some clarifying questions about Nick Montfort's IF theory essay

30 views
Skip to first unread message

Jacek Pudlo

unread,
Apr 7, 2007, 10:19:15 PM4/7/07
to
Nick Montfort writes in http://nickm.com/if/toward.html

"To build a theory of interactive fiction that is useful in deeply
understanding how interactive fiction is experienced, and how better sorts
of works can be created, a stronger approach than that of the theory-bag is
necessary, one which distinguishes those elements of interactive fiction
that result from it being (1) a text-accepting, text-generating computer
program; (2) a potential narrative, that is, a system which produces
narrative during interaction; (3) a simulation of an environment or world;
and (4) a structure of rules within which an outcome is sought, also known
as a game."


(1) When I hear the phrase "text-generating," I cringe slightly and think
"Markov bot." This strikes me as an unfair association, since bots of this
kind have little or no understanding of semantics and can therefore generate
nothing but jibberish, or semi-parrot the input. No work of interactive
fiction that I am aware of *generates* text. Instead, works of IF output
*pre-written* text. The text is not created shortly before it is presented
to the player -- it is already in existence waiting, so to speak, for the
player to trigger the condition for its output.

(2) I have a similar problem with "produces narrative during interaction."
Unless Mr. Montfort is talking about Mr. Crawford's infamous Erasmatron, I'm
not sure how "produces" applies to an IF narrative. Surely the plot-line(s)
is(are) already in place, waiting to be triggered, just like the pre-written
text.

(3) How does "simulation of an environment or world" apply to, say,
_Photopia_ or _Rameses_? Yes, both works present the player with a world,
but is it productive to speak of them in terms of simulation? In one of the
more stylistically elevated paragraphs, Mr. Montfort fancies himself a
surgeon deftly applying instruments of precision to the body of IF. What
does the concept of simulation yield when surgically applied to _Photopia_?

(4) Continuing the medical metaphor, is the game component a necessary part
of the anatomy of IF? If not -- as Mr. Montfort suggests later in the
essay -- perhaps this should have been clarified in the introduction. Also,
"a structure of rules within which an outcome is sought" describes just
about any human endeavour, not just the ones referred to as "games." Even
so, "Game" is a hopelessly inclusive concept that "may be good enough for a
butcher," but is hardly befitting a surgeon.


Jeff Nyman

unread,
Apr 8, 2007, 11:42:09 AM4/8/07
to
"Jacek Pudlo" <ja...@jacek.jacek> wrote in message
news:DiYRh.38058$E02....@newsb.telia.net...

> Nick Montfort writes in http://nickm.com/if/toward.html
>
> (2) I have a similar problem with "produces narrative during interaction."
> Unless Mr. Montfort is talking about Mr. Crawford's infamous Erasmatron,
> I'm
> not sure how "produces" applies to an IF narrative. Surely the
> plot-line(s)
> is(are) already in place, waiting to be triggered, just like the
> pre-written
> text.

I suppose I can see "potential narrative" from the *game player's* point of
view because while the narrative is "built in", they don't know what it
is -- at least during a first pass-through of the game. Further, some
narrative paths might be closed to them because of actions they take (or
don't take). As such, the actual narrative they experience may not be the
"full" narrative. So the player explores the game and via their interaction
with the game world they do produce some aspect of the narrative that's
built in.

This notion of "generated narrative" seems to be based on a line in the
article: "The distinction between what can be simulated and what can be
narrated is particularly important to understanding the workings of
interactive fiction." I'm not convined one way or the other and the article
does nothing to really operationally define this distinction in a way that I
find testable.

There's another line that talks about "the interactor [having] no
opportunity to influence the course of the narrative that is being
produced." But then the article says: "In the sense that scholars of the
story and of narrative (that is, narratologists) use the terms, a work of IF
is not a narrative." Fine, I guess, if you're concerned what scholars of
story and narrative have to say. (That said, I would argue that a case could
be made that a text-based interactive fiction game can, in fact, be a
narrative in the sense that these scholars describe.)

