutually exclusive. You may, for
example, be fated to death in a lonely alleyway on June 16, 1998, but
you may have a great deal of autonomy in choosing between multiple
chains of events that lead there. The troops defending Berlin against
the Russian advance in 1945 could conceivable have surrendered earlier
or later, but their defeat was inevitable. Life, at least in the
retrospective view of history, seems to consist of great stretches of
inexorable movement punctuated by brief moments when (often unknown to
all participants) there is a fleeting opportunity to change course.

The greater the determinism of IF, the more fiction-like it is;
conversely, the more interactive IF is, the less like fiction it is.
Much IF up to this point has been deterministic in the extreme, even
in the more flexible cases where multiple plot lines are provided. I
think we can make IF much more interactive without danger of losing
fiction-like properties like plot, character, and pacing. But that
will require looking at new ways of coding higher-order story
properties rather than just the relatively low-level object
manipulation puzzles that have thus far been an IF mainstay.

--Eric


+-------------------------------------------------------------------+
| "I have come a very long way from myself only to realize that |
| identity is a skill and self-betrayal is a habit. Once lost, the |
| former is very hard to regain; once gained, the latter is very |
| hard to lose." ---I. Corvus, _The Europe of Our Dreams_ |
+-------------------------------------------------------------------+

Greg Ewing

unread,
May 12, 1998, 3:00:00 AM5/12/98
to

okbl...@usa.net wrote:
>
> > something like GIVE THE CHAIR A CAREFUL EXAMINATION.

This sentence did weird things
to my brain...

The room is bare except for a table and a chair.
An irridescent box lies open on the table.

> look in box

The box contains a careful examination.

> give careful examination to chair

--
Greg Ewing, Computer Science Dept, | The address below is
spam-protected.
University of Canterbury, | To decode it, replace "splodge" with "."
Christchurch, New Zealand | and "curly" with "@".
greg curly cosc splodge canterbury splodge ac splodge nz

okbl...@usa.net

unread,
May 12, 1998, 3:00:00 AM5/12/98
to

In article <6j7ojg$hhl$1...@nnrp1.dejanews.com>,
L. Ross Raszewski <rras...@hotmail.com> wrote:
>
> Uh... that's realyl an opinion question. there are those who believe that
> there is nothign but fate in real life too.
>
In which case, there really isn't any point to this posting, is there? You
were after all fated to do this (and everything else you do in your life).

Mechanistic determinism is swell until you decide you want free will.

[ok]

okbl...@usa.net

unread,
May 12, 1998, 3:00:00 AM5/12/98
to

In article <3557caab.168331169@news>,

er...@gadgetguru.com wrote:
>
> I agree. But I don't think we have to restrain ourselves when it comes
> to NPCs, insofar as we lack anything remotely approaching artificial
> intelligence. To be sure, NPCs could be a lot more elaborate than they
> are today, but the ultimate limiting factor there is that we cannot
> make them think. This almost ensures that (in those cases where there
> is a clearly defined plot) the NPCs will have to play a focusing and
> defining role with respect to the plot, since we lack the technology
> to do otherwise.
>

I disagree: Creating an AI engine to power NPCs in a game world isn't that
difficult--in fact, in some ways, it's easier than the kind of scripting that
IF authors have to do in TADS, Inform, et al.

But let's say Granny's main goal in life is getting through the day with a
minimum amount of boredom, and her daily targets include cleaning, making sure
there's food in the house, getting to temple by 4pm and to canasta by 7:30,
with modifications for (say) aborting the cycle if her grandchildren are in
town. This wouldn't be that hard to emulate and would certainly give the
illusion of--if not intelligence--then at least some kind of life beyond the
player.

When the player gets to Granny's house, though, and she's not there, he's
likely to ascribe some significance to that. And I guess this is what it
comes down to: One of the rules in writing a novel (yes, it's broken often
enough) is that you don't leave unanswered questions. As a result, if you
mention a watch on page 1, the reader is going to believe that the watch has
some significance (thematic or plot) to the rest of the story. If it doesn't,
he may feel cheated or misled.

