NYTimes editorial: On Halo 3 and IF
steve....@gmail.com
Daniel Radosh, "The Play's the Thing."
This link should work for the next few days, until the article goes
into the NYTimes archive (at which time you'll need to have a paying
subscription to view it). To follow this link you'll probably need to
sign up for a free account if you have not already done so.
http://www.nytimes.com/2007/09/28/opinion/28radosh.html?th=&emc=th&pagewanted=print
Radosh's main argument is that Spiffy Graphics only gets us so far,
and modern games tend to lack artistic merit because they don't even
try. He recommends rediscovering the possibility of game-art by
looking back at IF.
I share the view that IF *could* be and indeed *should* be the
forerunner in game design as an art form. It should be *the* place
where discoveries are made and inventions tested. It *should* be the
undisputed champion of avant garde game design. Whether in fact it
*is* -- well unfortunately that's not so clear. Indeed it looks like
IF dropped the ball. It seems that projects like Facade etc. are the
champs of gaming art, and IF isn't really doing anything of great
interest. IF is looking more and more like the kid genius who was too
lazy to do anything with his life.
Why, for instance, did the Oz project fizzle, and why wasn't its
spirit and interest collectively taken up by the IF community? Are we
doing anything interesting instead?
Deathworks
Okay, I haven't read the editorial since I don't care that much about
signing up for yet another free account somewhere (^_^;; . But I do
wish to respond to what you are saying about the IF community.
As far as I can tell, there is still quite some creative productivity.
Take Inform 7 and Emily Short as living examples. Even if I loathe the
manual, there is no denying that Inform 7 takes quite some steps to
encourage alternative designs by adding a lot of useful features into
its core library. And while I think that some of the examples in the
manual don't work very well as examples because of their lugagge,
there is no denying - at least from the admittedly small sampling I
have seen - that Emily Short's works are not of low quality.
Of course, you also have to be realistic about your demands. A good
piece of IF needs quite some time to get developed. And you can't
expect every single piece of IF to re-invent the genre in a major
fashion. New styles and approaches are meaningless if they are to be
used once only and then disregarded in favor of new trends, so the
intense exploration of the existing variants, adding small personal
idoisyncracies here and there, is what should be hoped for.
I don't know, but you really sound as if the IF community had fallen
apart. As if all that was produced was merely 6-command-3-room games
that would be better suited for simple action games.
Sorry, but I feel that that negative attitude is really uncalled for.
Deathworks
steve....@gmail.com
> Emily Short's works are not of low quality.
If you mean she's a competent writer, sure. If you mean her work
contributes useful ideas about how games can express art,
unfortunately -- no, not really so much.
And owing to her persistent interest in AI, character presentation,
narrative steering, etc., she's the *best* example of the kind of
thing we're lacking. (This certainly isn't a complaint against Emily!)
> As if all that was produced was merely 6-command-3-room games[.]
That's not really the complaint, though, is it? The complaint is that
mainstream *interesting* games are drawing more energy from A-Life
projects of the 1990's than they are from any form of interactive
narrative work, and that avant garde art-games, while they are
happening in related fields, are not happening in IF.
Deathworks
I still think that you are far too negative. If you want ideas about
how to do things, while they can certainly be expressed via games
implementing them, articles/magazines, and engines/libraries are also
things to consider. As for magazines, I am not deep enough into the
scene to say how the current state of affairs are (some of the
classics seem to be really dead). But I have just come across the AIF
community which seems to have a currently active magazine with article
titles that seem informative and interesting on the artistic side (I
have to admit that I was too busy the past few days to actually read
the magazine). Likewise, with Inform 7, as I said before, we have a
new impulse in the direction of new forms of expression. Of course,
this needs to be put to use, but with the improvement/debugging of
existing features and implementation of new ideas that seem to be
still on the drawing board, chances are good, despite its manual.
Finally, as for the quality of the games themselves, I think you are
asking simply for too much: It seems as if you required that each and
every game released was a great work of art. If that was that case,
that would be awesome, right. But that's simply unrealistic. Most
releases are non-profit resulting in limited ressources, especially
time constraints. Then, most people are neither professional
programmers or novel authors, deprieving them of the experience in the
fields that are touched upon by the medium quite heavily.
Of course, even a non-professional can make a great game, but it is a
rare occurrance for it to be their first release. At first, they do
some standard fare, relying on what is easy to implement and easy to
write. Then they start exploring their limits, trying new approaches.
Eventually, they will have cut a path to something that is really
unique and revolutionary. But in that path, there may be a dozen
released games that are just average.
Therefore, I do believe that you should not sound so demeaning towards
average non-innovative fare. It is the basis for a healthy community
that is able to support and bring forth the extraordinary. Without all
those small adventures where people just have their simple model
worlds with their nice small stories, you would never see any of the
revolutionary work to begin with.
Personally, I think your comments are unnecessarily aggressive towards
newbies or people who have not yet done something revolutionary. In
that sense, I feel that you are ignoring that writing IF is very much
a learning experience, where the authors explore new concepts and new
approaches. And in doing so, you seem to attack those who have yet to
reach a point where they awe they audience by something not done
before. Needless to say that I don't think that such criticism is
quite counter-productive, making authors stop and depriving them and
us of what precious things they would have created in the future.
Deathworks
James Jolley
marun...@yahoo.co.jp says...
> Hi!
>
> I still think that you are far too negative.
Lol that's ritch from someone who has done nothing but whinge about I7?
> Finally, as for the quality of the games themselves, I think you are
> asking simply for too much: It seems as if you required that each and
> every game released was a great work of art. If that was that case,
> that would be awesome, right. But that's simply unrealistic. Most
> releases are non-profit resulting in limited ressources, especially
> time constraints. Then, most people are neither professional
> programmers or novel authors, deprieving them of the experience in the
> fields that are touched upon by the medium quite heavily.
He isn't saying that. He is just saying that IF needs to do something
else to make it more mainstream. We've all been saying that.
>
> Of course, even a non-professional can make a great game, but it is a
> rare occurrance for it to be their first release. At first, they do
> some standard fare, relying on what is easy to implement and easy to
> write. Then they start exploring their limits, trying new approaches.
> Eventually, they will have cut a path to something that is really
> unique and revolutionary. But in that path, there may be a dozen
> released games that are just average.
True to a point.
>
> Therefore, I do believe that you should not sound so demeaning towards
> average non-innovative fare. It is the basis for a healthy community
> that is able to support and bring forth the extraordinary. Without all
> those small adventures where people just have their simple model
> worlds with their nice small stories, you would never see any of the
> revolutionary work to begin with.
I doubt that. Some people are just natural programmers.
>
> Personally, I think your comments are unnecessarily aggressive towards
> newbies or people who have not yet done something revolutionary. In
> that sense, I feel that you are ignoring that writing IF is very much
> a learning experience, where the authors explore new concepts and new
> approaches. And in doing so, you seem to attack those who have yet to
> reach a point where they awe they audience by something not done
> before. Needless to say that I don't think that such criticism is
> quite counter-productive, making authors stop and depriving them and
> us of what precious things they would have created in the future.
Again, same could be said for you, whining about I7 instead of actually
bothering to learn the system.
Eric Forgeot
> Okay, I haven't read the editorial since I don't care that much about
> signing up for yet another free account somewhere (^_^;; . But I do
> wish to respond to what you are saying about the IF community.
>
you don't need to sign-up in order to read the article, it seems to be in
direct access.
About what was said, I think people can enjoy all sort of games (fps, point
and click adventures or interactive fictions) as entertainment, like
reading a book. Is litterature considered to be art ?
But I must admit most fps games are rather pointless.
David Fisher
news:1190971626.9...@g4g2000hsf.googlegroups.com...
> Today's (Sep. 28, 2007) New York Times features an great editorial by
> Daniel Radosh, "The Play's the Thing."
> http://www.nytimes.com/2007/09/28/opinion/28radosh.html?th=&emc=th&pagewanted=print
(Here is the relevant section if people don't want to look it up):
Like cinema, games will need to embrace the dynamics of
failure, tragedy, comedy and romance. They will need to
stop pandering to the player's desire for mastery in favor
of enhancing the player's emotional and intellectual life.
There is no reason that gorgeous graphics can't play a role
in this task, but the games with the deepest narratives
were the text adventures that were developed for personal
computers in the 1980s. Using only words, these "interactive
fictions" gave players the experience of genuinely living
inside a story. The steps required to advance the plot,
though often devilishly perplexing, felt like natural
behavior rather than arbitrary puzzle-solving. Today's
game designers should study this history as a starting
point for an artistic revolution of the future.
> I share the view that IF *could* be and indeed *should* be the
> forerunner in game design as an art form.
What would you like to see? No hint of sarcasm; I am interested in your
vision of IF and how to get there.
I get the general idea, but I would love some more specifics. What doesn't
exist right now that could exist, that people aren't doing? Apart from
continuing in the direction of Facade, that is.
Maybe a competition could promote some of these things. Or would that not be
appropriate?
Are you working on any of these areas yourself right now? (Again, not
sarcastic or negative - genuinely interested).
David Fisher
Andrew Plotkin
> In article <1190975591.0...@w3g2000hsg.googlegroups.com>,
> marun...@yahoo.co.jp says...
> > Hi!
> >
> > I still think that you are far too negative.
>
> Lol that's ritch from someone who has done nothing but whinge about I7?
Nah, Deathworks is being critical. There's a difference.
--Z
--
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
It's a nice distinction to tell American soldiers (and Iraqis) to die in
Iraq for the sake of democracy (ignoring the question of whether it's
*working*) and then whine that "The Constitution is not a suicide pact."
Jim Aikin
>
> What would you like to see? No hint of sarcasm; I am interested in your
> vision of IF and how to get there.
>
> I get the general idea, but I would love some more specifics. What doesn't
> exist right now that could exist, that people aren't doing?
Tell a great story.
I can elaborate on this point, but really, it's up to you and your
fevered imagination to provide the elaboration.
The story needs to be supported by a good implementation, good
beta-testing, good writing, and so forth. Those are all highly
desirable. But if the story ain't happening, you're just whistling in
the wind.
Now, we can get into a long wrangle in this thread about which existing
IF has "great stories" or "good stories" or whatever. I'm sure the usual
candidates will be trotted out. That debate doesn't interest me. It's
all angels dancing on the head of a pin.
What interests me is telling a great story. A great story involves
lifelike characters we can care about (several of them) behaving in ways
that make would sense in the real world. In other words, a great story
requires plausibility.
Marrying a great story to a puzzle-oriented medium is not easy. But I
claim it's possible.
--Jim Aikin
Jim Aikin
>
> Why, for instance, did the Oz project fizzle, and why wasn't its
> spirit and interest collectively taken up by the IF community? Are we
> doing anything interesting instead?
What was the Oz project?
--JA
Deathworks
Having thought about this thread for a while, I have come to the
following suggestion, which has already been hinted at by some of the
comments:
How about collecting articles/comments on what you hope to see? You
know, have people describe what they would consider an interesting or
great idea/strategy to use in game design. Such articles could be
short, simple statements, or could be major essays with a length
rivalling a major novel, whatever the person sees fit. Anything could
be suggested and described, as long as it is done as affirmative as
possible, that is telling what you WANT to see and not what you DON'T
WANT to see.
I think in doing so you can help people create interesting games by
giving them hints of what they may add to their original designs, or
what they could to improve their games.
Well, that is basically how I see this thread.
Deathworks
Adam Thornton
JFGI.
OK, slightly less dismissively, it was an attempt to do the same sorts
of interaction-modelling-with-big-sets-of-hidden-and-possibly-correlated-
variables that the Erasmatron wanted to do.
It's great if what you want to do is Walcotkeepian simulation of fairly
complex behavior. For those of us who don't buy Aristotle's claim that
plot should emerge dialectically from character and who would rather
create stage sets than model worlds, it's not clear that it's going to
produce interesting entertainments ("stories", "interactive fictions",
"games", whatever).
However, I can see why Steve Breslin would find it the most exciting of
the available directions.
Adam
steve....@gmail.com
> > I share the view that IF *could* be and indeed *should* be the
> > forerunner in game design as an art form.
>
> What would you like to see?
Oh I should think it's pretty obvious.
We should have twenty entries for Innovation Comp -- not one single
solitary entry. (God bless Victor, though!) That's just pathetic
people. The message, loud and clear: innovation is of zero concern.
Wow that's bad. (Ok, shame on me too for not entering, but really it's
just embarrassing.)
Somebody should at least notice if the Art comp gets "plum forgot" two
years running. And when it does run, it would be nice if the criterion
were something more interesting than "duh that sounded purdy." Art
isn't supposed to be pretty; it's supposed to be interesting and
thought provoking. Jarring and discomforting is much more on the right
track than "pleasant" -- for crying out loud.
Actual concerted work by groups of people on major groundbreaking
projects, like believable agents, drama management, variation of
voice, complex simulations of various kinds. For instance a great
demonstration of I7's (alleged) power would be a knowledge-based
directed agent; instead we get an overblown snooty argument now into
its second year that I7 has this neat new power to handle multiple
dispatch in action processing, which is completely trivial in any
language anyway, and not a pressing matter anyhow. (Oooh, my language
can handle this hypothetical and highly unusual case slightly better
than your language.) It's like all the energy to do good work has been
replaced by schoolyard desire to look smartypants.
No cooperation and no interest in the genre -- hmm, maybe they're
related?
The yearly comp is pretty lame, mostly because the entrants don't have
a vision for groundbreaking work. It's all very conservative. The
contest seems to be: how perfectly sophomoric can you write? It's not
*creative*; it's really not *writing*. You know what it is? "Creative
Writing."
ifMud disbanded would be a good move also. But then again maybe their
hyper-conservatism and political baloney is not a cause but rather a
symptom of the hyper-conservatism and political baloney of the larger
community.
> Maybe a competition could promote some of these things.
Sounds fantastic. Good luck getting anybody to enter!
> Are you working on any of these areas yourself right now?
Arguably. But I'm doing non-IF design work at the moment, for reasons
which are perhaps obvious.
steve....@gmail.com
> >What was the Oz project?
> of interaction-modelling-with-big-sets-of-hidden-and-possibly-correlated-
> variables that the Erasmatron wanted to do.
No that's just terribly dismissive, tellingly so. Oz was a set of
around eight widely divergent projects, basically a manifest of most
of the potential research directions for dynamic interactive
storytelling. The website:
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/oz/web/oz.html
Tons of fantastic and widely divergent articles linked from there.
Second of all, though, why the hell should we be slamming Erasmatron?
How the hell is it a smart thing to repeatedly castigate a worthwhile
experiment performed in good faith. Come on, you know there's no such
thing as a failed experiment, in the sense that even failure is highly
instructive. And it's not like the alternative (braindead, safe,
boringly programmed non-innovative go-nowhere conservatism) is any
more interesting or successful. Quite the contrary, it's the true
failure: the non-experiment.
If this community were at all healthy, it would be celebrating
interesting failures, pursuing their possibility in hopes of discovery
and fresh ideas, not lumping all forms of innovation into a single
group of "that which we shall not consider further."
This antagonism towards innovation is the main reason IF is a failed
genre. And Adam Thornton is on the interested and open-minded end of
the spectrum. -- Hey, nothing personal Adam. ;)
Josh Lawrence
> Today's (Sep. 28, 2007) New York Times features an great editorial by
> Daniel Radosh, "The Play's the Thing."
>
> This link should work for the next few days, until the article goes
> into the NYTimes archive (at which time you'll need to have a paying
> subscription to view it). To follow this link you'll probably need to
> sign up for a free account if you have not already done so.
>
>
> Radosh's main argument is that Spiffy Graphics only gets us so far,
> and modern games tend to lack artistic merit because they don't even
> try. He recommends rediscovering the possibility of game-art by
> looking back at IF.
>
> I share the view that IF *could* be and indeed *should* be the
> forerunner in game design as an art form. It should be *the* place
> where discoveries are made and inventions tested. It *should* be the
> undisputed champion of avant garde game design. Whether in fact it
> *is* -- well unfortunately that's not so clear. Indeed it looks like
> IF dropped the ball. It seems that projects like Facade etc. are the
> champs of gaming art, and IF isn't really doing anything of great
> interest. IF is looking more and more like the kid genius who was too
> lazy to do anything with his life.
>
> Why, for instance, did the Oz project fizzle, and why wasn't its
> spirit and interest collectively taken up by the IF community? Are we
> doing anything interesting instead?
I think the Oz project fizzled because after many years of work, they
didn't have a lot to show that was terribly inspiring. Don't get me
wrong, I'm intrigued by the goals they set for themselves, but take a
look at the sample transcripts from the finished projects Lyotard
(simulating a cat):
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/oz/web/lyotard2.html
and Playground (simulating social interactions among kids trading
cards):
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/oz/web/playground.html
Both are mildly interesting, but the agents (cat and kids) are far
from "believable" and the result very far from an experience of
involving art. Are there other examples of text-based output from an
Oz project that show more promise than these, perhaps in one of their
published papers? I'd be interested to look at them if so. But
overall, the promise of Oz finally seems to have bared some fruit in
Facade, which you mention, and which IIRC is based on some of the Oz
work, but Facade has its own problems, too. The concept is great, but
in actual interaction with the game/simulation I found there were so
many misinterpretations of my input by the parser, and so many
illogical reactions from the characters, that the seams of
artificiality showed as badly (if in different ways) as they do with a
standard IF NPC. In fact, the expectations of greater interaction/
intelligent reaction are set higher than the IF standard by what
Facade promises and focuses on, and so the flaws are actually all the
more apparent. Now, I do think Facade is a noble attempt, but it
should be seen as a deeply-flawed first step towards the goal of
believable agents and dynamically-generated interactive drama, not as
an unqualified success, no matter how much attention it's deservedly
gaining for its innovation.
On the subject of artistic IF, I think there is a good deal of work
that falls in the category of either 1) artistically exploring a
serious theme or 2) artistically playing with the genre itself.
Examples of category 1) that I have recently played include Phototopia
(death), Fate (as the title says), Shades of Grey and Tapestry
(complex moral choices), and Lost New York (impact of urban change on
communities). These all are attempts to grapple with serious themes
in a way that is rarely allowed in mainstream commercial games, and
this method of artistic expression is one which has great potential in
IF. For artistic pushing of the boundaries of IF, Emily Short has a
nice catalog of various innovations on these pages:
http://emshort.wordpress.com/reading-if/narrator-and-player-character/
http://emshort.wordpress.com/reading-if/notable-interface/
(and so on:) http://emshort.wordpress.com/reading-if/
I suspect that for you, something must be technically innovative to be
interesting art, and therefore any works of category 1) that are
implemented in a traditional IF/narrative way you dismiss. You are
certainly entitled to use that criteria, but I find it pretty limiting
-- it can also lead to overpraise of a work like Facade which is very
exciting in its innovative aspects but very disappointing in its
aesthetic-dramatic ones.
I agree it seems like the number of innovative new works of IF has
decreased recently, and I agree that's a shame. I think it may be a
side-effect of authors grappling with the newest major systems, T3 and
I7, and once more people have mastered those systems, we'll see more
innovative works. For example, I'm really excited by Nate Cull's
implementation of Planner in I7, and think that both the I7 and Tads
versions of Planner could lead to some very interesting, richer
behaviors and reactions from NPCs in general, and hope that more
authors will dig into Planner.
In terms of perusing a Facade-like goal of deep, believable AI agents
with whom the player interacts to dynamically create a drama, I don't
think it's simple laziness that's involved in that route not being
pursued within the IF world. It's that that route is an incredibly
difficult one, and people who have literally spent years full-time
pursuing it (the Oz people, Chris Crawford with Erasmatron/now
Storytron) have YET to produce a satisfying example of it. In
general, I suspect more fruitful, truly engaging works of art (and, at
the same time, better games) will be made by having more modest goals,
exploring innovations within specific areas in individual games within
the traditional IF mold, enhancing the medium in a progressive, rather
than revolutionary, manner. Though if someone pulls off a grand
revolutionary change that also works as powerful art, more power to
them!