Then the article says: "An IF work is an interactive computer program, but
not directly a narrative..." I find this to be similar to a statement like
"a novel is a book, but not directly a narrative." My point is that I'm not
sure what I'm really learning from this.

> (3) How does "simulation of an environment or world" apply to, say,
> _Photopia_ or _Rameses_? Yes, both works present the player with a world,
> but is it productive to speak of them in terms of simulation? In one of
> the
> more stylistically elevated paragraphs, Mr. Montfort fancies himself a
> surgeon deftly applying instruments of precision to the body of IF. What
> does the concept of simulation yield when surgically applied to
> _Photopia_?

I would think it does make sense to call them "simulation" because you are
simulating a world. In Rameses, if I remember right, you're a college kid or
something along those lines whose basically lacking some backbone and goes
along with the flow. So your simulating an environmental situation or a
person who responds in a certain way to an environment. In Photopia you're
simulating the actions of a person through these various ... what? dream
states? time zones?

My point is that simulation is, defined in one sense, as the representation
of the behavior or characteristics of one system through the use of another
system. This, to me, doesn't just have to be a simulation of a "world model"
(like a world where magic works) but can also be a simulation of inner
states or a simulation of how we think about things or a simulation of how
we might perceive a given world.

But, see, I guess this is where for me it becomes a question of: is this
much thought really all that necessary?

Maybe that's your point, too? I'm not sure but in reading the article you
cite it reminds me a lot of the "theories" and "models" and "formal logic"
that are often put forward for the testing and quality assurance
disciplines. It's not so much that I think it's wrong to do so, but rarely
have I found it terribly helpful or enlightening in the day-to-day practice
of performing testing or assuring quality. Likewise with actually writing
text-based interactive fiction games or playing them.

In reading the article, I'm struck by the use of terms like "extradiegetic",
"diegetic", etc. Yes, I know what they mean. But I'm reminded of what Robert
Alter said in his "The Art of Biblical Narrative": "I am particularly
suspicious of the value of elaborate taxonomies and skeptical as to whether
our understanding of narrative is really advanced by the deployment of
bristling neologisms like analepsis, intradiegetic, actantial."

To be fair, I gave the article a full (but somewhat quick read). I need to
go through it again with a more critical eye but these were my initial
thoughts.

- Jeff


signw...@lem.signwriter.name

unread,
Apr 8, 2007, 1:54:57 PM4/8/07
to
On 8 Apr, 16:42, "Jeff Nyman" <jeffny...@gmail.com> wrote:
> But, see, I guess this is where for me it becomes a question of: is this
> much thought really all that necessary?
>
> Maybe that's your point, too? I'm not sure but in reading the article you
> cite it reminds me a lot of the "theories" and "models" and "formal logic"
> that are often put forward for the testing and quality assurance
> disciplines. It's not so much that I think it's wrong to do so, but rarely
> have I found it terribly helpful or enlightening in the day-to-day practice
> of performing testing or assuring quality. Likewise with actually writing
> text-based interactive fiction games or playing them.
>
> In reading the article, I'm struck by the use of terms like "extradiegetic",
> "diegetic", etc. Yes, I know what they mean. But I'm reminded of what Robert
> Alter said in his "The Art of Biblical Narrative": "I am particularly
> suspicious of the value of elaborate taxonomies and skeptical as to whether
> our understanding of narrative is really advanced by the deployment of
> bristling neologisms like analepsis, intradiegetic, actantial."

I haven't read the essay, but my family bought me a copy of 'Twisty
Little Passages' from a library sale last week, which I believe
contains a revised and expanded version of the same material.

Sure, in most contexts the verbiage might be overkill, but my
understanding is that the book and the earlier essay were an attempt
to introduce a theory of IF and terminology for discussing it
*specifically* within an academic context. Try telling academia that
exactingly-defined jargon is unnecessary :)

Jeff Nyman

unread,
Apr 8, 2007, 2:20:35 PM4/8/07
to
<signw...@lem.signwriter.name> wrote in message
news:1176054897.0...@n59g2000hsh.googlegroups.com...