So, the more the NP characters move as if alive, the harder it is to focus the
player's attention. And it's =far= easier to move characters by a set of
"intelligent guidelines" than it is to either: a) leave them in one place for
the player to find, moving them only according to plot points or; b) move them
about randomly and let the player take his chances.

> IMHO, autonomy and fate are not mutually exclusive. You may, for
> example, be fated to death in a lonely alleyway on June 16, 1998, but
> you may have a great deal of autonomy in choosing between multiple
> chains of events that lead there. The troops defending Berlin against
> the Russian advance in 1945 could conceivable have surrendered earlier
> or later, but their defeat was inevitable. Life, at least in the
> retrospective view of history, seems to consist of great stretches of
> inexorable movement punctuated by brief moments when (often unknown to
> all participants) there is a fleeting opportunity to change course.

It does seem like that, but the fact that it's harder to unscramble an egg
than it was to scramble it in the first place--the fact that actions have
consequences is not the same as fate. If I'm "fated" to die in a lonely
alleyway on June 16 (gee, thanks for the month to live) then my free will
apparently doesn't include jumping off a building today! How is autonomy with
conditions placed on it autonomy?

> The greater the determinism of IF, the more fiction-like it is;
> conversely, the more interactive IF is, the less like fiction it is.
> Much IF up to this point has been deterministic in the extreme, even
> in the more flexible cases where multiple plot lines are provided. I
> think we can make IF much more interactive without danger of losing
> fiction-like properties like plot, character, and pacing. But that
> will require looking at new ways of coding higher-order story
> properties rather than just the relatively low-level object
> manipulation puzzles that have thus far been an IF mainstay.

Probably the most important point of this thread; thanks for making it.

Eric O'Dell

unread,
May 13, 1998, 3:00:00 AM5/13/98
to

On Tue, 12 May 1998 20:56:51 GMT, okbl...@usa.net wrote:

>I disagree: Creating an AI engine to power NPCs in a game world isn't that
>difficult--in fact, in some ways, it's easier than the kind of scripting that
>IF authors have to do in TADS, Inform, et al.

By artificial intelligence I meant actual intelligence, not just
cleverly designed rule-based heuristics. Where complex systems of
rules are concerned, there is plenty of code ready to be put to use
--- something like NRC's Fuzzy CLIPS might be very interesting in an
IF environment.

>When the player gets to Granny's house, though, and she's not there, he's
>likely to ascribe some significance to that. And I guess this is what it
>comes down to: One of the rules in writing a novel (yes, it's broken often
>enough) is that you don't leave unanswered questions. As a result, if you
>mention a watch on page 1, the reader is going to believe that the watch has
>some significance (thematic or plot) to the rest of the story. If it doesn't,
>he may feel cheated or misled.

But this is because the player is accustomed to the conventions of
non-interactive fiction. IF is in many ways more like a movie than a
novel, and in a movie one often sees thousands of extraneous objects
and actors. If the author is skilful in letting the player know what
the conventions of the medium are, this need not be an insuperable
barrier.

I was thinking about the things you've said about keeping the player
focused while I was out driving tonight, and it occurred to me that
with at least one type of plot, the player becomes somewhat
self-focusing: the save the world (or the boat or whatever) from
destruction plot. If the player has a pressing time-constrained need
to save the world (and thereby himself), he will be much more
receptive to plot cues from the author. The time constraint need not
be extremely tight --- the author can simply make the player feel like
he has a very limited amount of time to defuse the cobalt bomb, alert
the world to the terrorist plot, find the orbital elements of the
asteroid, whatever.

Time constraints might be applied to many other kinds of stories with
similar effect. There must certainly be other devices that I'm not
clever enough to think of besides time, but there's at least one good
device.

okbl...@usa.net

unread,
May 14, 1998, 3:00:00 AM5/14/98
to

In article <3559294b.258104161@news>,

er...@gadgetguru.com wrote:
>
> By artificial intelligence I meant actual intelligence, not just
> cleverly designed rule-based heuristics. Where complex systems of
> rules are concerned, there is plenty of code ready to be put to use
> --- something like NRC's Fuzzy CLIPS might be very interesting in an
> IF environment.
>
Well, I don't think there's any way around "cleverly designed rule-based
heuristics" unless you believe that intelligence is somehow a function of
structure and you can duplicate the necessary structure.