Josh
Adam Thornton
<steve....@gmail.com> wrote:
>Somebody should at least notice if the Art comp gets "plum forgot" two
>years running. And when it does run, it would be nice if the criterion
>were something more interesting than "duh that sounded purdy." Art
>isn't supposed to be pretty; it's supposed to be interesting and
>thought provoking. Jarring and discomforting is much more on the right
>track than "pleasant" -- for crying out loud.
Can you point me to the actual set of IF Art Show reviews that are "duh
that sounded purdy" ?
Adam
Josh Lawrence
>
> Second of all, though, why the hell should we be slamming Erasmatron?
> How the hell is it a smart thing to repeatedly castigate a worthwhile
> experiment performed in good faith. Come on, you know there's no such
> thing as a failed experiment, in the sense that even failure is highly
> instructive.
Well, one reason to slam Erasmatron was the huge hype Crawford built
around it compared with the actually horrible experience it provided
to the user, both in terms of user-interface and as interactive
drama. If you were to regard it solely on the aspect of what
instructive lesson its failed experiment taught you, the lesson would
likely be, "this is a horrible route for interactive art to
take" (even though I believe that's at least partly due to Crawford's
implementation, instead of the concept itself).
I had some hopes that Storytron would be better. But I lost much hope
when I realized how he solved the "if you're dynamically generating
NPC actions and narrative, how are you going to dynamically generate
all the necessary coherent, well-written prose to describe those
actions and narrative" problem -- by REPLACING coherent, well-written
prose with horrible "Deiko" units of grammar and sentence flow-
charts. Take a look at this image:
http://www.storytron.com/overview/images/screen_big.jpg
I originally thought this was part of the interface for programming an
NPC action. But this is how an NPC speaks to you (on the left of the
image) when you're playing Storytron. And supposedly you're to
communicate back using the same language. That is how the game
interacts with you, while supposedly creating an engaging story for
you to take part it. As communication and art, it's wretched.
Stephen Bond has an excellent elaboration on this point here:
http://www.plover.net/~bonds/storytron.html
Josh
Adam Thornton
<steve....@gmail.com> wrote:
>Second of all, though, why the hell should we be slamming Erasmatron?
>How the hell is it a smart thing to repeatedly castigate a worthwhile
>experiment performed in good faith. Come on, you know there's no such
>thing as a failed experiment, in the sense that even failure is highly
>instructive. And it's not like the alternative (braindead, safe,
>boringly programmed non-innovative go-nowhere conservatism) is any
>more interesting or successful. Quite the contrary, it's the true
>failure: the non-experiment.
Not only do I feel that the Erasmatron is not a particularly worthwhile
direction to be going in, I feel that the implementation is all wrong.
Although this sounds odd coming from someone who has repeatedly declared
his preference for faking it over modelling it, Crawford leaps straight
to an implementation without taking the time to develop a model, and it
cripples what he's trying to do.
Take a look at Chapter 16 of _Interactive Storytelling_, particularly the
bit about "Hard-wired versus Soft-wired Approaches." If there's any
*hope* of generating compelling stories with an Erasmatron-like
approach, you *have* to solve the soft-wired problem. Otherwise all
you're going to end up with is repetitive mechanically-generated text
about boring events. Yes, of course it's a huge problem. Yes, it's
probably functionally equivalent to solving AI. Yes, that's half of why
I am convinced this is the wrong direction.
I mean, hell, just *read* Chapter 16 to see what's wrong with the
Erasmatron. Each Actor may only have a single Role in any given
process. DiscoverJulietsBody appears to be an atomic verb rather than a
chunk of semantics assembled from a generic verb and a specific direct
object (if Crawford has a theory of linguistics, I can't find it, and
since he's talking about storytelling, in which, you know, language is
used to represent the world, you'd think he'd care about HOW language
does that. Hint: I think recursion is pretty important there). Look at
the "Territoriality" kludge he hacks up and then justifies in the name
of "dramatic reality" versus "physical reality"--all he's doing is
derailing his simulationism when it gives him a dramatic result he
doesn't want (it seems odd to me that he'd be squeamish about Dick
barging into Jane's shower when six pages earlier he's talked about
necrophilically violating Juliet's corpse, by the way). In which case,
why the simulationist framework in the first place? And by the time
he's finished buckling you into the straitjacket of his development
vision in Chapter 17, you really have to wonder whether the sorts of
stories it's possible to tell with an Erasmatron are worth telling at
all.
>If this community were at all healthy, it would be celebrating
>interesting failures, pursuing their possibility in hopes of discovery
>and fresh ideas, not lumping all forms of innovation into a single
>group of "that which we shall not consider further."
But you see, this is itself lumping all criticism into a single pile.
I'm not the community, and you're perfectly welcome to consider deep
simulation all you want. If you can put together something that tells a
compelling story with it, I'll be impressed. But--as Josh pointed
out--the Oz project (which I believe to have had much more solid
theoretical underpinnings than Crawford)--never got to the point of
creating compelling stories. Nor, as far as I am aware, has the
Erasmatron.
>This antagonism towards innovation is the main reason IF is a failed
>genre. And Adam Thornton is on the interested and open-minded end of
>the spectrum. -- Hey, nothing personal Adam. ;)
I've got nothing against innovation. I think Inform 7 is very
innovative and I enjoy writing in it. I don't think that Oz- or
Erasmatron-like dramatic simulationism is going to be a fruitful avenue
to pursue until there's real AI. And that's not going to be a short
time coming.
Adam
S. John Ross
Quoted to highlight irony amid the calls for "innovation." :)
"Let's think outside this here box, people! Let's get our avant-garde
on! Let's push some boundaries and ... hey! Where are you going! I'm
all alone in the box! Where is everybody?"
Emily Short
> Not only do I feel that the Erasmatron is not a particularly worthwhile
> direction to be going in, I feel that the implementation is all wrong.
> ... I don't think that Oz- or
> Erasmatron-like dramatic simulationism is going to be a fruitful avenue
> to pursue until there's real AI. And that's not going to be a short
> time coming.
I agree: I've looked at the output of a few other projects that have
also tried some form of dramatic simulation or other, and the results
have generally been disappointing as well.
Still: I've sunk many hours into work with the reactive agent planner
and other related approaches to goal-seeking. I found that after a
week or two of work I could just about get my NPCs to wander around
taking beers out of the refrigerator and putting their wet clothes
into the dryer. With more effort, I could bring this behavior into the
realm of dialogue and characterization -- I could make them get mad
and storm off, or avoid each other when they didn't like one another
-- but the output was really boring to read, because the auto-
generation precluded having any kind of interesting content. To do
that would have needed a vast modeling of knowledge and emotion, a
sophisticated prose generator, and probably massive databases of
dialogue templates; and it's still not clear to me that once you'd
done all that, the outcome would have been *good*.
That suggested to me that a) goal-seeking NPCs were a dead end
unlikely to produce anything of narrative interest; that b) they'll be
terrific, but only if someone produces orders of magnitude more
content than I am going to have time to construct in a lifetime; or
that c) I was not defining the problem narrowly enough. C is the only
answer that allows me a way forward, so somewhat arbitrarily, that's
the conclusion I settled on: that *just* putting energy into the
procedural side of things, working on getting my NPCs to do
increasingly complex tasks, was *never* going to arrive at an
interesting story.
Instead, I decided that it's necessary to back up and take a look at
what we might want to do with AI in IF, *specifically*. As in, where
can an interactive narrative genuinely benefit from such components?
How do we design systems so that all the simulation effort is going
into making the NPCs do interesting and compelling things, or managing
the drama to produce interesting and compelling next steps, rather
than into dull Sim-like activity? That in turn leads into questions
about what the player's range of action can and should be in an
interactive story; which is what I've mostly been working on lately.
*If* I have a clear and well-defined vision of what should be
happening in a given scene (say), then I have a reasonable shot at
writing an NPC or drama-management system that will positively
contribute to that experience: perhaps by judging what the player
hasn't yet learned and needs to know before the scene ends; perhaps by
selecting the dialogue that best continues thematic explorations
opened earlier in the work; perhaps by making something happen that
will challenge the player's current position on an issue, whatever
that position was expressed as being. Those are the kinds of AI
problems that present themselves when I frame the problem in terms of
the narrative, rather than pursuing smart NPCs or managed drama as an
end in themselves.
This isn't to say that no one should try a given experiment; but my
own reading is that some of the obvious ways forward *have* been
tried, the experiments were worthwhile and revealing, and the result
was that those ways don't work -- or, at least, that they don't work
unless they're applied to a well-defined problem.
Moreover, they don't produce *art* unless they're applied to a well-
defined problem that's relevant to the larger story-telling task at
hand.
Poster
> David Fisher:
>>> I share the view that IF *could* be and indeed *should* be the
>>> forerunner in game design as an art form.
>> What would you like to see?
>
> Oh I should think it's pretty obvious.
> We should have twenty entries for Innovation Comp -- not one single
> solitary entry. (God bless Victor, though!) That's just pathetic
> people. The message, loud and clear: innovation is of zero concern.
> Wow that's bad. (Ok, shame on me too for not entering, but really it's
> just embarrassing.)
What is the Innovation Comp? I post here fairly regularly, too, so you
can't say that the opinion is from yet another drive-by poster.
> Somebody should at least notice if the Art comp gets "plum forgot" two
> years running. And when it does run, it would be nice if the criterion
Am I required to care?
> were something more interesting than "duh that sounded purdy." Art
> isn't supposed to be pretty; it's supposed to be interesting and
But art can be "pretty", if satiating the senses with a vision of beauty
is what is meant. In other words, pleasing art can be just as much art
as disturbing art. Your refusal to even acknowledge that as a
possibility reveals that you're just looking for somewhere to grind that
axe.
> thought provoking. Jarring and discomforting is much more on the right
> track than "pleasant" -- for crying out loud.
> Actual concerted work by groups of people on major groundbreaking
> projects, like believable agents, drama management, variation of
> voice, complex simulations of various kinds. For instance a great
And how would you define "concerted" or "groundbreaking"? Honestly, it
sounds like you're a dictator that showed up, surprised that there's no
army to command.
> demonstration of I7's (alleged) power would be a knowledge-based
> directed agent; instead we get an overblown snooty argument now into
> its second year that I7 has this neat new power to handle multiple
> dispatch in action processing, which is completely trivial in any
> language anyway, and not a pressing matter anyhow. (Oooh, my language
> can handle this hypothetical and highly unusual case slightly better
> than your language.) It's like all the energy to do good work has been
> replaced by schoolyard desire to look smartypants.
>
> No cooperation and no interest in the genre -- hmm, maybe they're
> related?
No interest in the genre? By whom? Game developers? The culture at
large? If you mean developers, then please shut up now. Do you really
want to know how many of us have toiled, and how many hours, and how
many lines of code we've written?
Like it or not, most IF games are the product of one person with a
vision. As for cooperation goes, there's plenty of cooperation, but it
probably wouldn't meet your standards. Beta-testers, editors, and
software houses (see TextFyre) are examples.
> The yearly comp is pretty lame, mostly because the entrants don't have
> a vision for groundbreaking work. It's all very conservative. The
I'm assuming you're passingly familiar with the Dreamhold.
> contest seems to be: how perfectly sophomoric can you write? It's not
> *creative*; it's really not *writing*. You know what it is? "Creative
> Writing."
I worked as an editor on "The Baron", which was translated into English.
That wasn't sophomoric. I've beta-tested a few other games and the
writing I've seen hasn't been sophomoric, either. And I'm a writer, too.
As far as themes go, and not to blow my own horn overmuch, but my own
work deals with the modern working environment and the toll it takes on
the individual.
> ifMud disbanded would be a good move also. But then again maybe their
> hyper-conservatism and political baloney is not a cause but rather a
> symptom of the hyper-conservatism and political baloney of the larger
> community.
This is the only valid point so far. There is a tendency towards
"hyper-conservativism" in terms of game systems design, and that's
mostly the fault of the big names. I've brought up the inability to
patch Z-code games as other non-IF games (like say Civilization, Doom,
et cetera) are patched. The fact that Glulx still isn't
system-independent is a problem (though to his credit, Plotkin is
working on the sound aspect).
However, all of this is KNOWN. What are you doing about it, besides
filling up the newsgroups with just another rant? I tell you what I'm
doing about it -- I'm trying to integrate Gwindows and ORLib. (That's
not to brag, but only to underscore the point that people are doing
things.)
Would I love to write a game system from the ground up that addressed
some of the shortcomings of the current systems? Sure. However, my
interest in actually writing IF takes precedence. In short, the problems
I've found are largely vexing and irritating, but they're ones I can
live with, for now. I suspect that most authors feel the same way.
When I get done with my WIP, there's a good three or four more ideas
I've got raring to go.
>> Maybe a competition could promote some of these things.
>
> Sounds fantastic. Good luck getting anybody to enter!
>
>> Are you working on any of these areas yourself right now?
>
> Arguably. But I'm doing non-IF design work at the moment, for reasons
> which are perhaps obvious.
>
Fortunately, you've broken away from that other work to throw bombs
here. *mock applause coupled with a heartfelt rolling of eyes*
What the IF community Absolutely Does Not Need is another whiner or more
whining. Do something, rather than talk about doing something -- that's
the key to respect and value.
-- Poster
www.intaligo.com Building, INFORM, Seasons (upcoming!)
Adam Thornton
Emily Short <ems...@mindspring.com> wrote:
>That suggested to me that a) goal-seeking NPCs were a dead end
>unlikely to produce anything of narrative interest; that b) they'll be
>terrific, but only if someone produces orders of magnitude more
>content than I am going to have time to construct in a lifetime; or
>that c) I was not defining the problem narrowly enough. C is the only
>answer that allows me a way forward, so somewhat arbitrarily, that's
>the conclusion I settled on: that *just* putting energy into the
>procedural side of things, working on getting my NPCs to do
>increasingly complex tasks, was *never* going to arrive at an
>interesting story.
This may not be the fault of the NPCs and may have nothing to do with
AI.
Go to a bar sometime. Sit down next to someone. Talk to him or her.
Buy him or her a beer or seven. You will get a story.
And you know what? Odds are, it's going to be a boring story, and not
one you care much about.
Sure, everyone's got a story (and maybe a screenplay). Most of them,
alas, are not of interest to anyone except the author. Without some
reason to care that Mary got mad because Bob put her clothes in the
refrigerator and her beer in the dryer, their story--however engrossing
to them it might be--will be uninteresting to us.
Adam
steve....@gmail.com
> I think the Oz project fizzled because after many years of work, they
> didn't have a lot to show that was terribly inspiring.
I don't know if the program lost its funding. The faculty lead went
into private business, taking one of the major student; nice guys,
haven't heard about them for a couple years. Another major student
went on to co-author Facade. So maybe it wasn't showing impressive-
enough results, or maybe it was over-ambitious. At least it described
the objectives pretty well, and it did a lot of the tough groundwork.
Very gutsy and invigorating, inspiring in will and vision, with the
view that results come after a sufficient amount of work is done; thus
the opposite of what happens on RAIF, at least recently. No guts, no
vision except the negative "what's the point" and "it's too hard." --
Which sounds to me like a desire to rationalize surrender.
> Are there other examples of text-based output from an
> Oz project that show more promise than [Lyotard], perhaps in one of their
> published papers?
No. I think the project was more interested in identifying goals and
possible approaches. These are the kind of problems that take years
and sometimes decades to solve. To clarify the questions, goals, and
means -- these are useful steps. But no, they didn't have a strong
success producing output. (They did, however, produce specifications
for multiple output generators. They proved that the problem is
workable.)
> I'd be interested to look at them if so. But
> overall, the promise of Oz finally seems to have bared some fruit in
> Facade, which you mention, and which IIRC is based on some of the Oz
> work, but Facade has its own problems, too.
Yes. I think "The Party" is coming out next year; that's going to
answer a lot of the questions Facade posed and produced. Before we
start slamming Facade (which is the next obvious move), let's
recognize that this is once more an extremely valuable exploration, an
entirely valid experiment, and for christ sake forget about evaluating
it negatively as a game that didn't quite realize its potential and
promise.
> On the subject of artistic IF, I think there is a good deal of work
> that falls in the category of either 1) artistically exploring a
> serious theme or 2) artistically playing with the genre itself.
> [...]
> I suspect that for you, something must be technically innovative to be
> interesting art, and therefore any works of category 1) that are
> implemented in a traditional IF/narrative way you dismiss.
Well, I don't consider them useful for the development of the field: I
don't think they'll be generalizably useful to gaming's development
towards art. And anyway it's all too easy to address a theme or issue.
Successful art rarely takes such an obvious approach.
> I'm really excited by Nate Cull's
> implementation of Planner in I7, and think that both the I7 and Tads
> versions of Planner could lead to some very interesting, richer
> behaviors and reactions from NPCs in general, and hope that more
> authors will dig into Planner.
I seriously doubt it will happen. I think it's just too difficult;
nobody has any motivation to do it; there's no apparent interest in
doing something new; and the community seems strikingly opposed to
experimentation. I see no reason to hope for a future break in the
trend.
> In terms of perusing a Facade-like goal of deep, believable AI agents
> with whom the player interacts to dynamically create a drama, I don't
> think it's simple laziness that's involved in that route not being
> pursued within the IF world. It's that that route is an incredibly
> difficult one, and people who have literally spent years full-time
> pursuing it (the Oz people, Chris Crawford with Erasmatron/now
> Storytron) have YET to produce a satisfying example of it.
Where the hell did you get 'satisfying' as a criteria for anything?
That's either uselessly vague or anathematic of art, and is in any
case a completely bullshit criterion for good work in an emerging
genre. Try 'interesting' or 'thought provoking' on for size.
> In
> general, I suspect more fruitful, truly engaging works of art (and, at
> the same time, better games) will be made by having more modest goals,
> exploring innovations within specific areas in individual games within
> the traditional IF mold, enhancing the medium in a progressive, rather
> than revolutionary, manner.
That's why nobody is looking to IF for inspiration.
Jim Aikin
>
> Go to a bar sometime. Sit down next to someone. Talk to him or her.
> Buy him or her a beer or seven. You will get a story.
>
> And you know what? Odds are, it's going to be a boring story, and not
> one you care much about.
>
> Sure, everyone's got a story (and maybe a screenplay). Most of them,
> alas, are not of interest to anyone except the author. Without some
> reason to care that Mary got mad because Bob put her clothes in the
> refrigerator and her beer in the dryer, their story--however engrossing
> to them it might be--will be uninteresting to us.
Depends on how you define "interesting." I would argue that the
interestingness of a story is situationally dependent. If you know Mary
and Bob and have an inkling about their relationship friction from
earlier encounters, you will quite likely find Mary's narrative
engaging. Or, in a completely different manner, if Mary is an engrossing
storyteller, you will care in spite of yourself! If you've been shut in
a hospital room for a month and this is your first night out, almost any
story you hear might be exciting. Mary could tell you about her pet rat
getting his tail caught in the treadmill, and you'd howl with laughter.
A true storyteller is someone who can engage us to the point where we
care about people we've never met and have no special reason to care
about -- people who, in point of strict fact, don't exist.
The methods by which this is accomplished are not a secret. You can buy
whole books devoted to telling you how to do it. (I can recommend
several, if anyone is interested.) The difficult part is actually using
the ideas that are in the books.
Truman Capote had a line somewhere in the introduction to one of his
books about "the whip God gave me." I wish I remembered the exact quote.
The point he was making was that the route to great storytelling lies
through _relentless_ self-criticism. Of course, his eventually got the
best of him. He dried up completely. So there's something to be said for
the adage, "Get it right, but get it written."
--JA
steve....@gmail.com
> auto-generation precluded having any kind of interesting content.
That is undeniably an overstatement. You are overstating because
you're afraid to accept partial successes. You say it would be zero
interest, when in fact its only fault is that it would be a work
towards future work. Plus I already described like three years ago a
perfectly workable algorithm for action-report grouping. You never
even tried it and now you're saying what, that the whole of NPC
automation is entirely without hope? Bullshit you don't believe that.
You're just trying to sound off.