>
> Sure, in most contexts the verbiage might be overkill, but my
> understanding is that the book and the earlier essay were an attempt
> to introduce a theory of IF and terminology for discussing it
> *specifically* within an academic context. Try telling academia that
> exactingly-defined jargon is unnecessary :)

You're correct on that last point. :) Having had to write papers where "time
travel" had to be referred to as "global causality violation" or where
"extra dimensions" had to be referred to as "scaled M-Theory branes", I can
understand how that works. But the article claims it can also help to more
"deeply understand" IF and to "write" better IF, implying (the hope for) a
reach beyond academia, unless the assumption is that only academics write
interactive fiction. The last part of the article even states that all this
discussion "can result in a better understanding of interactive fiction and
should also allow the many of us who author IF works to make advances in our
art."

So, I gave the article a more thorough parsing.

My thoughts: What's meant here is a hypothesis, not a theory. I don't see
this as a theory at all. An article that's so careful with it's terminology
would presumably know the difference between theory and hypothesis. That
being said, it's not even really a hypothesis. At no time has the
"hypothesis" been given operational specificity. That is, the hypothesis
isn't stated in such a way as to permit testing by observing certain
characteristics that are produced independently of the hypothesis, and thus
to permit assessment of whether it's confirmed or disproved. There's not
even a null hypothesis stated, from what I can see.

I think what you have here is a heuristic fiction: an "IF Model" of sorts.
Heuristic fictions, unlike hypotheses, aren't evaluated by whether they're
proved or disproved, but rather by their usefulness. According to the
article this usefulness can be measured in terms of how well the model is

"useful in deeply understanding how interactive fiction is experienced, and

how better sorts of works can be created."

= = = = = = = = = = = = = =
Read the below only if you're interested in how I came to my conclusion.
= = = = = = = = = = = = = =

I had to consider what the essay was supposedly all about, so I could
understand what theory (well, hypothesis) was being proposed. To wit:

"This essay aims to begin some useful discussion of the elements of
interactive fiction from a theoretical standpoint." (stated in the
Conclusion).

"To build a theory of interactive fiction that is useful in deeply
understanding how interactive fiction is experienced, and how better sorts

of works can be created..." (stated in Theorizing Interactive Fiction)

"...a theory that distinguishes formal aspects from those related to
interpretation..." (stated in Theorizing Interactive Fiction)

"...narratology can inform a rigorous, formal theory of interactive fiction,
a theory that remains sensitive to the many-faceted nature of this new media
form." (stated in Theorizing Interactive Fiction)

*******

Some operational definitions are provided.

"A work of interactive fiction is, among other things, a computer program
that accepts text input from a user and produces text output in reply."
(stated in Interactive Fiction and the Interactor) Monfort leaves "among
other things" undefined here and that's important because he at one point
takes Roger Carbol to task for a definition that "does not distinguish games
from non-games, as any definition should." Well, neither does his definition
above ... except for the vague "among other things."

"Using text more specifically, to mean 'strings of words,' interactive
fiction indicates a category of text-based works, works that can contain
other media elements but where text and textual exchange are central."
(again stated, Interactive Fiction and the Interactor). I agree with this in
that he's allowing for text-based games that do have graphic and sound
components but where text is still central. (That said, what about
graphics-based interactive fiction such as the Monkey Island games or the
"Broken Sword" series? Montfort was saying that "text can also be considered
semiotically to be any set of signifiers" and that can include spoken text
in such games or even the text subtitles that are used in such games. So,
again, I don't think he operationally defines himself as well as he thinks
he does.)