>
> But this is because the player is accustomed to the conventions of
> non-interactive fiction. IF is in many ways more like a movie than a
> novel, and in a movie one often sees thousands of extraneous objects
> and actors. If the author is skilful in letting the player know what
> the conventions of the medium are, this need not be an insuperable
> barrier.
>

Agreed, but it's a different skill (for the author).

>
> Time constraints might be applied to many other kinds of stories with
> similar effect. There must certainly be other devices that I'm not
> clever enough to think of besides time, but there's at least one good
> device.
>

No, actually, there aren't any other devices. :-) Time provides the necessary
element to create suspense. Fiction without suspense gives one no reason to
continue on. (Mystery can do this to a degree but there is often a sense of
urgency even in a mystery.) In fact, suspense is probably the most lacking
element in IF--maybe the very reason that IF isn't that popular.

I've seen time-based puzzles appear on a some people's "things I hate the
most" list, too. But along with Granny not being home, I think it falls in
the category of "things that can be handled with the acquisition of new
techniques".

Greg Ewing

unread,
May 15, 1998, 3:00:00 AM5/15/98
to

> In article <3559294b.258104161@news>,
> er...@gadgetguru.com wrote:
> >
> > By artificial intelligence I meant actual intelligence, not just
> > cleverly designed rule-based heuristics.

I'm far from convinced that real humans are other
than an extremely large and rather haphazard collection
of rule-based heuristics...

Eric O'Dell

unread,
May 18, 1998, 3:00:00 AM5/18/98
to

On Thu, 14 May 1998 18:16:32 GMT, okbl...@usa.net wrote:

>> By artificial intelligence I meant actual intelligence, not just

>> cleverly designed rule-based heuristics. Where complex systems of
>> rules are concerned, there is plenty of code ready to be put to use
>> --- something like NRC's Fuzzy CLIPS might be very interesting in an
>> IF environment.
>>
>Well, I don't think there's any way around "cleverly designed rule-based
>heuristics" unless you believe that intelligence is somehow a function of
>structure and you can duplicate the necessary structure.

I didn't mean to espouse any particular theory of mind or
intelligence. I just meant to note that the failure of AI research
(thus far) to produce anything remotely resembling a human mind
necessarily limits what we can expect from NPCs.

>> Time constraints might be applied to many other kinds of stories with
>> similar effect. There must certainly be other devices that I'm not
>> clever enough to think of besides time, but there's at least one good
>> device.
>>
>No, actually, there aren't any other devices. :-) Time provides the necessary
>element to create suspense. Fiction without suspense gives one no reason to
>continue on. (Mystery can do this to a degree but there is often a sense of
>urgency even in a mystery.) In fact, suspense is probably the most lacking
>element in IF--maybe the very reason that IF isn't that popular.

I believe you're probably quite correct on this, with the caveat that
the threat (or opportunity) provided by the passage of time also
requires some sort of emotional identification on the part of the
reader/player. The save the world from annihilation plot is as common
as it is because the player can be relied upon to have some sense of
self-preservation. If the development of NPCs as literary characters
was more advanced, it would be possible to appeal to nobler motives
than pure self-interest. Some recent games have, thankfully, begun to
head in this direction.

>I've seen time-based puzzles appear on a some people's "things I hate the
>most" list, too. But along with Granny not being home, I think it falls in
>the category of "things that can be handled with the acquisition of new
>techniques".

Speaking for no one but myself, the flaw with most time-based puzzles
is that the time limits are too short to solve the puzzles without
violating the player's right to solve puzzles without the need for
information from past lives. Individual timed puzzles are, IMHO, to be
used sparingly if at all. But the story as a whole ought to have some
generous but firm limitation. To refer back to Pirates!, one had a
certain number of years before advancing age and failing health forced
one to cash in one's chips. The player need not be in a panicked rush,
but he ought not to be able to win the game by waiting for the villain
to retire.

Reply all
Reply to author
Forward
0 new messages