> To do
> that would have needed a vast modeling of knowledge and emotion, a
> sophisticated prose generator, and probably massive databases of
> dialogue templates;
Your adjectives ('vast', 'sophisticated' (in the sense of highly
programmatically complicated), and 'massive') are all carefully chosen
to exaggerate the problems. You can choose between facing the issues
and thinking about them further, or describing them as prohibitively,
impossibly complicated and simply give up. You chose to give up, and
now you're rationalizing.
> and it's still not clear to me that once you'd
> done all that, the outcome would have been *good*.
Even if it fell on its face, it would be not *good* but *great*
compared to your standard IF claptrap.
> That suggested to me that a) goal-seeking NPCs were a dead end
> unlikely to produce anything of narrative interest; that b) they'll be
> terrific, but only if someone produces orders of magnitude more
> content than I am going to have time to construct in a lifetime; or
> that c) I was not defining the problem narrowly enough.
If you mean that you weren't considering specific solutions, yes. But
you're letting the problem lay you flat; you gave up way too early.
You can't bear to produce something that isn't "successful" (in the
sense that it's within convention). That's a recipe for innovation
paralysis.
> This isn't to say that no one should try a given experiment; but my
> own reading is that some of the obvious ways forward *have* been
> tried, the experiments were worthwhile and revealing, and the result
> was that those ways don't work -- or, at least, that they don't work
> unless they're applied to a well-defined problem.
Do me a big favor. Revise that paragraph. Take out all the hedging and
bullshitting and lip-service and slipknots make your point. I think
you're saying that you tried AI but you didn't know what you were
doing and you still can't figure it out.
Adam Thornton writes:
> [A]ll you're going to end up with is repetitive mechanically-generated text
> about boring events. Yes, of course it's a huge problem. Yes, it's
> probably functionally equivalent to solving AI. Yes, that's half of why
> I am convinced this is the wrong direction.
> [...]
> I don't think that Oz- or
> Erasmatron-like dramatic simulationism is going to be a fruitful avenue
> to pursue until there's real AI.
Okay, here we go again: "Yes of course it's a huge problem." The
familiar avalanche of "why trying isn't worth it" hyperboles and
rationalizations. Think of three solutions and get back to me, or just
admit you're simply uninterested in pushing the genre.
You can make progress without "solving AI": First, "solving AI" is a
very dumb concept. But even assuming this is a discrete goal, do you
imagine the process of its solution will be immediate? If not, how do
you think the process will occur, if not people doing work which does
not immediately generate "huge solution"-style results? What's with
the complaint that things would be a lot easier in an AI-complete
world? The whole point is that the work is worthwhile because it's not
easy. You don't want to work on the problem until it's already solved.
Jim Aikin
>
> Very gutsy and invigorating, inspiring in will and vision, with the
> view that results come after a sufficient amount of work is done; thus
> the opposite of what happens on RAIF, at least recently. No guts, no
> vision except the negative "what's the point" and "it's too hard." --
> Which sounds to me like a desire to rationalize surrender.
This line jabbed my naughty streak, so I hopped over to Baf's Guide to
find out what games you've written lately, Steve.
Having taken a 7-year break between my first and second releases, I'm
not in a position to cast stones. So I won't. I'll just remark that when
I read the phrase 'rationalize surrender,' I was moved to look it up.
> Yes. I think "The Party" is coming out next year; that's going to
> answer a lot of the questions Facade posed and produced. Before we
> start slamming Facade (which is the next obvious move), let's
> recognize that this is once more an extremely valuable exploration, an
> entirely valid experiment, and for christ sake forget about evaluating
> it negatively as a game that didn't quite realize its potential and
> promise.
You know, I'm a card-carrying liberal, but this line reminds me more
than a bit of the accusation conservatives sometimes level at liberals.
We're accused of saying things like, "This is a completely valid and
important innovation in education. Never mind that the students didn't
learn anything!"
How exactly are we to evaluate the validity of an experiment, other than
by asking whether it realized its potential? The proof of the pudding is
in the eating.
> Where the hell did you get 'satisfying' as a criteria for anything?
> That's either uselessly vague or anathematic of art, and is in any
> case a completely bullshit criterion for good work in an emerging
> genre. Try 'interesting' or 'thought provoking' on for size.
No, let's stick with 'satisfying.' I will certainly do so. In my
opinion, anything _other_ than 'satisfying' is bullshit. The proof of
the pudding, etc.
--Jim Aikin
Jim Aikin
>
> You can make progress without "solving AI": First, "solving AI" is a
> very dumb concept. But even assuming this is a discrete goal, do you
> imagine the process of its solution will be immediate? If not, how do
> you think the process will occur, if not people doing work which does
> not immediately generate "huge solution"-style results? What's with
> the complaint that things would be a lot easier in an AI-complete
> world? The whole point is that the work is worthwhile because it's not
> easy. You don't want to work on the problem until it's already solved.
Uhh.... Okay, hotshot. Let's stipulate that "working on the problem" is
a desirable activity irrespective of whether the problem is definitively
solved. I'm even willing to let you stipulate what "the problem" is,
which is probably inadvisable.
My question is: How exactly are YOU "working on the problem"? If you've
made some sort of significant contribution to the science/art of
generating lifelike NPC behavior automatically, please -- tell us about
it! That would be soooo much more productive than all this kvetching.
--JA
Adam Thornton
<steve....@gmail.com> wrote:
>Okay, here we go again: "Yes of course it's a huge problem." The
>familiar avalanche of "why trying isn't worth it" hyperboles and
>rationalizations. Think of three solutions and get back to me, or just
>admit you're simply uninterested in pushing the genre.
I am simply uninterested in pushing the genre *in that direction*.
Knock yourself out.
> You don't want to work on the problem until it's already solved.
Damn straight. Let those people who think that creating AI is a fun
thing to do, do it. Me, I want to do some sort of interactive
storytelling, and I don't think that reactive-actor simulationism
(except for generating background color, where it can be awfully useful)
is going to be helpful to me until there's actual Artificial
Intelligence.
Adam
Adam Thornton
>No, let's stick with 'satisfying.' I will certainly do so. In my
>opinion, anything _other_ than 'satisfying' is bullshit. The proof of
>the pudding, etc.
The proof of the pudding is truly marvellous, but the margin is too
small to contain it.
Adam
Josh Lawrence
> Josh Lawrence wrote:
>
> > On the subject of artistic IF, I think there is a good deal of work
> > that falls in the category of either 1) artistically exploring a
> > serious theme or 2) artistically playing with the genre itself.
> > [...]
> > I suspect that for you, something must be technically innovative to be
> > interesting art, and therefore any works of category 1) that are
> > implemented in a traditional IF/narrative way you dismiss.
>
> Well, I don't consider them useful for the development of the field: I
> don't think they'll be generalizably useful to gaming's development
> towards art. And anyway it's all too easy to address a theme or issue.
> Successful art rarely takes such an obvious approach.
>
Well, that's a point where we strongly disagree. I think it's easy to
pick a theme to address, I think it's very difficult to address a
theme well. "Shades of Grey" for example, set up some interesting
moral quandaries but then partially nullifies them by clearly
preferring one course of action of over another in the endgame.
"Tapestry" has this same problem -- in the course of the game the
player is forced to make agonizing choices between actions that are
not obviously better than others, but then the endgame seems to
suggest that certain actions were morally superior than others (this
interpretation of the endgame has been debated, however). But even
though these games' endings seem flawed, the experience of agonizing
over the choices as the player was a rich one -- and so these games
did successfully advance the art of the genre by bringing that to the
table. But it's clearly a theme, only one of many, that could be
explored further and with more nuance, and IF is well-suited to this
task.
> > I'm really excited by Nate Cull's
> > implementation of Planner in I7, and think that both the I7 and Tads
> > versions of Planner could lead to some very interesting, richer
> > behaviors and reactions from NPCs in general, and hope that more
> > authors will dig into Planner.
>
> I seriously doubt it will happen. I think it's just too difficult;
> nobody has any motivation to do it; there's no apparent interest in
> doing something new; and the community seems strikingly opposed to
> experimentation. I see no reason to hope for a future break in the
> trend.
>
This is curious - you think it's too difficult, yet this what you're
urging that the community desperately needs to pursue? I do agree
it's difficult and time-consuming, but I think it can be fruitfully
explored if the scope with which Planner is employed in an individual
game is limited. For instance, have it control the actions of a
villain who runs around the same (limited) environment as the player,
dynamically reacting to the player's actions by altering the
environment, but have it so that the player (until perhaps the
endgame) sees these dastardly deeds but cannot directly interact with
or talk to the villain -- that way you leverage Planner to have an
interesting antagonist who dynamically reacts, but you don't have to
bogged down in implementing direct interaction/conversation with this
antagonist. Or use Planner to create interesting actions/interactions
with non-human characters -- say, a number of fantastic animals a
player has to manage within a surreal zoo, or within a bizarre self-
contained alien ecosystem, or programming several disparate robots in
a robot factory. By not attempting to replicate complex human
behavior, the task becomes much more doable, but still allows for
richer and more interesting behavior than we normally see.
Basically, instead of using Planner to try to generate an interesting
experience from a bunch of independently active agents "turned loose",
I'm talking about carving out specific areas of a game to deploy
Planner to enhance the experience in particular away -- similar to the
limited, strategic deployment of NPC AI Emily discusses above.
Of course, I don't have a good track record of completing games, but
I'd think it would be rewarding to try this approach.
> > In terms of perusing a Facade-like goal of deep, believable AI agents
> > with whom the player interacts to dynamically create a drama, I don't
> > think it's simple laziness that's involved in that route not being
> > pursued within the IF world. It's that that route is an incredibly
> > difficult one, and people who have literally spent years full-time
> > pursuing it (the Oz people, Chris Crawford with Erasmatron/now
> > Storytron) have YET to produce a satisfying example of it.
>
> Where the hell did you get 'satisfying' as a criteria for anything?
> That's either uselessly vague or anathematic of art, and is in any
> case a completely bullshit criterion for good work in an emerging
> genre. Try 'interesting' or 'thought provoking' on for size.
>
I mean "satisfying" as in "That was satisfying as a work of art
because it was rich in meaning, and successfully dealt with whatever
aesthetic/thematic/experiential goals it set for itself", not
"satisfying" as in "That was a satisfying hershey bar". This
satisfaction is separate from whatever other specific emotional and
intellectual responses that the work generates. A work can be
throught-provoking, disturbing, even depressing...and still be
satisfying as a work of art. If you like, replace "satisfying" with
"successful". Facade is interesting for its innovative approach, but
does not succeed as a satisfying work of art because it fails to
deliver on its promise of believable characters and a moving,
dynamically-generated drama. The same applies, to a much greater
degree, to the sample Erasmatron gameworld I tried long ago. I will
be interested to try "The Party" when it comes out, to see if they
achieve something more artistically successful with the Facade
approach.
Josh
Jim Aikin
Very good. It took me fifeen full seconds to get it.
--JA
Brian Slesinsky
> You can choose between facing the issues
> and thinking about them further, or describing them as prohibitively,
> impossibly complicated and simply give up. You chose to give up, and
> now you're rationalizing.
Or you can make a rational decision about which projects are worth the
effort. Dumping on people who disagree about the cost-benefit ratio
is very unlikely to change anyone's mind. If you have some actual
ideas that you think are worth pursuing, how about talking them up and
doing some cheap experiments to show that they're promising? Or
maybe raise some money and hold a contest or something?
- Brian
David Fisher
news:yKGdnVpHd5w7HGDb...@giganews.com...
> steve....@gmail.com wrote:
>>
>> We should have twenty entries for Innovation Comp -- not one single
>> solitary entry. (God bless Victor, though!) That's just pathetic
>> people. The message, loud and clear: innovation is of zero concern.
>> Wow that's bad. (Ok, shame on me too for not entering, but really it's
>> just embarrassing.)
>
> What is the Innovation Comp? I post here fairly regularly, too, so you
> can't say that the opinion is from yet another drive-by poster.
It was a strange idea for a competition I had (in March 2007) ...
http://www.ifwiki.org/index.php/Innovation_Comp
David Fisher
David Fisher
news:1191013296.8...@d55g2000hsg.googlegroups.com...
> On Sep 28, 3:50 pm, a...@fsf.net (Adam Thornton) wrote:
I found this post (and Adam's previous post) very insightful ... I don't see
all of the problems as being quite so insurmountable, though.
Steve mentioned cooperation ... there is already tons of cooperation
happening in IF (contributed extensions, etc), but it's a good point to
make. If someone makes some progress in a difficult area of IF but their
method isn't discussed/shared, then it isn't possible for the next person to
build on that - they have to re-create the method from scratch again.
(Probably preaching to the converted here. I'm not aware of any jealously
guarded IF secrets ... or maybe that's because they're secret :-) ).
> How do we design systems so that all the simulation effort is going
> into making the NPCs do interesting and compelling things, or managing
> the drama to produce interesting and compelling next steps, rather
> than into dull Sim-like activity?
...
> Those are the kinds of AI
> problems that present themselves when I frame the problem in terms of
> the narrative, rather than pursuing smart NPCs or managed drama as an
> end in themselves.
I totally agree with this. The aim should be to create a coherent,
interesting narrative, rather than to just have autonomous NPCs running
around. This kind of idea can apply to all kinds of things in the story
besides NPC behaviour, too.
There was an experiment along these lines using Anchorhead done by Mark
Nelson and Michael Mateas:
http://www.lcc.gatech.edu/~mateas/publications/NelsonMateasAIIDE05.pdf
This even includes doing things like changing where an object is found in
order to make the plot move along more smoothly. Potentially you could also
change the geography, the characters ... anything. My only worry with this
kind of thing is that the world won't seem consistent on replay (or after
Undo), and that no two players get the same world. This might seem like a
minor thing, but it reminds me of a review of a game where different endings
had different (and contradictory) explanations for the events that happened,
which left the player feeling very disoriented.
David Fisher
<<Chateau d'IF ~~ http://dfisher.serverspeople.net>>
James Jolley
>
> Damn straight. Let those people who think that creating AI is a fun
> thing to do, do it. Me, I want to do some sort of interactive
> storytelling, and I don't think that reactive-actor simulationism
> (except for generating background color, where it can be awfully useful)
> is going to be helpful to me until there's actual Artificial
> Intelligence.
>
>
>
look at it but I would have thought that the rule bassed I7 coupled with
this would allow you great flexibility.
>
Deathworks
I am quite pleased to see people share some information about articles
or projects that have been done in the past. I really liked the
comment and thoughts Emily Short gave in her response, finding them
very useful and inspiring (although I don't know whether I will be
able to incorporate anything in the game I am currently programming as
finishing a game is big enough a challenge for me, sorry). Anyhow, I
was quite appalled by the flame with which Steve responded to Emily.
This may sound harsh, but no matter how often I re-read that reply, I
could not help but feel that it was simply vicious and unfair.
In addition, I find Steve's position rather difficult to understand.
On the one hand, we have large dismissive statements about how the
community as a whole has failed and how bad things are at the moment.
On the other hand, he says we should not judge some title called
"Facade" on its own merit but rather wait for some planned sequel to
bring out its potential. So, should we judge things by what they are
or what they might potentially be? If you are judging by potential, I
don't see how you can condemn the IF community as a whole. And if you
judge by actual facts, that defense of "Facade" is meaningless.
Sure, I am a drive-by poster here. As much as it annoys myself, I
can't really guarantee that I will be around next month. Maybe another
depressive mood has moved me onward somewhere else, maybe not. Still,
I am trying my best to contribute. I am working on my game, I am
asking for information about Inform 7 because I have problems with the
manual and I have decided to collect suggestions on it here so that
something may change.
And by the way, I have been active in various online communities.
People saying "There is not enough stuff that is what I like" are
plenty around them. You may change the phrasing, but when you start
saying "only my way is the right way" and really push that point,
that's just the same thing. Asking for something to be revolutionary
but at the same time stringently defining what is good and what is no
good sounds rather contradictory to me.
Finally, I fail to see the benefit of revolution over evolution. As
long as there are some results, either one is equally valid in my
perception and so neither should be a reason to condemn a community.
Sure, I have to admit that at a few times, I have whined about the
Inform 7 manual. I have to admit that I gave in to that weakness
because of the hard time I was having finding specific information in
it. I also admit that one or two of the comments I made were unfairly
exaggerated. Still, my general intent and what I am trying to do is
contribute to a community who has already given me quite a lot (like
Inform 6 and Inform 7, the game "theatre", and much much more).
Deathworks
steve....@gmail.com
> > You can choose between facing the issues
> > and thinking about them further, or describing them as prohibitively,
> > impossibly complicated and simply give up.
>
> effort.
This isn't about people having different priorities for innovation;
it's about people having zero interest in any innovation. AI isn't
going to be everybody's interest -- it's only one example of what
could be happening but isn't. I'm sure there are lots of other viable
directions for innovation.
The problem is not that people are pursuing *other* directions; the
problem is that people are pursuing *zero*. Nobody is interested in
the previous research, and nobody is doing any new research.
IF would own innovation in the gaming field, if IF developers were
interested. As it is, it's dead and irrelevant.
jon.i...@gmail.com
innovation in IF - not insurmountable, but pretty big. The first is
the lack of "time": something like Facade could do a totally different
type of interactivity than most IF, and writing "drama" is quite hard
when tied to a turn-by-turn clock. If any gets round to making a
practical I7 real-time extension, I'd be keen to have a play with it.
Secondly, the IF interface is incredibly broad. No-one can pretend the
IF parser is intelligent, and although the XBox pad is dumb by
comparison, it can often offer a more intuitive interactive
experience. Writing immersion in IF is a very time-consuming process
because there's a huge verb-space to handle.
In the end, though, I think we've slightly priced ourselves out of the
market. Authors want to write games which are polished, well-written,
slick and responsive. This "wastes" development time that could be
used on creating and deploying new characterisation and interaction
strategies, and makes those new strategies extremely difficult to
render properly. I think to aid experimentation we need to cut back
our expectations a little.
jon
Harry Giles
> I'm of the opinion there are weighty factors preventing serious
> innovation in IF - not insurmountable, but pretty big. The first is
> the lack of "time": something like Facade could do a totally different
> type of interactivity than most IF, and writing "drama" is quite hard
> when tied to a turn-by-turn clock. If any gets round to making a
> practical I7 real-time extension, I'd be keen to have a play with it.
>
> Secondly, the IF interface is incredibly broad. No-one can pretend the
> IF parser is intelligent, and although the XBox pad is dumb by
> comparison, it can often offer a more intuitive interactive
> experience. Writing immersion in IF is a very time-consuming process
> because there's a huge verb-space to handle.
>
> In the end, though, I think we've slightly priced ourselves out of the
> market. Authors want to write games which are polished, well-written,
> slick and responsive. This "wastes" development time that could be
> used on creating and deploying new characterisation and interaction
> strategies, and makes those new strategies extremely difficult to
> render properly. I think to aid experimentation we need to cut back
> our expectations a little.
I think there's one other major obstacle to overcome, and that's how
much time we as IF authors have to commit. The vast majority of us are
enthusiastic amateurs with lives in other areas of work to lead. What
Steve is talking about probably takes more time and energy than many
of us are willing to commit. Amateurs and people under time
constraints do tend towards conservatism for these reasons alone, I
suppose. In a way Steve's criticisms do hit home, though I think
anyone would object to their delivery -- and I do think there are
reasonable reasons for people's limitations.
*On the other hand*, one thing I do think it's worth paying more
attention to is Steve's comments about co-operation. The vast majority
of IF games are solo authored. This seems to me to be a little odd for
a number of reasons: first, co-authoring of any form of fiction is an
exciting and dynamic process, which can lead to exciting results, so
I'd think more people would be willing to try it out; second, IF
requires a very specific skill-set (an ability to write fiction
combined with an ability to code combined with an ability to recognise
IF's specific literary requirements), and so solo authoring massively
limits the pool of potential authors; thirdly, co-operation is
*exactly* the way to meet the final obstacle you list: the desire for
polish "wasting" devlopment time. It seems to me that teams of writers
would be able to accomplish a great deal in terms of scale and
innovation, with different people working in different areas. I
thought the approach Textfyre has taken -- although that's not leading
in the direction of innovation, at least not in the sense Steve means
-- was extremely sensible.