*******

The article in the section "Interactive Fiction and the Interactor" talks
about how an article by Emily Short ("What's IF?") does not apply to a work
like "The Space Under the Window." Why? Because Short's article appears to
say: "IF *tends* to represent, in some form, an environment or imagined
world whose physical space we can explore." Montfort says that "Space Under
the Window" is a work of hypertext rather than interactive fiction. I
disagree in that the game does present a world. The interaction with the
world may be "different" from the "standard" way of communicating with such
a world, but that doesn't mean a world isn't presented. It is. Montfort
says: "Whether a work simulates a world or not can be determined from
outside, by an interactor studying the work." I agree. As the "interactor"
(player, for those who like to keep it simple) of this game, I found it a
simulation of a world. A limited simulation, yes, but a simulation
nonetheless.

In "Interactive Fiction and the Interactor", Montfort talks about the
distinction of "work of IF" and "game of IF." The problem is that this is
not operationally defined in any useful way. Monfort seems to suggest that
for IF to be called a "game" it would have to have some "quest or intrigue"
or some "final reply that is a 'winning' one" or just any sort of "final
reply at all." But, consider that a game can be "an amusement or pastime."
There's nothing in there that says it *must* have a quest or intrigue or any
sort of final reply. (Montfort seems to agree, I guess. Later he says: "A
final reply is not required for a work to be interactive fiction, and some
works, by design, do not produce a final reply." So I'm not sure what the
point here was at this point in his article.)

Montfort, on this topic uses Ian Finley's game "Exhibition" as an example.
He describes it as a "simulated space" where a player (interactor) can chat
with other characters about a work of art. He says that there's "no way to
win or lose [the game]." Then: "calling this a 'game' is unfair to
Exhibition, which is not actually a game." What is it then? He doesn't say.
I guess we're just supposed to think it's a "work", a term he hasn't clearly
defined in any sort of operational sense, at least one that makes sense in
terms of what a game means. His comparison to "Hollywood Hijinks: is
interesting though in that he says this latter game "simulates a treasure
hunt in a mansion and has a very definite an explicit goal."

This section ends by saying: "'Work' has real advantages as a term, however,
in discussions where we are trying to be as clear and precise as possible,
rather than simply using those terms which have become customary." I find
that ironic because I found this section to be anything but "clear and
precise" in telling me why I should have worried so much about this
distinction or why it would apply to an effective "theory" of IF, either in
terms of "deeply understanding" it or in helping me produce a better
game/work of it.

*******

In "Cycles, Exchanges, and the IF World" Monfort says: "a work of IF is not
a narrative." He says this relative to how scholars use the term and then he
quotes one: "the representation of real or fictive events and situations in
a time sequence." Let's be simple (but accurate) about narrative. It's "a
story or account of events, experiences, or the like, whether true or
fictitious." In both senses, however, I can see how text-based interactive
fiction is a narrative.

Now, granted, Montfort seems to be talking about interactive fiction as
"producing" narratives as opposed to "being" a narrative. Perhaps. But to a
player, does that distinction matter? Is this part of what he meant when he
said "deeply understanding how interactive fiction is experienced"? If so,
I'm curious how many players agree (or care one way or the other).

Montfort also says: "Similarly, interactive fiction is not a story in the
sense of the things that happen in a narrative..." A story, being simple
about it, is "the plot or succession of incidents" of some work of fiction.
Of course, Montfort goes for the scholarly definition again and presents
one: "the content plane of narrative as opposed to its expression or
discourse; the 'what' of a narrative as opposed to its 'how'" I don't know
if this is leading to "deeply understanding" anything or how this will help
produce better works. "Clarifying" all this, Montfort says: "An IF work is
always related to story and narrative in their narratological sense, even if
a particular work does not have a "story" in this ordinary sense." Note that
none of this is presented with any sort of backing whatsoever, not even
references.