Why is it we've a culture of solo authorship? I think partly because
that culture still dominates in literature. I think partly because
we've few chances to meet in the flesh, which always helps precipitate
conversation, though one would think, contra Steve, that IFMud would
lead to more collaboration. And I think partly because writing IF
remains competition-driven, which tends to encourage solo authorship.
Naturally, I'm biased here, and inexperienced: both IF's I've written
were collaborative: the first (unfinished) I wrote the text for with
someone else writing the code; in the second (for IFComp 2007),
collaborative partly to manage the scale of the project and partly
because it's just a pleasing way to write, we both took on both tasks.
But I do think it's worth considering more collaboration as a route to
more innovation. Once this second game's finished, I'm planning to do
a call-out on some imagined projects I have to see if anyone would
like to collaborate on them. I hope they will.
Another approach would be to announce a Collaboration Competition at
some point, of course . . .
Excuse me n00b thoughts, but I hope they've been at least vaguely
interesting.
Lx.
Deathworks
Harry Giles wrote:
...
> attention to is Steve's comments about co-operation. The vast majority
> of IF games are solo authored. This seems to me to be a little odd for
> a number of reasons: first, co-authoring of any form of fiction is an
> exciting and dynamic process, which can lead to exciting results, so
> I'd think more people would be willing to try it out; second, IF
> requires a very specific skill-set (an ability to write fiction
> combined with an ability to code combined with an ability to recognise
> IF's specific literary requirements), and so solo authoring massively
> limits the pool of potential authors; thirdly, co-operation is
> *exactly* the way to meet the final obstacle you list: the desire for
> polish "wasting" devlopment time. It seems to me that teams of writers
> would be able to accomplish a great deal in terms of scale and
> innovation, with different people working in different areas. I
...
(I left out the rest for brevity, but I am also taking all that
followed into account)
I find this a very interesting view. However, I have to admit that I
am currently having a bit of a hard time warming up to the
colaborative idea even though I can clearly see its benefits (add that
there is someone you can share even the slightest progress with).
Please allow me to describe my thoughts and feelings on this. Maybe
you, or someone else, can give me advice or tell me how to re-evaluate
the situation that a more colaborative approach would make sense to
me:
First of all, as a sidenote, I have a lot of unfinished projects of
various kinds on my harddisk, including several Inform 6 ones.
Usually, it is the classic case of biting of more than I can chew,
which is why I am currently careful to make sure that I try to define
things as being small.
Anyhow, the main issue I have with a collaborative approach is the
intimacy between me and my stories and my characters. They are all
something very personal to me and I easily get touchy about them. So,
I would be very protective about my characters and always kind of
paranoid that co-authors could mess up their personalities (the best
way to the end of friendship).
Another problem I would probably have would be finding someone willing
to cooperate with the really weird settings, or rather the weird fan
service the game is bound to include - yes, my games are to be adult
content even if they may not be explicit - the characters and the
settings are all aimed at that. However, this is not to say that the
eye-candy is all. Beyond that, I am open to various things, but it is
very probable that the weird stuff will leak in. In a way, it is that
stuff that gives me the strength to write the game. (Let's say, I will
be more than content if I will not be totally condemned by the IF
community for the game I am currently working on.)
Finally, I am a bit worried about the organization aspect:
distributing work and tasks and what not, deadlines, handling delays,
co-ordinating conceptual re-examination and all that kind of stuff.
So, any suggestions or thoughts on that would be appreciated as I
believe, despite the doubts described above, that colaboration could
greatly enhance my chance to finally finish a game (or at least be
part of a crew who finished a game).
Deathworks
Emily Short
> Emily Short writes:
> > auto-generation precluded having any kind of interesting content.
>
> That is undeniably an overstatement. You are overstating because
> you're afraid to accept partial successes. You say it would be zero
> interest, when in fact its only fault is that it would be a work
> towards future work. Plus I already described like three years ago a
> perfectly workable algorithm for action-report grouping.
Sorry; refresh our memories? I'm sure people would be glad of it.
> You never
> even tried it
Er, well. Since (I confess) I don't recall what your algorithm was,
I'm not sure; but I did get far enough to solve some of these problems
for simple actions, so as to collate movement and action commands in
Mystery House Possessed and in a few examples in the I7 manual.
The problem I had was that it was impossible to see how to get from
those to something *narratively* interesting: the example I gave was
getting the characters to engage in automated yet plausible dialogue.
And *that* I do think is a problem too hard for me to even work on. I
don't think we can get to plausible conversation constructively --
that is, by writing more and more sophisticated sentence-generators
based on grammar, and adding nuance. Every attempt I've seen along
those lines (mine and other people's) has been visibly false,
obviously clunky and inhuman. My remaining ideas for a viable solution
involve doing statistical analysis of large databases of interaction
(both at the level of pragmatics -- how do people fight? how do they
express interest in each other? etc. -- and at the sentence level --
how are things phrased? what forms express what nuances?), then
refining dialogue generation techniques from those templates.
This would require not only lots of manpower (to tag and process the
content), but also expertise in linguistics and computer science that
I don't have (to determine *how* it should be tagged and processed),
and physical resources I can't offer; not to mention a source database
of information I don't know how to acquire. I suppose I could make it
my life's mission to get a degree in a relevant field, followed by
grants to cover that kind of research -- but since I already have
other competing life missions, that's not going to happen.
> and now you're saying what, that the whole of NPC
> automation is entirely without hope? Bullshit you don't believe that.
> You're just trying to sound off.
I didn't say that -- for that matter, I've released several games
involving various types of NPC automation over the past couple of
years. I said that automation pursued for its own sake, as a primary
goal, without a more careful definition of the domain of action, was
not going to get me where I want to go: I'm not going to be able to
start with NPCs walking from room to room and (even with quite a lot
of modeling effort) make progress to NPCs who have compelling fights
and love affairs. This is because the problem of portraying fights and
love affairs involves different challenges from portraying characters
who walk from one room to another. If I want to arrive at that
outcome, I'm going to need to start by working on a set of problems
that are at least relevant to the outcome; it's no good working on a
different set of problems entirely.
> > and it's still not clear to me that once you'd
> > done all that, the outcome would have been *good*.
>
> Even if it fell on its face, it would be not *good* but *great*
> compared to your standard IF claptrap.
Here's where we disagree (again): I don't believe that technical
sophistication is the only measure of excellence in IF, nor the only
form of innovation. "Rameses" was a very innovative piece, not because
it did anything remotely difficult to code, but because it pointed out
new ways to make use of the interactivity of IF.
I also don't think "exploring a theme or issue well" is easy.
> But
> you're letting the problem lay you flat; you gave up way too early.
> You can't bear to produce something that isn't "successful" (in the
> sense that it's within convention). That's a recipe for innovation
> paralysis.
This paragraph made me laugh.
Anyway, if you think there's work that needs to be done here, why are
you looking for me to do it? Why don't you? By your own arguments, you
have the knowledge, the tools, and the passion for the work.
> > This isn't to say that no one should try a given experiment; but my
> > own reading is that some of the obvious ways forward *have* been
> > tried, the experiments were worthwhile and revealing, and the result
> > was that those ways don't work -- or, at least, that they don't work
> > unless they're applied to a well-defined problem.
>
> Do me a big favor. Revise that paragraph. Take out all the hedging and
> bullshitting and lip-service and slipknots make your point. I think
> you're saying that you tried AI but you didn't know what you were
> doing and you still can't figure it out.
You've always been fond of rewriting me; but in this case, I'm happy
to agree with the content of your remark: yes, I didn't know what I
was doing. However, a large part of not knowing what I was doing lay
in not defining the problem properly.
What I'm trying to do now is understand the value of interactivity in
interactive narrative well enough to apply technical innovations where
they will actually advance the art as art.
Jim Aikin
>
> Why is it we've a culture of solo authorship? I think partly because
> that culture still dominates in literature. I think partly because
> we've few chances to meet in the flesh, which always helps precipitate
> conversation, though one would think, contra Steve, that IFMud would
> lead to more collaboration. And I think partly because writing IF
> remains competition-driven, which tends to encourage solo authorship.
I can only speak for myself. I generally have very specific ideas about
what I want to do. Those ideas are largely intuition-driven. At the
stage where I'm designing a story, I might be inclined to respond to a
collaborator's idea with, "No, that will never work," without really
wanting to take the time to explain in detail why I can see at a glance
that it will never work ... or at least, it will never work given the
unstated intuitive vision toward which I'm groping.
That said, suggestions are often incredibly useful. I'm certainly not
saying that I think I can get it all perfect without outside input.
(Compare Last Resort with Lydia's Heart if you want an example of how
outside input can influence the development process.) But the question
of how one negotiates with that input -- i.e., who's in charge -- is
delicate.
I'd love to have help with the coding, of course. I can do that stuff,
but it's not quite so pleasant.
--JA
Emily Short
> innovation in IF - not insurmountable, but pretty big. The first is
> the lack of "time": something like Facade could do a totally different
> type of interactivity than most IF, and writing "drama" is quite hard
> when tied to a turn-by-turn clock. If any gets round to making a
> practical I7 real-time extension, I'd be keen to have a play with it.
Hm, interesting. I don't know how hard this would be; there is, of
course, a real-time library for I6, but long enough ago that I don't
think it could be effectively merged with the I7 library.
Personally, I thought the real-time aspect of Facade was one of its
weakest points, in that it meant I constantly ran out of typing time,
and my comment wasn't understood in the intended place in the
interaction. I also wasn't crazy about... Border Zone, wasn't it? And
the handful of other real-time games I've played since have been more
frustrating than fun, so I'm pretty dubious about the efficacy of
combining real-time with a typing interface. But that's just me,
obviously! If someone turns out something cool, I'll be interested to
try it.
> Secondly, the IF interface is incredibly broad. No-one can pretend the
> IF parser is intelligent, and although the XBox pad is dumb by
> comparison, it can often offer a more intuitive interactive
> experience. Writing immersion in IF is a very time-consuming process
> because there's a huge verb-space to handle.
This is where it becomes imperative to whittle the verb space down to
what's relevant for the game you're writing, and then relentlessly
communicate to the player what his scope of action is.
> In the end, though, I think we've slightly priced ourselves out of the
> market. Authors want to write games which are polished, well-written,
> slick and responsive. This "wastes" development time that could be
> used on creating and deploying new characterisation and interaction
> strategies, and makes those new strategies extremely difficult to
> render properly. I think to aid experimentation we need to cut back
> our expectations a little.
I think expectations can be managed somewhat if you pick and choose
the context in which you release something and the notes you give with
that release. IF art show pieces are usually extended a bit of
latitude (though it's true that there's an emphasis on thoroughness of
implementation that might go against what you describe). But I think
something like "Attack of the Robot Yeti Zombies" did a fine job of
framing itself so that people knew what it was going to be; and if you
release something as an experiment/toy/philosophical exploration,
people tend to look at it and discuss it in those terms. I'm thinking
of the Suprematism examples released some months ago; Jesse McGrew's
dining philosophers simulation, posted on ifMUD; Victor Gijsbers'
Figaro example; there are probably others that aren't leaping to mind
right now.
Point is, it *is* viable to share sketches with other authors, and to
get discussion and feedback, without working up a full game showing
off the completed concept in completely polished form.
James Jolley
ems...@mindspring.com says...
> > I'm of the opinion there are weighty factors preventing serious
> > innovation in IF - not insurmountable, but pretty big. The first is
> > the lack of "time": something like Facade could do a totally different
> > type of interactivity than most IF, and writing "drama" is quite hard
> > when tied to a turn-by-turn clock. If any gets round to making a
> > practical I7 real-time extension, I'd be keen to have a play with it.
>
> Hm, interesting. I don't know how hard this would be; there is, of
> course, a real-time library for I6, but long enough ago that I don't
> think it could be effectively merged with the I7 library.
>
> Personally, I thought the real-time aspect of Facade was one of its
> weakest points, in that it meant I constantly ran out of typing time,
> and my comment wasn't understood in the intended place in the
> interaction. I also wasn't crazy about... Border Zone, wasn't it? And
> the handful of other real-time games I've played since have been more
> frustrating than fun, so I'm pretty dubious about the efficacy of
> combining real-time with a typing interface. But that's just me,
> obviously! If someone turns out something cool, I'll be interested to
> try it.
I'd certainly try it but they have a lot of disadvantages for those
using screen readers.
Emily Short
> I thought the RAP was realtime or is it still turnbassed then?
RAP is turn-based, yeah. You can control how many things the NPCs are
allowed to do in a turn -- whether they're allowed to string together
a bunch of actions or just do one step of a process at a time -- but
that's different.
Mike Roberts
> Why is it we've a culture of solo authorship?
My theory is that so much IF is solo-authored because IF is the kind of game
that solo authors can still write. My evidence is anecdotal, but many
people here have mentioned that it's one of the things that attracts them to
the form. If we could just as easily solo-author something like, say, Halo
3, I suspect a lot of us would be off doing that instead. Then we'd have
artistic fulfillment *and* $170 million.
(And, conversely, people who like to work collaboratively and want to write
games have more options than solo authors for the kinds of projects they can
pursue effectively. Many of those non-IF options have larger built-in
audiences, and some even have salaries. Maybe some of these
collaboration-minded people still choose IF, but maybe some of them don't;
my guess is their representation here is fractional.)
--Mike Roberts
mjr underscore at hotmail dot com
James Jolley
ems...@mindspring.com says...
working in order. I remember in one of your articles you saying it was
abstract and that kind of worried me a little. Is it really all that
hard to use?
Emily Short
> In article <1191091622.786373.261...@57g2000hsv.googlegroups.com>,
Weeell. The I6 implementation of RAP was hard to get working (I found)
because, while it provides a *structure* for planning, it did not
provide any actual plans: there was no code included to make NPCs do
things even as basic as picking up an object or opening a door. This
was made harder because I6 did not have an idea of actions that could
be performed by NPCs. This may be what I meant in an earlier article,
especially if it was written a couple of years ago: the framework
existed for complicated work, but it still required that the author do
quite a lot of development before he saw any interesting results.
The Inform 7 implementation is easier in a couple of ways: first, it
*does* come with a second extension with basic plans (and an example
game to show how it all works); second, it's now possible to say "try
Fred opening the selected door" (or whatever), without having to write
your own code for every NPC action. (It is also, I think, currently
incompatible with the latest release of I7: I feel obliged to mention
this, before you rush off to try using it. I have looked briefly at
what would be needed to make it compatible with the latest version,
but it looked like it was going to need a bit of effort to sort out,
and my plate has been more than full, so I figured I'd let that go
unless Nate wanted to work on it or someone else asked.)
I have never tried the TADS 2 and TADS 3 implementations of RAP, so I
don't know how they compare.
James Jolley
ems...@mindspring.com says...
(It is also, I think, currently
> incompatible with the latest release of I7: I feel obliged to mention
> this, before you rush off to try using it. I have looked briefly at
> what would be needed to make it compatible with the latest version,
> but it looked like it was going to need a bit of effort to sort out,
> and my plate has been more than full, so I figured I'd let that go
> unless Nate wanted to work on it or someone else asked.)
> That is a shame. Is it both the planner and basic plans extensions that are incompatible? If this is the case, perhaps they should be removed or a message put up there.
Emily Short
> In article <1191095770.240463.51...@22g2000hsm.googlegroups.com>,
> emsh...@mindspring.com says...
There is already a message: if you check the description of Basic
Plans, it says "(Incompatible with versions of Inform after 4Uxx.)" I
think it worth leaving up, though, because I suspect it would not be
impossible to modify this to be compatible again, and it seems better
to let people see the current version than to take down all reference
to it and make them code from scratch.
Rubes
> Personally, I thought the real-time aspect of Facade was one of its
> weakest points, in that it meant I constantly ran out of typing time,
> and my comment wasn't understood in the intended place in the
> interaction. I also wasn't crazy about... Border Zone, wasn't it? And
> the handful of other real-time games I've played since have been more
> frustrating than fun, so I'm pretty dubious about the efficacy of
> combining real-time with a typing interface. But that's just me,
> obviously! If someone turns out something cool, I'll be interested to
> try it.
I also didn't like how real-time text entry worked in Facade for the
same reasons, so that was something we tried to keep in mind as we
adapted Vespers to real time. So we made a conscious effort to blend
"real-time" with "turn-based" play. I'm not sure if it truly works or
not, but I feel like we've done a decent job removing that sense of
time-restricted input. There are still a few tricky spots to deal
with, but we'll try to work around those.
Brian Slesinsky
I think it's important for most games to be one person's vision.
However, code review (that is, reading the code and making suggestions
and minor improvements) would be a nice way to get started and make a
contribution in less time than it takes to write a full game. (I
think that would be more fun than beta testing, but then I'm weird.)
But I think that only works for a project that seems "worth it"; that
is, by an established author who's writing something that's likely to
be a successful game. It's the opposite of doing a research
experiment where you have no idea how it will turn out.
- Brian
Adam Thornton
Jim Aikin <midig...@sbcglobal.net> wrote:
>Adam Thornton wrote:
>> Sure, everyone's got a story (and maybe a screenplay). Most of them,
>> alas, are not of interest to anyone except the author. Without some
>> reason to care that Mary got mad because Bob put her clothes in the
>> refrigerator and her beer in the dryer, their story--however engrossing
>> to them it might be--will be uninteresting to us.
>
>Depends on how you define "interesting."
[...]
>A true storyteller is someone who can engage us to the point where we
>care about people we've never met and have no special reason to care
>about -- people who, in point of strict fact, don't exist.
I think we're in violent agreement: do you see computers being able to
do this effectively until AI is, for lack of a better term, solved?
I think my point--and, possibly, Emily's--is that storytelling is a rare
skill among *humans*. Most *people* suck at it. Expecting a machine to
be able to do it well seems to me to be barking up a tree in the wrong
forest altogether.
Adam
Adam Thornton
Uh, I can't tell what you mean by "that."
Rule-based I7 and actor-plan-models certainly would make NPCs, with
their own in-game goals, easier.
The questions are: would this make a good game if these characters were
foregrounded (I believe "no"), and would this add entertaining
background color to deceive the player into thinking the game was
smarter than it was (I believe "yes, with caveats"). The caveats being
mostly along the lines of: don't let your autonomous actors mess with
the overall plot. You have to be careful, for instance, that the NPC
Sue, in an effort to remove all the knives from the game so that her
friend Mary, who's a cutter, can't hurt herself, doesn't steal the
Elvish Sword Of Great Antiquity you need to dispatch the troll and then
toss it in the incinerator.
Of course, some would argue that my notion of a fixed overall plot is a
straitjacket and that we ought, instead, to roll with whatever story
happens by allowing actors with differing goals to interact. In fact,
this is essentially the Aristotelean position I am so adamantly set
against, and I think it's going to lead to "games" (with no defined goal
in mind I'm not sure it qualifies as a "game") in which a whole lot of
nothing happens, and I don't think that's going to be interesting.
Adam
Adam Thornton
<steve....@gmail.com> wrote:
>The problem is not that people are pursuing *other* directions; the
>problem is that people are pursuing *zero*. Nobody is interested in
>the previous research, and nobody is doing any new research.
But this is demonstrably and self-evidently bullshit.
To pick one, obvious, glaring example, about which I know damn well you
know, I7 is pretty major new research into how you can create IF, about
how authors think about the stories they're creating, and about how to
blur the distinction between a producer and a consumer of IF.
You may not like it, but that doesn't make it not an experiment, and on
a pretty grand scale, too.
Adam
Sabahattin Gucukoglu
news:1191060714.2...@22g2000hsm.googlegroups.com...
> This isn't about people having different priorities for innovation;
> it's about people having zero interest in any innovation. AI isn't
> going to be everybody's interest -- it's only one example of what
> could be happening but isn't. I'm sure there are lots of other viable
> directions for innovation.
>
> The problem is not that people are pursuing *other* directions; the
> problem is that people are pursuing *zero*. Nobody is interested in
> the previous research, and nobody is doing any new research.
>
> IF would own innovation in the gaming field, if IF developers were
> interested. As it is, it's dead and irrelevant.
Here's the problem: you want people to do something they don't conceive
themselves capable of doing on a realistic scale without a return and
without any startup help from yourself or your minions. The return is
that, like you, we all want in our heart of hearts for IF to become as rich
and famous a proposition as the latest crap from
<insert-vendor-of-shoot'em'up-with-no-new-or-conceptive-ideas-in-it>. If
what I've read of you so far is anything like what I think you've been
saying, though, you *think* that somehow your ideas for convoluted, lengthy
pieces of half-baked "Game experience" with arguably unrewarding output to
show for it will somehow make IF more rich and famous. You think, somehow,
that the reasons for IF not being rich and famous have less to do with, for
instance, the medium of the story, the plain text nature of most works of
IF or the lack of multimedia in a lot of games or a different, more
attractive interface. Infocom, you suppose, was killed not by the
appearance of graphics but by the lack of a model for new character
interactions based on some academic research, rather than the fine,
well-tested and artistically proper techniques IF is famous for.
Don't ask us to admit defeat until you're sure you're targeting the real
failures of IF - at least as they are seen from the perspective of the
unwashed masses. When you're sure you've got something real we can improve
on, please let us know. You might, for instance, tell us that you've
emerged from your nuclear-attack-proof bunker and that your consultation
with the lizard alliance has resulted in a list of five important steps we
could take to get IF into the mainstream as a reality and not a memory of
that which never became anything at a cost no more prohibitive than the
startup investment in a company that would shepherd and requisition
anything needed to make it happen and sell products based on the results
that won great acclaim for being new, improved, innovative and - more
importantly - worth the whole rebuke you brought up, which is that you're
not convinced that, despite the fact that IF *should* be in the limelight,
it isn't. Otherwise, I'm afraid I simply don't see what you've been on
about the whole time or why you're just not taking the bate from the
others. Don't get me wrong - I do want IF to improve. Really. But come
on, put your money/time where your mouth is.
Cheers,
Sabahattin
--
Please DO NOT reply to sender.
Sabahattin Gucukoglu <mail<at>sabahattin<dash>gucukoglu<dot>com>
Address harvesters, snag this: fee...@yamta.org
Phone: +44 20 88008915
Mobile: +44 7986 053399
http://sabahattin-gucukoglu.com/
Josh Lawrence
I noticed an incompatibility when I tried Planner with 4U65, and it
seemed to be fixed by changing this line in the Executing Actions
section of Planner:
Carry out someone trying doing something:
to
Instead of someone trying doing something:
At least, the Alchemy demo included with Basic Plans seemed to run as
expected after that, and still does with 4X60. I haven't yet done
other tests outside the Alchemy demo to see if there's other issues,
though. I sent an email to Nate about it at the time, to which he
responded, so he knows about it, even if he hasn't uploaded a fixed
version.
Josh
Emily Short
> I think we're in violent agreement: do you see computers being able to
> do this effectively until AI is, for lack of a better term, solved?
>
> I think my point--and, possibly, Emily's--is that storytelling is a rare
> skill among *humans*. Most *people* suck at it.
I meant something more specific: that modeling low-level activity is
not going to produce an interesting story as emergent behavior, and
that the closer the model comes to modeling interesting narrative
content, the harder it is to auto-generate prose about the model. This
is my observation both of my own work and of the (considerably more
advanced) projects by other people.
(Modeling low-level NPC activity is not wholly pointless: it can, I
think, furnish more plausible backgrounds for a story, or provide
puzzles where the player has to work with/around/against the NPC's
manipulation of his surroundings. These are potentially worthwhile
things to do, and I'm glad to see that there are more tools than there
used to be for writing this kind of thing. But this is also not an
area for ground-breaking research. AI of this kind is well known in
commercial games, as witness the extensive chapters on pathfinding and
planning in computer game design books.)
So, storywise, you get this trade-off: you can model NPCs doing minor
actions (and describe these relatively well, albeit with some effort),
but this isn't really that interesting except as background to the
real story; or you can model NPCs doing narratively significant
things, but describe them so wretchedly that no one will care anyway.
It is possible to create a model whose purpose is to track NPC states
(whether through active goal-seeking or by following the NPC's passive
reactions to player action) and have those trigger pre-written
outcomes when the time is right, while allowing the player
considerable diversity of behavior in reaching those outcomes. This
approach is more promising, I think, and Facade comes closer to
getting it right than anything else I've seen.
Here you run into a big content-creation problem, though -- I think
one of the problems with Facade is that there wasn't enough content,
as absurd as that may sound given how much the authors put into it.
More significantly, you also run into problems with pacing, player
direction (conveying to the player how he can make a difference and
exercise his agency within this model world), and narrative structure
(creating a plot arc that feels organic). I think all of these
problems stem from having the goal of making an open-ended work where
the player can "do anything" and see what happens, rather than one in
which the player's range of action is well-defined and relevant to the
story/stories we want to tell.
To address Facade specifically, the desire to let the player do
anything reasonable meant not only accepting natural language input
(much of which couldn't be meaningfully resolved within the story and
therefore got ignored or, worse, misinterpreted, reducing the player's
sense of agency) but also letting the player make some physical
gestures (interacting with props, hugging and kissing characters). But
the interface did not (I felt) do a very good job of communicating
what these options were, and as a player I had a hard time taking
actions with any sense of intention. The experience rapidly distanced
me from emotional involvement and made me treat Grace and Trip like
toys rather than like people -- obviously the opposite of what we
want, in a work like this.
I say all this not to rag on Facade, which I think was a terrifically
worthwhile experiment; but to suggest that I think that it is not
possible to continue in the same direction and arrive at a satisfying
result merely by dint of greater effort.
> Expecting a machine to
> be able to do it well seems to me to be barking up a tree in the wrong
> forest altogether.
I think the more useful question to ask is: what part of the story I
am telling could be enhanced by a procedural component?
There are intriguing answers to that question in many cases; but I
think we have to start there, rather than with "how could a computer
mimic a human storyteller?" or "how can I make my NPCs automatically
behave like real people?". I am not at all convinced that the first
question can be answered in general, and I suspect that if we managed
the second (though how terribly far in the future?), we would find the
results rather unliterary.
jon.i...@gmail.com
> Personally, I thought the real-time aspect of Facade was one of its
> weakest points, in that it meant I constantly ran out of typing time,
> and my comment wasn't understood in the intended place in the
> interaction. I also wasn't crazy about... Border Zone, wasn't it?
That's true, but what I'm getting at is that text-input probably isn't
the best interface. The discussion, after all, was on whether IF was a
good testing ground for computer games / interactive storytelling in
general, and my feeling is that there are some things we can do well
and some things we can't. Time-based input is a pretty good device,
after all, and works a treat in action games and creating "immersion",
which we strive for. I would love to do a time-based dialogue-based
game, but a text-interface would hinder that - if the player is meant
to respond fast, they need a quick way of doing it. Some kind of
scrolling menu would be good, or a voice-recognition microphone. These
are valid experiments ... but I don't think we can do them.
> I'm thinking
> of the Suprematism examples released some months ago; Jesse McGrew's
> dining philosophers simulation, posted on ifMUD; Victor Gijsbers'
> Figaro example
Interesting. I didn't know about these (probably through my own
ignorance). More of a culture of this would be excellent (I feel in
the past there's often been a "don't waste a great idea" mentality).
jon
Emily Short
> That's true, but what I'm getting at is that text-input probably isn't
> the best interface. The discussion, after all, was on whether IF was a
> good testing ground for computer games / interactive storytelling in
> general, and my feeling is that there are some things we can do well
> and some things we can't. Time-based input is a pretty good device,
> after all, and works a treat in action games and creating "immersion",
> which we strive for. I would love to do a time-based dialogue-based
> game, but a text-interface would hinder that - if the player is meant
> to respond fast, they need a quick way of doing it. Some kind of
> scrolling menu would be good, or a voice-recognition microphone. These
> are valid experiments ... but I don't think we can do them.
Ah, hm.
So do you think we've tapped out the experiments that are possible
with the largely text-in, text-out interface? (I don't think so, but
I'm curious.)
Adam Thornton
Sabahattin Gucukoglu <fee...@yamta.org> wrote:
>You might, for instance, tell us that you've
>emerged from your nuclear-attack-proof bunker and that your consultation
>with the lizard alliance has resulted in a list of five important steps we
>could take to get IF into the mainstream as a reality and not a memory of
>that which never became anything at a cost no more prohibitive than the
>startup investment in a company that would shepherd and requisition
>anything needed to make it happen and sell products based on the results
>that won great acclaim for being new, improved, innovative and - more
>importantly - worth the whole rebuke you brought up, which is that you're
>not convinced that, despite the fact that IF *should* be in the limelight,
>it isn't.
Faulker and Cormac McCarthy (at least _Blood Meridian_-vintage McCarthy)
salute you.
I'll buy a drink for the first person to diagram that sentence.
Adam
David Fisher
news:1191113418.1...@d55g2000hsg.googlegroups.com...
> Some kind of
> scrolling menu would be good, or a voice-recognition microphone. These
> are valid experiments ... but I don't think we can do them.
Take a look at "Madam Bovary on the Holodeck" (warning: 1.7 Megabytes):
http://www-scm.tees.ac.uk/users/f.charles/publications/conferences/2007/mm07_cavazza.pdf
(Not that many people have the resources to try something like that).
David Fisher
jon.i...@gmail.com
Er... I think there are some experiments I'd like to make outside the
text-in text-out, yes, and I think there are things that text would
prototype well, which would actually be more playable with a different
interface (making for good experiments but less good "end results"). I
there are still things we can do in the text-world: but often they're
experiments discussing the text medium itself. I guess what I'm saying
is, I think the interface mechanism is a definitive choice.
jon
steve....@gmail.com
> > Emily Short writes:
> > > auto-generation precluded having any kind of interesting content.
>
> > automation is entirely without hope?
>
Well then what the hell does "precluded having any kind of interesting
content" mean?
> > I already described like three years ago a
> > perfectly workable algorithm for action-report grouping.
>
> Sorry; refresh our memories? I'm sure people would be glad of it.
Eliminate the royal plural, 'lest' you sound like a right cocky bitch.
And from whose ass did you pull "glad of it"?! (Graham all around,
'methinks.') Anyway, it was an email to you. If you had any authentic
interest you'd be able to recall it.
To others who might be interested: it's perfectly simple, and exactly
the kind of thing you'd figure out if you needed to solve the problem
in your own work. You catch the set of actions before they are
reported, sort them into (custom) groups, and utilize the group-
defined report-rubrics to print the messages.
Maybe you forgot because it's easy, or maybe you pretended to forget
because you're trying "to strenuously imply" that it's unworkably
complicated.
> This paragraph made me laugh.
You laugh when someone points out that fear of failure leads to
innovation-paralysis? Hm, let me guess... was it a nervous laugh? An
"it's funny because it's true" laugh?
Where's the woman who was bold enough to release "Best of Three"?
> You've always been fond of rewriting me[.]
No, I do not enjoy it, and I have not until relatively recently found
it necessary. I would much rather you speak plainly. Circumlocutionary
dodges are hostile and disruptive to intelligent discussion, and it is
only ethical (although socially difficult) to identify them as such.
> What I'm trying to do now is understand the value of interactivity in
> interactive narrative well enough to apply technical innovations where
> they will actually advance the art as art.
Oh right, you're reflecting thoughtfully. The hell you are. You're
just trying to sound off. You've lost your nerve. And no, I'm not
trying to bait you into regaining it; I'm just pointing out that
you've bullshitted yourself into a hole.
On the other hand, and as I said at the beginning, I believe that you
yourself have already done more than enough. As far as IF innovation,
you've earned your tenure. This does not mean that you're forgiven for
setting up blockages for others' future work (e.g., cowardly
describing viable research projects as 'vast', 'sophisticated' (in the
sense of highly programmatically complicated), and 'massive' -- yes,
all in a single sentence). But you certainly get a clean pass for
ceasing your own research.
steve....@gmail.com
> I'm of the opinion there are weighty factors preventing serious
> innovation in IF - not insurmountable, but pretty big. The first is
> when tied to a turn-by-turn clock. If any gets round to making a
> practical I7 real-time extension, I'd be keen to have a play with it.
This has been available to TADS for, what, a little shy of a decade?
> Secondly, the IF interface is incredibly broad.
Broad-ness as an obstacle to innovation?
> I think we've slightly priced ourselves out of the
> market. Authors want to write games which are polished, well-written,
> slick and responsive. This "wastes" development time that could be
> used on creating and deploying new characterisation and interaction
> strategies, and makes those new strategies extremely difficult to
> render properly. I think to aid experimentation we need to cut back
> our expectations a little.
Yes I strongly agree. It's a complex criticism: not only does
conservative-shine rob innovation for time, but it shames innovation
into foreclosure.
steve....@gmail.com
> > Nobody is interested in
> > the previous research, and nobody is doing any new research.
>
> But this is demonstrably and self-evidently bullshit.
>
> To pick one, obvious, glaring example, about which I know damn well you
> know, I7[.]
You could define 'research' broadly enough to include that. I am under
the impression that it doesn't provide any new capabilities. "Easier
for people scared of formal symbolism" is, arguably, a capability, but
not of the order I was thinking. Inspiring, new capabilities in
artistic treatment of interaction. In my sense, the innovation
inspires experts in other fields; coddling novices is nice, but I
don't really see it's related.
Maybe you can pick one other "obvious, glaring example" -- or failing
that, *any* example, obvious or not. Good luck. :)
Adam Thornton
>Eliminate the royal plural, 'lest' you sound like a right cocky bitch.
>And from whose ass did you pull "glad of it"?! (Graham all around,
>'methinks.') Anyway, it was an email to you. If you had any authentic
>interest you'd be able to recall it.
OK. Refresh *my* memory. *I'd* be glad of it. I don't get Emily's
email.
>Oh right, you're reflecting thoughtfully. The hell you are. You're
>just trying to sound off. You've lost your nerve. And no, I'm not
>trying to bait you into regaining it; I'm just pointing out that
>you've bullshitted yourself into a hole.
I'm just going to channel Autumyn DC and say: "bullshat."
Adam
steve....@gmail.com
> >Emily Short wrote, quoting me:
> >> Sorry; refresh our memories? I'm sure people would be glad of it.
> >Eliminate the royal plural, 'lest' you sound like a right cocky bitch.
> >And from whose ass did you pull "glad of it"?! (Graham all around,
> >'methinks.') Anyway, it was an email to you. If you had any authentic
> >interest you'd be able to recall it.
>
> OK. Refresh *my* memory. *I'd* be glad of it.
Then read the very next paragraph beginning, tellingly enough, 'To
others who might be interested:'.
Wow, I hope you don't have to go to work in the morning. :)
Adam Thornton
Alas, I do. But, as Winston Churchill said, "In the morning, I shall be
sober."
Fortunately, my work has nothing whatsovever to do with IF, Which is,
of course, much of the problem with your exhortations to the IF world.
Pay me my usual salary to do reactive-planner NPC agents for a year, and
I'll consider it.
Adam
Emily Short
> Emily Short wrote, quoting me:
>
> > > Emily Short writes:
> > > > auto-generation precluded having any kind of interesting content.
>
> > > [Y]ou're saying what, that the whole of NPC
> > > automation is entirely without hope?
>
> > I didn't say that[.]
>
> Well then what the hell does "precluded having any kind of interesting
> content" mean?
If I say that autogenerating prose about conversational interaction
makes the conversational interaction boring, that is not the same as
saying that there is never any point to NPC automation under any
circumstances. I outlined elsewhere why NPC automation might be useful
in some cases, and also why I think focusing on it as a primary goal
is not going to produce the kind of storytelling results I want to
achieve.
> > > I already described like three years ago a
> > > perfectly workable algorithm for action-report grouping.
>
> > Sorry; refresh our memories? I'm sure people would be glad of it.
>
> Eliminate the royal plural, 'lest' you sound like a right cocky bitch.
...
> Anyway, it was an email to you. If you had any authentic
> interest you'd be able to recall it.
I assumed you must be referring to something different and more
complicated than your private email to me, since you followed up by
saying that I hadn't tried it; whereas, in fact, I have published
multiple works that collect actions and put the reports together, and
I know you've seen at least some of them. This happens in Mystery
House Possessed; it also happens in a couple of pieces of released I7
source code. For that matter, Savoir-Faire did something similar to
describe the results of throwing objects, before this ever came up
between us.
Anyway, it would be perfectly reasonable for you to have articulated a
more complicated structure of action reporting than we discussed back
then; the hard part is not coming up with the idea "list actions, then
assemble report from them", but in the details of that reporting: in
deciding how to group elements, how to assign priority to different
content, when and how to glue on subordinate clauses, how to
distinguish between necessarily contemporary and necessarily
sequential actions, and so on.
> To others who might be interested: it's perfectly simple, and exactly
> the kind of thing you'd figure out if you needed to solve the problem
> in your own work. You catch the set of actions before they are
> reported, sort them into (custom) groups, and utilize the group-
> defined report-rubrics to print the messages.
I can expand on this a bit, I suppose. I wrote in my review of Battle
of Walcott Keep something which got quoted in your email as
====
> [T]he generation of dull, repetitive, or unidiomatic prose is [a] hard [problem]. I think the solution probably involves representing all the actions that are going to occur in a given turn as data, then running some algorithms on that data to discard anything too ineffective to be worth reporting, cluster related actions together into compound sentences and related sentences into paragraphs, introduce variations of phrasing, etc.
====
and then after further exchange I said
====
>Anyway, RAP is something I'm interested in, which is why I wrote so much detail in my review. Like I said, I've tried to work on it some myself, and I have come up with a module (in Inform, obviously) that does store data and then print it, with some event-stacking and combination; it's not anywhere near perfect, though.
====
to which you replied:
====
> Yes, that's very cool. In fact, we were thinking of doing something like filtering RAP-actions through a database. My ideas on this are pretty trivial and undeveloped. We decided against pursuing this, as I may have said, for to expose the mechanics of the expo.
====
and then you elaborated later on the design you envisioned:
====
> Actions are executed implicitly (no "default messages"), and then registered in a stack of actions.
> Each message_generator object (defined by the system or customized/defined by the user) is passed this stack (by the message_handler), and returns the number of actions it can report. The one which returns the highest number wins. (This way, the more sophisticated message_generators override the less sophisticated ones where possible.) That message_generator is then instructed to generate a message for that group of actions. The group of actions reported are removed from the stack, and the process repeats.
> Actions which are left over (i.e., which resolve no capable message_generator) are then reported with the default report. (e.g., "Bill takes the bag.")
====
I did something slightly different from this detailed proposal --
partly because I was working with a rule-based system for most of the
relevant works, but mostly because I found that just optimizing for
quantity-of-actions-reported didn't quite handle everything I wanted
to deal with. Among other things, it wasn't always clear how to decide
between two equally voracious description rules which ought to get
priority. Would I rather have a paragraph that says
"Bob and Pete walk into the dining room. Bob is carrying a gun."
or one that says
"Bob walks into the dining room, carrying a gun. Pete walks into the
dining room."
?
Though in each case we have a 2-fact report followed by a 1-fact
report, the latter text is, I think, clearly inferior unless I want to
suggest that Pete *follows* Bob, in which case I probably want to
change the sentence a bit to say so explicitly. As they are, the first
paragraph sounds like something a native English speaker might write;
the second sounds defective.
So I found that I preferred an event-grouping algorithm that applied a
slightly more explicit hierarchy.
I also wanted to be able to have a reporting rule like
"Bob and Lily walk into the dining room, arm in arm."
take precedence over
"Bob and Lily walk into the dining room."
when Bob and Lily are lovers: the fact of their being lovers is not
part of the action I'm reporting, but an external circumstance, so
again I need to do more than count actions-reported. Inform 7 rule
precedences mostly prioritized these things automatically for me in
Mystery House Possessed, and then I think in a few places I hand-
tweaked the order to achieve a desired effect.
Anyway. If "plan and perform actions silently in the world model, then
group and describe them to the player" is all you meant, then yes, of
course I've done it; there is in fact sample code (perhaps not ideal,
but one potential approach), and a brief write-up of why this task is
interesting, in the latest version of the I7 documentation (in the
Recipe Book, under Other Characters > Traveling Characters). I haven't
forgotten how.
I think there *is* more progress to be made on the action-collating-
and-printing front, but it no longer lies in the area of raw
experiment (can we collate these things and print them? does that give
better results than just printing a string of reports?) but in the
areas of tool-building (how do we write libraries that make it easy
for lots of authors to design specialized reporting of NPC behavior?)
and refinement (if we have multiple authors doing this in multiple
games, what kinds of different reporting approaches emerge? how do
different authors nuance their reporting for the sake of a particular
game? what does this teach us about categories of report content?).
Nick Montfort's dissertation gets into more detail about the
refinement side -- how specifically can reports of actions vary, why
might we want to vary them, how do we categorize the distinctions. I
know you were not especially excited by what he had to say, but I
found it well worth reading the relevant chapters.
I also now think this method of prose generation is not limited in
value to NPC behavior, and that list -> group -> report misses an
important step. I wound up writing a room description extension that
does something similar, but the steps are in that case
1. list items to print
2. cull items that are to be suppressed
3. order list by item priority
4. group and report, running the grouping rules to look for the most
interesting thing it can say about the first list item; that might
include mentioning other list items as well, in which case they are
removed and not consulted again; the definition of "most interesting"
might have to do with how many other list items can be included, but
other circumstantial features might also determine precedence.
The preliminary filtering and sorting of the list make it easier to
run comprehensible grouping rules on the result, because the grouping
rules do not themselves have to contain information about the relative
priority of the things they describe.
The same is true to a degree with NPC actions, but I think the room
description example is in some ways easier as a technical place to
start because it avoids some of the temporal challenges
(distinguishing necessarily contemporary from necessarily sequential
events, for instance). Still, the more I fiddle with this kind of
problem of producing an interesting description based on a
juxtaposition of elements -- items in a room, actions by an NPC,
whatever -- the more often the algorithm winds up looking like that.
List; cull; sort; group; print. Or group & print as one move,
sometimes.
This question of modeling the discourse level in some way separately
from the world model is an area of considerable thought in the IF
community just at the moment. We have Nick's dissertation; we also
have the input of people like Ron Newcomb and Jeff Nyman (to name only
two) explaining how standard prose writing constructs might translate
to an IF model. This discussion has largely been in criticism of how
Inform 7 doesn't go far enough to support a writing (as opposed to
programming) mindset, but the suggestions are ones that haven't come
up before with other IF languages, or not in the same degree of
detail. This suggests to me that thinking about IF creation in a new
way engenders new thoughts about what IF might do.
However, I'm not convinced that that sort of task is relevant at all
as research towards AI-in-story-telling in commercial (which is to
say, usually graphical) games: the industry is great at pathfinding
and chains-of-small-tasks AI, and because the presentation layer is
visual, it doesn't have to worry about constructing prose; it can
represent actions in analog, by *showing* characters moving around.
While this poses some challenges, they're totally different challenges
and fairly well-covered by existing expertise. This question of how to
glue a lot of actions (or anything else) together in a way we can
describe verbally is a problem for IF only because we stick to the
text medium.
> Maybe you forgot because it's easy, or maybe you pretended to forget
> because you're trying "to strenuously imply" that it's unworkably
> complicated.
Describing simple actions is not unworkably complicated. But it's also
not the interesting part of the problem. Unworkably complicated is
making a model to auto-generate an argument between two characters and
then manufacture an appropriate set of bile-spewing remarks between
the two of them, taking into account all their past personal history,
their causes of irritation just and unjust, the nuances of their
diction, and so on: humans have a gift for this sort of thing that I
don't expect to be able to capture in code.
If you can't auto-generate, you have to go with the pick-from-finite-
body-of-prewritten-prose solution, which limits the value of modeling
all this in the first place. Or you wind up with Chris Crawford's
Storytron lingo for talking to characters, which does not appeal to
me.
> You laugh when someone points out that fear of failure leads to
> innovation-paralysis? Hm, let me guess... was it a nervous laugh? An
> "it's funny because it's true" laugh?
Nah; I laugh because all my work fails. There is nothing I write that
I don't eventually hate and want to disown. The process is about
ignoring that fact for as long as it takes to release each new thing.
> Where's the woman who was bold enough to release "Best of Three"?
Mm. Technically innovative work is not the only kind that is scary to
release.
> > What I'm trying to do now is understand the value of interactivity in
> > interactive narrative well enough to apply technical innovations where
> > they will actually advance the art as art.
>
> Oh right, you're reflecting thoughtfully.
Possibly not where you've seen it, since you're not privy to
unreleased WIPs and libraries, may not be reading my blog, and (as far
as I know) were not present at the ACM Hypertext conference in
Manchester a couple of weeks ago; but yeah, I have been working semi-
systematically on this question.
The methodological premise: it is more useful to consider how
interactivity can contribute to narrative than it is to ask how we can
cram a successful narrative into an interactive format. Discussions of
interactive storytelling, especially those from the mainstream gaming
industry, often take the latter approach. But this posits a conflict
between story and the player's agency in it; it also encourages us to
think in terms of process first, especially of how to make the
computer behave like a story-telling human, rather than in terms of
desired result first, and then of how to make the computer accomplish
the aspects of that task that are procedural, while leaving the rest
of the work to a human author.
So, then: what does narrative gain from interactivity? My list, as
I've been drawing it up lately, contains five things:
-- exploration; the player's freedom to select and pursue themes in
the work, and to acquire expository narrative at a pace of his own
choosing; the author's corresponding freedom to include much more
material than any one player is likely to need or desire. [Several
recent works play with this: Floatpoint a bit, but even more some of
Eric Eve's work; both Blighted Isle and Elysium Enigma leave it very
much up to the player how much of the backstory he cares to find out,
and which pieces he finds relevant. This is not really a new concept,
though; Worlds Apart has lots of tangential content. The interesting
question is how to make the tangential content an important part of
the whole, even if the player only experiences part of it, and without
making the player feel *obliged* to experience all of it. (If he feels
obliged to find every piece, and especially if he must do so in a
given way, then this is not an aspect in which the work is really
interactive: it only means that the task of reading all the pieces is
harder than it would be in an un-interactive book.)]
-- challenge; in IF this mostly involves the well-charted territory of
puzzle, but challenge contributes to a story by increasing the
player's investment, both of time and emotional energy; a work too
exclusively composed of choices and outcomes, without challenge or
exploration, is like to be seen as ineffective and badly paced. The
Baron suffers from this a bit; and I see the lack of challenge and
exploration as likely problems with the storytelling universe Chris
Crawford wants to create, which is All Choices, All The Time. I don't
think this means that we can't have puzzle-less IF, but I do think it
means that puzzles (or something like them) are more organic to the
storytelling aspect of the medium than earlier theorists really wanted
to admit.
-- complicity; a word often bandied about in theoretical IF
discussions, but the player's willingness to undertake the actions
required in a story (I believe) commits him to the outcome in a way
that he would not be committed merely by reading the end of the story
in a book. It also, of course, means that there are probably some
stories that will not be told in IF form: I doubt most of us would get
through Lolita if we had to act the narrator's part, with all the
sordid violations that entailed. (See the judges' reactions to
Rendition in the latest IF comp: I think the ones who found it most
successful were the ones who were too repelled to finish it. I tried
it, but also quit after a move or two. This is obviously an edge
case.)
-- the constraint of character (which I sometimes list as "role-
playing", but what I mean is "inhabiting a role of the author's
choice", not at all the same thing as "creating a role of the player's
choice"). Rameses is the classic example of the character the player
explores from the inside; less obviously, Varicella, or the Kissing
Bandit, or Act of Misdirection, in all of which the player learns to
adapt attitudes, strike poses, make gestures, etc., appropriate to the
other rather than to himself. The ability to find the limits of the
character dynamically (if we do it well) answers the objection that
people often find themselves having with static literature: why does
the protagonist do *that*? Why doesn't he just ______ instead? What,
is he stupid? In IF you can ask these questions of the story and --
assuming the author has framed the problem and written the responses
well enough -- get an answer. ("I don't recognize that verb." is, of
course, a disappointing let-down, which is why I'm secondarily
obsessed with designing to communicate range of action to the player.
If you teach the player how to play the interesting parts of the game,
he's less likely to run into the boundaries.)
And finally, of course, the much debated
-- choice: challenging the player to prefer one outcome of the story
over another, and in the process to consider the implications of doing
so. (Insert a mass of dialogue with Victor Gijsbers, Stephen Bond, et
al., here, much of which has taken place on RAIF.)
Now: I accept that this may all strike you as not interesting, or it
may seem at most to be a way to analyze what is already done, rather
than to facilitate what may be done in the future. But I think it
does, actually, point at some opportunities for modeling and
procedural approaches.
Some of this is not new:
-- If we want our challenges to provide pacing and structure, then we
design them to adjust themselves dynamically to the player's progress
(by throwing in new clues, or redirecting the story when something
gets too sticky). Commercial game designers are already on to this
idea. We could make progress on this in IF by modeling a little more
explicitly what the player knows and how suggestions are leaked to him
-- and also perhaps (by noting what he's mentioned and tried to
interact with) guess what he thought was significant and what he
passed over in the first reading of the text.
This is another reason why I'm intrigued by the discussion of
discourse modeling, and by the idea in Nick Montfort's dissertation of
a "suggester" voice as a distinct aspect of the narrator. (I had a
couple minor comments on this here:
http://emshort.wordpress.com/2007/08/13/two-readings-of-possible-interest/
). I think in the case of IF we could do some very interesting things
by explicitly tracking what we think the player has already been shown
or figured out, and drawing his attention gently and then more plainly
towards the things missed.
Bronze was a rather limited try in that direction, one that attempts
to model the puzzle-functionality of the game so that it can
procedurally calculate how close the player is to solving a given
puzzle and what he needs to do next; but it serves that information up
in the rather clunky form of explicit hints, rather than through
subtle revision of the narrative itself.
A few IF games (including a few of my own) have tried the alternate
tack of making puzzles that come in different difficulty settings, but
such games are harder to design, harder to write walkthroughs for,
harder for players to communicate about. If the fundamental solution
of a puzzle changes from playthrough to playthrough, you can't tell me
how to solve the version facing me in mine; you only know how you
solved yours. Dreamhold does a better job with this idea than most
I've seen, but even there I thought that the easy version of some
puzzles meant that things were left lying around as red herrings that
were essential in the hard version; this was inevitable but
unfortunate.
But if we instead focus on the way in which a puzzle is described (and
the rate and persistence with which further clues are offered to the
player) as the key determiner of its difficulty, and make the
"suggester" prose vary separately from the rest of the text, it
becomes much easier to design consistent but flexible puzzles and also
to scale the challenge during play-time (rather than forcing the
player to choose a level at the outset).
I'm not trying to kill the pure puzzle game here: there are games
where I don't want extra hints or spoilers or spoonfeeding, but want
to take however long I take to solve it the hard way. That's fun too.
I just played such a game tonight, and it was really satisfying. But
in the context of storytelling, a lot of the time we're going to want
to adjust the difficulty, and that's going to require both a model of
what the puzzle is and a way to translate that information into hints
and suggestions woven into the narrative and action responses.
-- Re. exploration: if we want exploration of an area, or a
conversation, to be responsive to the themes the player has already
uncovered, then we need both a way to track those themes and the drama-
management to direct what gets found; this again is stuff that has
been discussed some here, and discussed some in the commercial and
academic arenas.
The hard part here is not how to steer the exploration or conversation
(it's technically easy enough to plant objects in the next place the
player happens to look, for instance, and only somewhat more difficult
to vary directions of NPC speech depending on how we want a scene to
go). The hard part is largely in the design: what are our important
themes? how do we articulate and re-articulate them? how do we decide
what the goal states are? Floatpoint was partly an experiment into
this question -- how do we set up a game in which every time the
player starts to choose one outcome, he is confronted with a counter-
argument against what he's decided to do? A design experiment, not a
coding experiment.
-- Re. choice, the procedural challenge is to manage the combinatorial
explosion of the narrative if we let the player meaningfully branch
the story too many times; the design challenge is to come up with a
kind of drama that is manageable. I really admire the Facade approach
of trying to build narrative tension, etc., in an organic way, but for
the purposes of choice-rich narrative, I'm more interested in drama
management at a macroscopic level, where the individual scenes are
handcrafted by the author but the underlying model determines which
scenes are eligible to happen next. One can envision both a fuzzy form
of this (rate all the remaining scenes with a number indicating its
degree of relevance and appropriateness to the dramatic pacing; pick
the one that scores highest) and a non-fuzzy form (as the author, I
know all the possible states the game can get into; I just don't know
in what order the player will achieve them; at any given moment there
is precisely one correct "next scene", and the program is just sparing
me the bother of writing out all the options as an explicit tree.)
My recent work on this (in an unreleased WIP) takes a sort of middle
path: at any given point in the unfolding of the story there is only
one major event that is appropriate to have happen next, given the
choices the player has made in the story, *but* that major event may
need some additional set-up, if the player has missed pertinent facts
in the exposition so far or otherwise hasn't been appropriately
grounded to experience it.
So the story generator selects the event, looks at the prerequisite
requirements for that event, and supplies appropriate lead-in scenes
first if needed. The result is a certain amount of variation in
texture -- some scenes are relatively quiet and uneventful, existing
only to set the desired atmosphere or let the player stumble across a
piece of information, while others are the big set-pieces. It... well,
it has a long way to go yet, because though the model is promising so
far, it needs a huge load of content, and there are points where the
segues aren't terribly smooth, and passages where the player can feel
railroaded by the set-up-then-follow-through sequence. Still, the
framework of the model is built and a large web of scenes created;
that's my present take on drama-management, but it's arranged around
the question of how we usher the player through a long series of
significant choices without the game flying apart at the seams, and
how we ensure sufficient build-up and pay-off for choices to feel
individually important.
Anyway: so far, so (relatively) pre-chewed. It's complicity and
constraint-of-character (or expression-of-character, if you like)
whose implications I find the most interesting, and the least
procedurally developed, in IF or in any other game context.
-- Re. constraint/expression of character: yes, a lot of this has to
do with designing well; communicating to the player what he can do in
this world and where his agency lies; providing appropriate negative
feedback for reasonable actions that this particular character won't
perform (like all Rameses' reasons for not saying the various things
that are on his conversation menus); providing interesting optional
verbs (like the Kissing Bandit's TWIRL MOUSTACHE) to let the player
get into his role; hinting to the player what the verbs might be; and
so on.
But I wonder whether there is something more we can do, a way we can
be clever about turning false moves into true ones, the unimplemented
actions into viable parts of the story. Someone recently asked about
whether it was possible to *force* the player to express the player
character's emotions (e.g., by getting him to hit things with a
baseball bat when enraged). I'm not sure we can do that, exactly, but
I'm interested in the idea that any given action in an IF model might
be categorized in terms of its expressive content. Actions might then
not have a single default response, but several, or even dozens; we
might look at the context of the narrative, determine whether the
player's last command (otherwise meaningless) is one that could be
treated as a gesture of frustration, decide that frustration is a
reasonable thing for the player character to be feeling at the moment,
and narrate accordingly. Under other circumstances, the game might
find a different interpretation for the same gesture. The narration
would not necessarily have any effect on the progress of the story per
se, but it would give meaning to behavior that would otherwise have
been pitched out as meaningless. This is a bit akin to Facade's
attempt to clump player input into meaningful categories before
responding to it, only in this case we're not dealing with a mass of
natural language input, so attaching (some) meaning to given input may
be easier. It's just another application of Grice's maxim of
relevance, this time between the player and the parser: the parser is
to assume that everything the player does or attempts to do is a valid
part of the performance of the player character, rather than that some
of it is an error (an assumption carried over from a more antagonistic
model of the player/parser relationship).
I don't envision this giving rise to a selection of default responses
that could be used across the board from game to game, though possibly
the abstract action categorization could be shared from one game to
another, at least partially.
I also realize that it would be possible to go too far with this and
get something weirdly clownish, in which the player character's every
slight twitch was blown up into some drama-queen enactment. A light
touch would be needed with the writing. But I think it would be
interesting, and worth a try.
-- Questions of complicity are most interesting when we push the
player just to the point where he's uncomfortable, but not to the
point where he stops playing: it's always those games that have made
me squirm in exactly this way which I remember as especially,
magically compelling; and as particularly distinct from the experience
of any written book.
This one, I don't know what to do with yet. Suppose the goal is to
determine how the player is reacting to a scene and adjust the
intensity accordingly; how do we do that (barring bio-feedback
devices)? Some progress might be made by doing statistical studies of
time to enter commands, use of meta-commands, etc., to characterize
the behavior of players under situations of stress vs. situations in
which they did not feel concerned about what was happening; but here
again we're trailing off into a kind of experiment for which I haven't
really got the resources, and I'm not quiiiite sure what I'd do with
the results, either. Still, now and then I turn this over in my head.
Maybe brinksmanship with the player's tolerance isn't the right goal
anyway; if I think about it the wrong way it starts to sound less like
literature and more like an S&M session. I don't know. I understand
complicity least, of the items on this list, so it fascinates me most.
Oh well: such are my recent thoughts, deep or not. There's some other
stuff about the persuasive value of simulating choices and their
consequences, the ethical presentation of problems, etc., but it's a
bit further afield from what we were talking about, and less likely to
yield technical challenges immediately.
But you didn't answer my other question; why aren't *you* doing this
work you think needs to be done? Or, if the answer is something you'd
rather not share, how about this instead: what are the near-term goals
that you dream of seeing fulfilled? What do you think is the next
step? What *would* you be working on, if whatever-it-is didn't prevent
you from being able to?
Harry Giles
> "Harry Giles" <harry.lodest...@gmail.com> wrote:
> > Why is it we've a culture of solo authorship?
>
> My theory is that so much IF is solo-authored because IF is the kind of game
> that solo authors can still write. My evidence is anecdotal, but many
> people here have mentioned that it's one of the things that attracts them to
> the form. If we could just as easily solo-author something like, say, Halo
> 3, I suspect a lot of us would be off doing that instead. Then we'd have
> artistic fulfillment *and* $170 million.
Mm, I suspect you're right. As literary culture still encourages the
Romantic ideal of the author, that spills over into game-writing as
well . . .
Lx.
Harry Giles
> Hi!
>
> Harry Giles wrote:
>
> ...> attention to is Steve's comments about co-operation. The vast majority
> > of IF games are solo authored. This seems to me to be a little odd for
> > a number of reasons: first, co-authoring of any form of fiction is an
> > exciting and dynamic process, which can lead to exciting results, so
> > I'd think more people would be willing to try it out; second, IF
> > requires a very specific skill-set (an ability to write fiction
> > combined with an ability to code combined with an ability to recognise
> > IF's specific literary requirements), and so solo authoring massively
> > limits the pool of potential authors; thirdly, co-operation is
> > *exactly* the way to meet the final obstacle you list: the desire for
> > polish "wasting" devlopment time. It seems to me that teams of writers
> > would be able to accomplish a great deal in terms of scale and
> > innovation, with different people working in different areas. I
>
>
> Anyhow, the main issue I have with a collaborative approach is the
> intimacy between me and my stories and my characters. They are all
> something very personal to me and I easily get touchy about them. So,
> I would be very protective about my characters and always kind of
> paranoid that co-authors could mess up their personalities (the best
> way to the end of friendship).
This is one of the most common problems I hear people give: that they
feel not only intellectually but also emotionally attached to their
characters. I'm going to rant a bit now:
My main artistic work, which is theatre, is a medium where the culture
of artistic collaboration is well established. More specifically, I do
clowning, of a kind where co-operation is *essential*. Our
performances are improvised, spontaneous, and collaborative. Put a
group of clowns in a street situation and you never know how they're
going to react -- and nor do they. The one thing that is certain is
that they will react *as a group*. When we're being clowns, there is
constantly a dialogical background to the behaviour: we're constantly
aware of what others are improvising, adapting ourselves to that
imporvisation. No clown can attempt to rule the performance, or the
performance will collapse. It has to be a collaborative performance.
So I'm used to some of the things that makes collaboation work. Even
in static media, like fiction, we can apply some of the lessons from
collaborative improvisation. And one of the first is to loosen the
intellectual possessiveness you have over your creation. To continue
to feel that what you create is your idea and yours alone, and that
its success depends on the succesful communication of your idea, is to
adopt an unstable and incoherent position. Just as a simple example:
a work of art which preaches to the audience, which demands the
audience feel a particular way, is usually an unsuccessful work of
art; a work of art with ambiguity (the Mona Lisa's mysterious smile?)
will provoke far more discussion and appreciation.
What this involves is cultivating humility -- adopting a humble
attitude to your art. Loosening your grip so that that your ideas
alone don't matter *too* much. When you do this, you open yourself up
to allowing other people's ideas to contribute to the creation in a
dynamic way. When you have a team of people working on an artwork, it
never goes the way any single one of them planned: the artwork emerges
as a product of dicussion, debate, compromise and co-operation. I
usually find at they very least that *better* ideas emerge from a
collaboration, because anything included everyone has to be happy to
include -- the idea has to pass a lot more muster. At the very least
the product is always exciting for everyone involved, because nobody
knew exactly what it was going to be like.
So you're right: it's impossible to write something like IF if you
feel that it has to be the work of a single creator. But I would argue
that people don't have to feel that way, and that if they do feel that
way they should at least try collaborative work. It feels entirely
different, and it may be only inhibition that's holding you back from
it.
> Another problem I would probably have would be finding someone willing
> to cooperate with the really weird settings, or rather the weird fan
> service the game is bound to include - yes, my games are to be adult
> content even if they may not be explicit - the characters and the
> settings are all aimed at that. However, this is not to say that the
> eye-candy is all. Beyond that, I am open to various things, but it is
> very probable that the weird stuff will leak in. In a way, it is that
> stuff that gives me the strength to write the game. (Let's say, I will
> be more than content if I will not be totally condemned by the IF
> community for the game I am currently working on.)
This is obviously a more specific problem to overcome. The problem is
that the AIF community has in the majority produced works that really
are embarrassing -- not through the fact that they involve sex, but
just through the lack of quality. This means that there's an automatic
connection made: AIF = rubbish. Fair on unfair, anyone writing IF with
an adult element will find the onus is on them to demonstrate they can
do it well. (On the other hand, most readers tend to be sceptical
about even new standard IF, so bad is the worst of it.)
I'm probably going to face this problem myself in a bit. The next IF I
have planned is going to be all about sex. It's not going to involve
an interactive sex scene, I don't think, but it will involve elements
of sex that I haven't really seen in IF outside of the IF-as-
pronography AIF scene. And I don't want to be writing porn. I'm going
to struggle to find a way of making it all work, especially given the
extreme sex-scepticism (however well-grounded) of the AIF community.
None of this helps you very much, I suspect. Especially since it leads
me to say: the best thing you could do is demonstrate that IF with
fanservice can be written well, and then more people will be likely to
want to collaborate on it.
> Finally, I am a bit worried about the organization aspect:
> distributing work and tasks and what not, deadlines, handling delays,
> co-ordinating conceptual re-examination and all that kind of stuff.
Oh, this is the easiest bit. Unless you live a solitary life you have
to do this kind of activity day-in, day-out. If the people you're
writing with are reliable, and you trust them (and that is a
prerequisite for literary collaboration), he organisational aspects
just fall into place, they really do. There are no significant
problems to overcome unless you end up with someone who doesn't co-
operate, and then you just have to end the project.
I hope some of these thoughts helped, or provoked something.
Lx.
Harry Giles
Oh, bum, I forgot something: the fact that there's an emotional
attachment for you to your characters makes eveyrthing I jut wrote
much harder, because its easier to be humble about an idea than an
emotion. I can only suggest here developing a greater sense of
detachment -- the aim, after all, is not to make a character you
relate to but one the reader relates to, and the former in no way
necessarily producs the latter.
Lx
Harry Giles
> Harry Giles wrote:
>
> > Why is it we've a culture of solo authorship? I think partly because
> > that culture still dominates in literature. I think partly because
> > we've few chances to meet in the flesh, which always helps precipitate
> > conversation, though one would think, contra Steve, that IFMud would
> > lead to more collaboration. And I think partly because writing IF
> > remains competition-driven, which tends to encourage solo authorship.
>
> I can only speak for myself. I generally have very specific ideas about
> what I want to do. Those ideas are largely intuition-driven. At the
> stage where I'm designing a story, I might be inclined to respond to a
> collaborator's idea with, "No, that will never work," without really
> wanting to take the time to explain in detail why I can see at a glance
> that it will never work ... or at least, it will never work given the
> unstated intuitive vision toward which I'm groping.
>
Hallo Jim -- I've just replied to Deathworks about this, so I won't
repeat myself here. But I thought I'd add something about this:
> That said, suggestions are often incredibly useful. I'm certainly not
> saying that I think I can get it all perfect without outside input.
> (Compare Last Resort with Lydia's Heart if you want an example of how
> outside input can influence the development process.) But the question
> of how one negotiates with that input -- i.e., who's in charge -- is
> delicate.
Hmm, well, I'm used to working in anarchist situations when it comes
to art (and much of life, for that matter). I've developed experience
and tools for non-hierarchical collaborative work. They weren't easy
to develop, but they do develop naturally when you put your mind to
it. There's plenty of literature to prompt you along, as well. But if
the idea frightens you, you can always have a creative hierarchy and
still have collaboration, so long as everyone agrees on the terms from
the beginning. Much theatre involves this more than IF-with-
suggestions does: an artistic director who has to order and make sense
of the input from actors, producers &c., while still allowing them
enough creative freedom.
Basically, I don't think art is a special creative actvitity: the same
structures that we use throughout life can be used in art.
Lx.
Deathworks
I guess the emotional bit is the one most difficult for me to
overcome. The thing is, I am powering my activity with the love I have
for the characters I create. So, without them being themselves, I am
likely to run out of steam pretty quickly.
In addition, I am extremely character based in my creations (thus far
non-interactive): After choosing a very general situation/plot, I
design or choose the characters who are to participate in the
situation and then I allow them to write the story for me. It is kind
of a relationship of mutual trust where I know that they will lead me
through the story if I simply ask them what they will be doing next.
This may be inefficient and selfish, but I am afraid I am not the kind
of person who can easily overcome such emotional bonds.
Actually, this might be not that much of a problem if I sticked to
NPCs. However, my characters are also rather idealized in various odd
ways (this is why I love them), which may easily make it difficult to
fit them in with more realistic, darker settings.
Anyhow, thank you for your responses and I will spend some time
thinking about this (I will be riding on the bus a few hours today, so
I really have the time to spare (^_^;; ).
About the AIF community, I have begun checking in on their portal and
have read the back-issues of their newsletter. My impression is
somewhat mixed (the general attitude towards sexuality is a bit
stronger than I feel comfortable with). I found some of the articles,
especially the 101 to be very interesting to read. So, I guess I will
officially place my game (if I ever get it finished (^_^;; ) into the
realm of AIF, despite some doubts in various directions.
Deathworks
Busman1215
>
> read more »
Let's all ignore Steve's frequent personal attacks and accusations of
people being this-and-that and saying this-and-that. Our time could be
spent better developing our next IFs. Steve doesn't know that good
manner and respect are essential in conversations with other people.
Emily, I really like the way you write. I wish I could write like you
do.
Busman
"The best way to do it is to do it." -- Amelia Earhart
Jim Aikin
improv theatre and clowning are a useful counterbalance to my control
freak tendencies.
As a collaborative enterprise, improv theatre is probably at one
extreme. For one thing, the output is gone in an instant -- nobody is
likely to scrutinize it months afterward (until you get good enough that
you're taped for the Comedy Channel).
I play in a 5-piece band, which is somewhere in the middle of the
spectrum: We have a bandleader (who in my opinion doesn't exercise
enough control). The musical arrangements are rehearsed, and different
people are loosely in charge of different pieces. If I wrote the tune, I
get to dictate in general terms what sort of parts I would like. It's
then up to the bass player and drummer to attempt to follow my
suggestions. So there's dialog, and a definite leader/follower dynamic,
but ultimately it's very free-form and spontaneous. (It's also pretty
sloppy, at times.)
Traditional theatre is a collaboration. Ultimately the director is in
charge of (and responsible for) everything, including changes in the
script, but the director has to depend on the actors and set designer to
come up with innumerable details. So that's sort of the conservative end
of the collaboration spectrum.
Playing in a symphony orchestra isn't collaboration at all, except in
the sense that the conductor "collaborates" with a dead composer.
A collaboration in creating a work of IF might proceed along any of
these lines. If I had time (I don't), I might be interested in doing an
improv-type collaboration, in which the one rule is that you don't get
to say "no." Whatever happens, you have to say "yes."
On the other hand, since I often say "no" to myself in writing of any
sort -- deleting sentences that don't work, rejecting poor puzzle
design, and so forth -- I suppose I would need to be able to say "no" to
others' ideas as well. I think the "say yes" rule might have to apply
mainly to the larger features of the game.
Or it might not work at all in this context.
--JA
Rikard Peterson
> Playing in a symphony orchestra isn't collaboration at all, except in
> the sense that the conductor "collaborates" with a dead composer.
That's up to the conductor. A conductor doesn't have to be a dictator.
Jim Aikin
Well, I can look at that in two ways, neither of which results in my
agreeing with you.
The snarky response is, "Name one who isn't."
The substantive response is, I've played in several community
orchestras, one of them quite good and others not so good. My experience
has been that, above all, rehearsal time is limited. An effective
conductor knows exactly what he or she hopes to get and can reasonably
expect to get from the group. The conductor asks for specific things.
If necessary the conductor may spend a few minutes working through a
given passage in order to get as close as possible to the interpretation
he or she has in mind, but I don't believe I've ever heard a conductor
ask an individual flautist or French horn player how a passage might
best be interpreted, much less an individual member of a string section.
I've heard them pause in rehearsal to consult with the concertmaster on
bowings, but that's a technical matter, and has little to do with
interpretation. If the concertmaster prefers a bowing that the conductor
feels will run counter to his or her intended interpretation, the
suggested bowing will be rejected.
So I guess I'd say, "No, you're wrong. A conductor _does_ have to be a
dictator." I don't think that's a bad thing, necessarily. If you want to
play in a symphony orchestra, you just accept it. It goes with the
territory. Orchestra players are footsoldiers, not generals.
If you want to improvise and make a real contribution to the
arrangements and interpretations, join a jazz group. That's what I do.
--JA
Adam Thornton
Jim Aikin <midig...@sbcglobal.net> wrote:
>The substantive response is, I've played in several community
>orchestras, one of them quite good and others not so good. My experience
>has been that, above all, rehearsal time is limited. An effective
>conductor knows exactly what he or she hopes to get and can reasonably
>expect to get from the group. The conductor asks for specific things.
Yeah, but when you said "a symphony orchestra," *most* of us imagined, I
wager, a professional symphony.
When you have professional musicians, not enthusiastic amateurs, as your
players, I imagine that their individual styles and beliefs are given a
lot more weight.
>If necessary the conductor may spend a few minutes working through a
>given passage in order to get as close as possible to the interpretation
>he or she has in mind, but I don't believe I've ever heard a conductor
>ask an individual flautist or French horn player how a passage might
>best be interpreted, much less an individual member of a string
>section.
Hell, I've seen this fairly often in both the Rice and Princeton scatter
bands. If there's a solo, or indeed sometimes when the music simply
features one section, the conductor will spend a while figuring out what
the soloist or featured section thinks they want versus what the
conductor wants, and it's very much a negotiated process.
>So I guess I'd say, "No, you're wrong. A conductor _does_ have to be a
>dictator." I don't think that's a bad thing, necessarily. If you want to
>play in a symphony orchestra, you just accept it. It goes with the
>territory. Orchestra players are footsoldiers, not generals.
Is this the case when you're talking about a professional orchestra?
Adam
George Oliver
> If I had time (I don't), I might be interested in doing an
> improv-type collaboration, in which the one rule is that you don't get
> to say "no." Whatever happens, you have to say "yes."
>
> On the other hand, since I often say "no" to myself in writing of any
> sort -- deleting sentences that don't work, rejecting poor puzzle
> design, and so forth -- I suppose I would need to be able to say "no" to
> others' ideas as well. I think the "say yes" rule might have to apply
> mainly to the larger features of the game.
>
> Or it might not work at all in this context.
>
> --JA
>
I'm interested in collaboration as well, and with Lx's (or Giles'?
Which do you prefer?) comments in mind, Jim, something that might aid
the situation of determining the creative agenda in a piece of improv
collaboration are some of the frameworks that have been developed
recently for roleplaying games. I've read many of your posts now in
RAIF threads so I have some idea of your general thoughts on creative
collaboration and IF.
The example I have in mind is a game called 'Polaris'. I'm going to
throw around some terminology here so please forgive me. You can learn
more about this game here:
http://www.story-games.com/codex/index.php?title=Polaris
In the basic version of this game four people sit around a table and
play characters in a scene. In a typical game there are four
protagonists, and during a scene there is one main protagonist.
However I won't bore you with all the details, the part I want to get
to is how Polaris determines what happens in the narrative. Much of
the time is called 'free play', where players go back and forth
describing the setting, saying dialogue, etcetera. When there is a
point where the players may differ on how the scene develops, the
players determine the narrative using conflict phrases, which are:
1. But only if...
2. And furthermore...
3. It shall not come to pass.
4. You ask far too much.
5. It was not meant to be.
6. That was how it happened.
7. It was no matter.
8. We shall see what comes out of it.
Now, obviously it is not this simple. The mechanics of the game are
designed to deal with these four players, and handles what happens
when a player uses, for example, 'It shall not come to pass'. But I
think the use of a framework like this has a lot of promise for improv
collaboration, and I'm betting that there are other frameworks out
there, in improv games or drama theory. Maybe someone else can help me
out here.
best, GO
James Jolley
> Rikard Peterson wrote:
> > Jim Aikin <midig...@sbcglobal.net> wrote in news:fdra72$hfq$1...@aioe.org:
> >
> >> Playing in a symphony orchestra isn't collaboration at all, except in
> >> the sense that the conductor "collaborates" with a dead composer.
> >
> > That's up to the conductor. A conductor doesn't have to be a dictator.
>
> Well, I can look at that in two ways, neither of which results in my
> agreeing with you.
>
> The snarky response is, "Name one who isn't."
>
> The substantive response is, I've played in several community
> orchestras, one of them quite good and others not so good. My experience
> has been that, above all, rehearsal time is limited. An effective
> conductor knows exactly what he or she hopes to get and can reasonably
> expect to get from the group. The conductor asks for specific things.
>
> If necessary the conductor may spend a few minutes working through a
> given passage in order to get as close as possible to the interpretation
> he or she has in mind, but I don't believe I've ever heard a conductor
> ask an individual flautist or French horn player how a passage might
> best be interpreted, much less an individual member of a string section.
I have had situation when I was asked my views on how to best interpret
a specific passage as a classical pianist but perhaps that is different?
Rikard Peterson
> Rikard Peterson wrote:
>> That's up to the conductor. A conductor doesn't have to be a
>> dictator.
>
> Well, I can look at that in two ways, neither of which results in my
> agreeing with you.
>
> The snarky response is, "Name one who isn't."
If you want a name, I'll offer my own, small and humble as it is. Even
though I do practically all my conducting in my role as music teacher
where all the musicians are my students (ages 9-19) and I know more about
music than they do, there is still some room for collaboration. But that
was not what I was primarily thinking about when I wrote my post. (Read
on for more.)
> The substantive response is, I've played in several community
> orchestras, one of them quite good and others not so good. My
> experience has been that, above all, rehearsal time is limited. An
> effective conductor knows exactly what he or she hopes to get and can
> reasonably expect to get from the group. The conductor asks for
> specific things.
I agree. That's how it should be, and that also fit my experience of good
conductors.
> If necessary the conductor may spend a few minutes working through a
> given passage in order to get as close as possible to the
> interpretation he or she has in mind, but I don't believe I've ever
> heard a conductor ask an individual flautist or French horn player how
> a passage might best be interpreted, much less an individual member of
> a string section.
Maybe not. I can't recall a conductor asking the orchestra for advice,
but any decent conductor will listen to the musicians' interpretations of
the music. It doesn't have to be verbal communication, but I've heard
that too. Then it's up to him to decide how it should sound, but
communication and collaboration is happening. (More or less, depending on
both person and situation.)
> [snip] If the concertmaster prefers a bowing that the conductor feels
> will run counter to his or her intended interpretation, the suggested
> bowing will be rejected. [snip]
Yes. The ultimate decision is the conductor's to make. That far we agree.
In a group of that size, there has to be someone to make the decisions or
it would all become very inefficient, so in a way I made a poor choice of
words. As you say, I was wrong. A conductor _does_ have to be a dictator,
and that's a good thing. But what I'm trying to say is that that doesn't
automatically remove all collaboration. It's done in a different way than
in a quartet or in a jazz group, but it can still be done.
Rikard
Jim Aikin
>
> Is this the case when you're talking about a professional orchestra?
I would imagine it depends on the conductor ... but with 100 musicians,
it would be a foolish conductor who spent much time consulting them
individually about anything.
I have a friend who plays full-time in the San Francisco Symphony. I'll
ask him and get back to you.
--JA
steve....@gmail.com
torpedoing of the discussion. Nobody has time to actually work through
your post, and you know it. If you really wanted to contribute, you'd
work to condense your ideas into a useful range.
Emily Short wrote:
> If I say that autogenerating prose about conversational interaction
> makes the conversational interaction boring, that is not the same as
> saying that there is never any point to NPC automation under any
> circumstances.
Why 'if'? Are you not saying, but only suggesting that you might have
been saying?
That's just the first sentence. The rest is equally disingenuously
complicated.
Anyway, I'm not going to carefully read a rambling, how do you say so
cute ... 'wudge' ... of quotations from emails, random thoughts
expressed roundabout at best, with a copy/paste WIP essay thrown in.
Nor is anybody else I'd guess. But I think I can scan it quickly and
extract the few points. In the future, please edit your own work. It's
disrespectful to your readers, to make them wade through a huge mess.
> I think there *is* more progress to be made on the action-collating-
> and-printing front [...] in the areas of tool-building[.]
Ok one decent point. Oops, that was my point. Seemed not so long ago
you were describing this work in various shades of hopeless. Ok, now
you seem to agree that it's not as entirely hopeless as you were
saying. Great. Anything else? Oh yes...
> Nick Montfort's dissertation[....]
Oh dear, you think this thing exists? Well, yes the dissertation in
all its coffee-stained glory is available for download -- not that
you'd want to read it. But the program he's pretending to refer to ..
'nn' .. do you imagine that it actually exists? The main tell: Nick is
perfectly capable of writing, but he doesn't do it here. My guess, he
wrote it in three weeks. You can tell he's talking about an entirely
hypothetical project.
Oh, you should have seen his powerpoint slideshow.
> I wound up writing a room description extension
Well god bless you, maybe there's some innovation in the IF arena
after all. People are writing extensions to order the objects reported
in a room description.
No, I don't think anybody is going to be particularly inspired by
this. It's trivially easy work. Were you really calling this an
example of an innovation?
Look, sister, innovation is probably going to be related to
storytelling or characterization -- something interesting, anyway. I
don't want to delimit it of course. Simulation-world complexity is
another strong direction.
But no, it would have to be an awful fine room description extension,
if it were going to be innovative for IF and gaming in general.
> [W]hat does narrative gain from interactivity? My list, as
> I've been drawing it up lately, contains five things:
>
> -- exploration
> -- challenge
> -- complicity
> -- the constraint of character
> -- choice
> Some of this is not new[.]
Some? Heh, most of it is not new. Actually, none of it is new. That's
the problem. This doesn't even stretch back to the 1990s. We're in
'80s territory here. Well, technically 1970s. But let's be extremely
charitable and call it 1990s. What the hell?
Further, your comments on each of these matters tends towards "well,
here's what professional games are doing, and we can emulate that if
we do X." My point is we should be setting the pace, not playing catch-
up. And we're not even playing catch-up, we're just talking about
playing catch-up. It's pathetic upon pathetic.
Here's an unordered and incomplete list of things you might have been
thinking about instead:
-- fostering empathy: towards finagling NPC believability
-- automating narrative variation
-- stronger realism in simulation: towards making the world sandbox-
playable
-- IF community cooperation on complicated projects
-- development and encouragement of discussions on research issues
But even these are 1990s considerations at best. The problem is we
don't even have a vision of 2000s issues. We've been entirely passed
by. This is now a bullshit genre. Genres which are less-equipped for
dealing with subjects proper to us, are now *way* ahead of us. This is
now a toy genre, and any inspiration people can draw will come from
related genres or from what we accomplished back in the 1980s.
[snip excerpts from a long essay]
> Oh well: such are my recent thoughts, deep or not.
Not really deep, no. Publish the essay separately. Don't copy it to
newsgroups because you think it will help you save face.
> But you didn't answer my other question; why aren't *you* doing this
> work you think needs to be done?
Because nobody cares. Hence my disgust with the current state of the
community's near-absolute non-interest in research and innovation.
Adam Thornton
<steve....@gmail.com> wrote:
>> But you didn't answer my other question; why aren't *you* doing this
>> work you think needs to be done?
>Because nobody cares. Hence my disgust with the current state of the
>community's near-absolute non-interest in research and innovation.
Well, fair enough. I'm sure you have better places to be, then.
Please don't let the door hit you in the ass on the way out.
Adam
steve....@gmail.com
> > [M]y disgust with the current state of the
> >community's near-absolute non-interest in research and innovation.
>
> Well, fair enough. I'm sure you have better places to be, then.
I have more-or-less given up, yes -- on the idea that IF can be a
genre of innovation. But I can still hope that the genre will
reawaken, and I will complain when it doesn't. Still, it's seeming
damn unlikely, especially now that other genres have caught up to the
same expressive potential. There are indeed more productive places to
invest one's efforts. (Yes, there are better places to be.)
But even as a toy genre, IF is still really useful and interesting. It
will always be. And it may expand its capability when new technologies
become available. But it's basically lost its leadership.
Two sentence history of the decline of IF: initially it lost its
popularity because it wasn't graphical, and graphics became cool. But
then at the crucial moment when we needed its skills, it lost its
credibility, for we found it was not especially good at producing what
it was supposed to be great at: interesting interactivity.
Emily Short
> First of all, Emily, I consider your post a more-or-less intentional
> torpedoing of the discussion.
It was an intentional torpedoing of your claim that nothing is going
on. You can read about what I'm doing, and what I think of other
people's related work, if you're interested (as you claimed to be). I
didn't really expect the majority of RAIFers to follow along, though I
supposed a few might, if they were curious about current projects and
hadn't already read about them on my blog.
And no, I didn't copy and paste it from anywhere; and no, it wasn't
especially complicated, disingenuous, or rambling. It *was* several
hours worth of work to write, so I didn't miss the point that it might
take someone a little while to read. If you don't have time to get
into it now, that's fine, of course; but if you just skim it so that
you can continue your angry tirade, I'm afraid I am not going to take
either your critiques or your tirade seriously.
> > Nick Montfort's dissertation[....]
>
> Oh dear, you think this thing exists? Well, yes the dissertation in
> all its coffee-stained glory is available for download -- not that
> you'd want to read it.
Have you? Not the powerpoint slides but the dissertation itself?
Thoughtfully and with an open mind?
> But the program he's pretending to refer to ..
> 'nn' .. do you imagine that it actually exists? The main tell: Nick is
> perfectly capable of writing, but he doesn't do it here. My guess, he
> wrote it in three weeks. You can tell he's talking about an entirely
> hypothetical project.
Which is odd, since I played with things he'd written with nn. Perhaps
you think that no one is doing anything simply because you don't
bother to look at anyone's work long enough to understand it? I don't
think nn is a fully fledged system or anything close to it, and I did
skim some pieces of the dissertation, not having an immediate
compulsion to follow the discussion of Genette; but I would recommend
Nick's section on the different aspects of IF narration to anyone who
is thinking of working on narrative variation. It's the only sustained
work on the topic I've seen anywhere, and it contains insights I've
never encountered before. I've even explained some of them, and how I
think they might be put to work, back in that post you didn't read.
> > I wound up writing a room description extension
>
> Well god bless you, maybe there's some innovation in the IF arena
> after all. People are writing extensions to order the objects reported
> in a room description.
To cull, order, group, and report, for the purpose of extending the
kinds of narrative variation possible. It's the most basic groundwork
only, but it's a prerequisite to anything that might follow. Make fun
of room descriptions if you want, but if narrative variation research
happens in IF, a lot of it is going to be focused on how the setting
is described. Setting is critical in IF, and there are all kinds of
things we might want to do with it. Actions are both more fiddly
(because of issues about relative timing) and less useful (because in
the vast majority of narrative work, major actions happen in cut
scenes and most necessary variation can be done with a handful of hard-
coded variables).
> No, I don't think anybody is going to be particularly inspired by
> this. It's trivially easy work. Were you really calling this an
> example of an innovation?
I don't think I originally called it anything. I was describing, in
detail, what I thought was wrong with the algorithm you outlined for
report generation, and how it might be improved. This is an area that
*you* identified as needing research.
> > -- exploration
> > -- challenge
> > -- complicity
> > -- the constraint of character
> > -- choice
> > Some of this is not new[.]
The replaced punctuation there was a colon, which means that I was
referring to the speculation that followed, rather than to the
premises that came above. Anyway, I don't claim that this set of terms
is new. What I haven't seen is much discussion of how all these
elements fit together, how they serve the purposes of a *story*
(rather than being orthogonal features of a game), and how we might
enhance the advantages that they bring to interactive narrative
through greater procedural complexity.
> Some? Heh, most of it is not new. Actually, none of it is new. That's
> the problem. This doesn't even stretch back to the 1990s. We're in
> '80s territory here. Well, technically 1970s. But let's be extremely
> charitable and call it 1990s. What the hell?
>
> Further, your comments on each of these matters tends towards "well,
> here's what professional games are doing, and we can emulate that if
> we do X."
Hardly at all.
But maybe I'm missing reading that you've done; if you have citations
to share where people talk about puzzle-modeling (that is, abstractly
tracking what the player is working on, what he knows, and what he
needs to do next), or about allowing the player to act expressively in
order to explore his character (rather than to invent a new one of his
choosing), or about the uses of complicity in any depth at all, I'd be
interested to see them. I've read a few things about puzzle-modeling,
but most of the things I've looked at seemed to imply that the
programmers were hard-coding all the possible checkpoints along the
way to solving each puzzle, and that there wasn't much attempt to keep
track of the player's probable knowledge-state.
> -- fostering empathy: towards finagling NPC believability
I've explained repeatedly (though you've chosen not to understand me)
why I think goal-seeking behavior, open-ended drama management, and
dialogue generation are not on their own going to create a set of
believable NPCs who do narratively interesting things.
Empathy, if we're going to get it, will come from giving the player
the agency to do emotionally significant things to and with NPCs, and
then experience the aftermath. There may be some procedural aspects to
this, ensuring that all the player's reasonable options are provided
for and none of them break the game. The procedural solutions may even
draw on some goal-seeking and drama-managing ideas, though in a
modified form. But creating appropriate agency is going to be first
and foremost a design problem. I know you have no respect for design
innovations, only for technical innovations, but the latter is
worthless without the former.
> -- automating narrative variation
Which I talked about at some length back there where you didn't read
it.
> -- stronger realism in simulation: towards making the world sandbox-
> playable
Since IF is the wrong medium for sandboxes, I consider this, in its
general form, a waste of time. I do plenty of simulationist work, on a
case by case basis, as needed for particular games; you stopped
describing this work as interesting and started describing it as
cowardly and dull about the time I shifted to using I7. The output
hasn't gotten simpler -- if anything, the later work is considerably
more complicated -- so I've drawn my own conclusion about what your
objection actually is.
> -- IF community cooperation on complicated projects
This is not a research topic.
> -- development and encouragement of discussions on research issues
This is not a research topic.
> > But you didn't answer my other question; why aren't *you* doing this
> > work you think needs to be done?
>
> Because nobody cares. Hence my disgust with the current state of the
> community's near-absolute non-interest in research and innovation.
If you -- clever, passionate, full of ideas, attuned to what goes on
in related fields, equipped with an intimate knowledge of your
preferred IF language, and (a rare bonus) emotionally armor-plated
against any form of negative feedback -- if *you* cannot be bothered
to work on these problems, then you cannot possibly expect your
complaints to have a motivating effect on anyone else.
Brian Campbell
> On Oct 2, 5:33 pm, steve.bres...@gmail.com wrote:
> > -- fostering empathy: towards finagling NPC believability
>
> I've explained repeatedly (though you've chosen not to understand me)
> why I think goal-seeking behavior, open-ended drama management, and
> dialogue generation are not on their own going to create a set of
> believable NPCs who do narratively interesting things.
I'm not sure that trying towards believability is the correct approach
to fostering empathy. Something I've noticed in a few media (though
haven't seen it directly in IF) is that you can achieve better
empathetic results if you avoid the uncanny valley (
http://en.wikipedia.org/wiki/Uncanny_Valley for those of you who don't
recognize the term). For instance, I think one of the reasons that
Pixar does a better job of making you feel for their characters than
movies like the Final Fantasy movie is that the characters don't feel
like dead, somewhat zombie-like people, but instead animated toys,
fish, cars, and so on. Even their human characters are more cartoony
than realistic.
I've seen other examples of this. People love the robot Keepon,
especially in that video that's gone around YouTube (
http://www.youtube.com/watch?v=3g-yrjh58ms ). Does Keepon look like a
human? No, but the small amount of humanity imparted in something so
simple feels so much more alive than a detailed, realistic looking
humanoid robot that wasn't quite right would.
Applied to IF, this may mean that empathy could be better achieved by
inanimate objects that have been imparted with some sort of agency or
personality, or maybe animals who become your companion throughout the
game, or perhaps robots (I haven't played Planetfall, but from what
I've heard Floyd may be one of the most memorable NPCs in interactive
fiction). The point here is that getting NPCs to act just like humans
may never work, since they'll always feel somewhat phony, but getting
non-humans to feel like they have a spark of humanity may be much
easier.
S. John Ross
> The point here is that getting NPCs to act just like humans
> may never work, since they'll always feel somewhat phony, but getting
> non-humans to feel like they have a spark of humanity may be much
> easier.
And it's just plain nice :) It's always been a favorite technique of
mine in pen-and-paper RPGs. In one of my current IF projects, there's
a Slime [Uresia] who fits that bill, although whether he's a friend or
an enemy depends on how quickly the player figures out what kind of
hell the little fella's been through lately. Plus a thing about hand-
puppets, But I dig the whimsy.
And half-tangentialy, I'm beginning to worry that many of the same
differences between myself and my colleages in PnP design will become
differences between myself and other IF writers. This entire thread
reads like a massive deja vu, many times over ... I need an
Excedrin :)
But the good news is, the GAMES are starting to ROCK ...
David Fisher
news:1191224629.2...@50g2000hsm.googlegroups.com...
Emily, I loved your post - thank you for taking the time to write down your
thoughts.
> -- exploration
...
> The interesting question is how to make the tangential
> content an important part of the whole, even if the player
> only experiences part of it, and without making the
> player feel *obliged* to experience all of it.
It sounds pretty tricky to let the player know that the tangential content
contains no important information without saying so explicitly. Maybe a clue
could be dropped in-game - if the player keeps on pursuing the non-essential
information, then the game/narrator or an NPC could say something about it
being unimportant.
Sometimes I wish I could ask the game a direct question about this sort of
thing. Wasn't there a game with a "red herring detector"?
> I see the lack of challenge and exploration as likely
> problems with the storytelling universe Chris Crawford
> wants to create, which is All Choices, All The Time.
I still love the idea of having lots of freedom of choice, but then
constraining the plot to follow up on the player's decisions. If the player
is discouraged from or unable to keep on changing the direction of the plot
once they have made a clear choice, then it won't feel so directionless
(which is how I understand your objection).
A small example for clarity: the PC is offered a quest, and takes it up. If
the PC pursues something else, or just neglects the quest in some way, the
game first gives a reminder, then some minor consequences, then some major
consequences for not following the quest.
The ultimate in IF for me be to be able to make meaningful choices that are
(from the player's point of view) woven into the plot. I wouldn't expect to
be able to start a shoe selling business, get married and watch my children
grow up or create a new invention, but I would love to have a significant
choice about how I relate to each NPC in the game (within the general
constraints of the PC's personality, if the PC is well defined), and to have
lots of different ways of solving puzzles (eg. if there is a forest, being
able to find some leaves or a stick if I need them, or chop down a tree to
block the road - or at least try to) ... things that you really could do if
you were in the PC's position, and which there is no compelling reason to
disallow (from the player's point of view). As long as the game continued to
move the plot-as-it-currently-stands along, I think something like this
would work extremely well.
I sometimes feel you are a bit dismissive of the idea of being able to "do
anything" in a game. Without the plot being managed properly, I agree it
could be boring and directionless; but if it included tracking and following
up on the player's choices so far in the game, I think it would be
fantastic.
David Fisher
quinti...@gmail.com
> I have more-or-less given up, yes
Maybe if you hadn't turned into such a monumental prick people would
care what you have to say. Instead, you've become a self-parody. I
mean, when Adam Thornton -- probably the most patient person I've seen
on this group -- asks you to leave, you know you've gone past the
point of no return.
"Hey, look, I'm Steve Breslin! Do as I say not as I do!"
Adam Thornton
<quinti...@gmail.com> wrote:
>Adam Thornton -- probably the most patient person I've seen
>on this group
Wait, what?
Adam
Stephen Granade
Perhaps this is your evil bearless twin, a man of patience and
forbearance who is unacquainted with Stiffy Makane or TPB.
Stephen
--
Stephen Granade
stephen...@granades.com
David Fisher
news:7dabr02...@sargent.dyndns.org...
> ad...@fsf.net (Adam Thornton) writes:
>
>> In article <1191419422....@50g2000hsm.googlegroups.com>,
>> <quinti...@gmail.com> wrote:
>>>Adam Thornton -- probably the most patient person I've seen
>>>on this group
>>
>> Wait, what?
>
> Perhaps this is your evil bearless twin
I love that concept. An identical twin, undetectable from the original,
except ... without a bear.
David Fisher
quinti...@gmail.com
> In article <1191419422.677873.47...@50g2000hsm.googlegroups.com>,
> Wait, what?
>
> Adam
Well I based it on you're being more willing to dialog with Jacek than
anyone else I've seen on the group (excluding Jacek's puppet accounts,
of course). You've also been willing to attempt a rational discourse
with Breslin when most people, myself included, have long ago given it
up as futile.
I meant no offense. :)
Emily Short
wrote:
> > -- exploration
> ...
> > The interesting question is how to make the tangential
> > content an important part of the whole, even if the player
> > only experiences part of it, and without making the
> > player feel *obliged* to experience all of it.
>
> It sounds pretty tricky to let the player know that the tangential content
> contains no important information without saying so explicitly. Maybe a clue
> could be dropped in-game - if the player keeps on pursuing the non-essential
> information, then the game/narrator or an NPC could say something about it
> being unimportant.
Well -- what I'm thinking of are things like databases of material the
player can look up, where there's no puzzle necessity to look up every
single entry. I think the trick here actually is to make it clear
relatively early on that there is a huge amount implemented and so the
player should not try to follow up every keyword he can think of (or
explore every drawer, or whatever), but should instead only pursue
things that either interest him personally or seem likely to advance
his progress towards a goal.
That requires that goals be clearly articulated and that progress
toward them be fairly sensible.
> Sometimes I wish I could ask the game a direct question about this sort of
> thing. Wasn't there a game with a "red herring detector"?
"First Things First".
> > I see the lack of challenge and exploration as likely
> > problems with the storytelling universe Chris Crawford
> > wants to create, which is All Choices, All The Time.
>
> I still love the idea of having lots of freedom of choice, but then
> constraining the plot to follow up on the player's decisions. If the player
> is discouraged from or unable to keep on changing the direction of the plot
> once they have made a clear choice, then it won't feel so directionless
> (which is how I understand your objection).
Well, no, that's not the sum of my objection. Part of my objection is
that if the bulk of interaction takes the form of choice-making, the
player tends to feel less invested in the individual choices and
perhaps unprepared to make them at all. We've seen this in various
forms; several players said both of Floatpoint and of Fate that they
felt like they needed more time to explore, get to know the territory,
and become familiar with the other characters, before being asked to
make significant moral decisions.
> I would love to have a significant
> choice about how I relate to each NPC in the game (within the general
> constraints of the PC's personality, if the PC is well defined),
Sure.
> and to have
> lots of different ways of solving puzzles (eg. if there is a forest, being
> able to find some leaves or a stick if I need them, or chop down a tree to
> block the road - or at least try to)
I am all for multiple solutions and trying to accommodate all the
puzzle approaches that the player character is rationally likely to
try.
Both of these are things I tried to do with "City of Secrets" -- you
can solve almost all of the puzzles (such as they are) in multiple
ways, you can make friends or enemies of various characters, etc.
(There were a few constraints necessitated by the plot I was told to
work with, but not *that* many.) So I'm not unsympathetic to what you
describe, as far as you've described it here.
The problem is -- and this becomes visible in the midgame of "City of
Secrets" -- that if you give the player too many options about what he
can do, he may feel unguided, get bored, and give up. I left a lot of
territory open to explore there and let the player decide which
characters he wanted to trust and which he wanted to work against; but
because not everyone clearly understood what the options were and how
to pursue them, many people found this unsatisfying.
> I sometimes feel you are a bit dismissive of the idea of being able to "do
> anything" in a game. Without the plot being managed properly, I agree it
> could be boring and directionless; but if it included tracking and following
> up on the player's choices so far in the game, I think it would be
> fantastic.
You've put quite a lot of constraints on your definition of "do
anything", though: if the player has to pursue one of a set of goals
that the author has chosen, that's already quite different from a
scenario in which the player is allowed to wander around and try
various stuff and the game world just naturally adjusts around all his
behavior to make interesting results come of it, whatever it might be.
In that extreme a form I haven't seen many implementation attempts, of
course, but I have seen people come onto RAIF and describe such a game
as their ultimate ideal; to which my response is, "that's not a game:
it's your life."
Sabahattin Gucukoglu
>
>> In article <1191419422....@50g2000hsm.googlegroups.com>,
>> <quinti...@gmail.com> wrote:
>>>Adam Thornton -- probably the most patient person I've seen
>>>on this group
>>
>> Wait, what?
>
> Perhaps this is your evil bearless twin, a man of patience and
> forbearance who is unacquainted with Stiffy Makane or TPB.
What's TPB, then?
Cheers,
Sabahattin
--
Please DO NOT reply to sender.
Sabahattin Gucukoglu <mail<at>sabahattin<dash>gucukoglu<dot>com>
Address harvesters, snag this: fee...@yamta.org
Phone: +44 20 88008915
Mobile: +44 7986 053399
http://sabahattin-gucukoglu.com/
Adam Thornton
<quinti...@gmail.com> wrote:
>On Oct 3, 10:17 am, a...@fsf.net (Adam Thornton) wrote:
>> In article <1191419422.677873.47...@50g2000hsm.googlegroups.com>,
>> Wait, what?
>anyone else I've seen on the group (excluding Jacek's puppet accounts,
>of course). You've also been willing to attempt a rational discourse
>with Breslin when most people, myself included, have long ago given it
>up as futile.
>
>I meant no offense. :)
None taken, I suppose. But you might want to guard your Cheerios for a
while, just in case.
Aadam
Quintin Stone
> None taken, I suppose. But you might want to guard your Cheerios for a
> while, just in case.
>
> Aadam
Duly noted.