Commands that the player issues to game/work of IF are said to be
"diegetic", meaning at the level of the characters, as it applies to their
thoughts and actions. Actually, you could argue they're not, though, because
they're issued by the player -- not the player character. If anything,
they're somewhat metadiegetic (or hypodiegetic), which means that part of
the diegesis (story) that's embedded in another one (and often understood as
a story within a story), as when a diegetic narrator themselves a story. So
here the player is "narrating" a story in the sense of generating
narratives. (In this sense, the game itself -- the programming behind it,
that is -- would be the extradiegetic level, the level of the narrative's
telling.)

Montfort says that "[inputs] such as those that save, restore, quit,
restart, change the level of detail in the room descriptions, or address
some entity that is not part of the IF world -- to ask for hints, for
instance -- are directives. A directive is, in narratological terms,
extradiegetic." Personally, I think this could still be thought of at the
metadiegetic level. Of course, in reality, it would probably be simpler to
just say such things are "non-diegetic."

Confused yet? All I'm trying to show is that you can argue a lot of
different ways with these terms, none of them really enlightening in my
opinion. Further, the goal was to operationally define these concepts. How
Montfort worded it is that "other theorists" should be able to apply the
terms in the same way. Well, I've just shown that this is, at the very
least, problematic with how the "theory" is being put forth.

*******

In "Player Characters, Non-Player Characters, and Other Persons", I think
there's at least a valid distinction here: "The difference is that a
character in interactive fiction must be an existent who acts within the IF
world. Being a part of the simulation, rather than being a part of the story
that the generated narrative tells, is essential for a character in
interactive fiction." As he later says of such characters, if they can be
called this: "They have a sort of existence within the IF world, but there
is no opportunity to interact with them."

Montfort attaches importance to this distinction. He says: "An opportunity
for the interactor's input to influence the behavior of a person -- not
simply to cue an appearance -- would seem to be important in designating
this person an NPC." But why is that important either to the author or
player? It's not, as far as I can see. How does that matter in terms of
deeply understanding or creating such works? The article doesn't say, even
though that's its stated intention in terms of driving towards this "theory
of IF."

I also think this "type" of character concept is interesting, however, in
terms of something he says in the next section (World, Rooms): "Rooms, like
characters, are simulated and are part of the IF world; they are not just
mentioned in some of the narrations that are produced." But, in many cases,
they are just mentioned: again from the player and author point of view. So,
in both cases, I'm not sure what this has bought us in terms of
understanding.

*******

The "Conclusion" states that "distinguishing the simulated from the
non-simulated, the IF world from the text that describes it, and the
diegetic from the extradiegetic" was part of this whole deal. The goal? To
"demonstrate that a better perspective on IF can result from making such
distinctions."

"Better perspective" from whose viewpoint? "Better" compared to what?
"Better" defined how? In terms of playing games? Writing them? Understanding
them, whether or not you write or play them?

Nothing is really clarified. We hear a lot about interactors, cycles,
traversals, courses, exchanges, transgressions, interactions, sessions,
directives: absolutely none of it meaningful (at least to me) or indicative
of "deeply experiencing" a game (work, if you prefer) of text-based
interactive fiction nor of how to create a better such game.

There's information here, no doubt. But a hypothesis? Not that I can see. A
theory? Definitely not. To be fair, his Conclusion does indicate this
somewhat, indicating that he's moving "toward" a theory, so I'm perhaps
being a little harsh. But, that said, there's really nothing indicated that
suggests a theory is being moved toward since, again, I can't see anything
testable toward a theory. It seems the emphasis is on whether or not people
will accept the terms.

He says: "The first test of the more formal concepts defined here will be if
they are understood and can be applied unambiguously, in a consistent way,
by different theorists. The only way of knowing whether this can be done or
not, and what the failings of the concepts presented here are, will be in
another theorist taking up these terms in discussion of other specific IF
works."

That's not a theory *of* interactive fiction, however. That's just a
dictionary for *talking about* interactive fiction. I think what the article
has done, taking away some of the jargon, is provide a relatively useful
start of a model for considering some aspects of text-based interactive
fiction that could potentially be used to operationally define such fiction
from other games and other types of fiction.

- Jeff


Jacek Pudlo

unread,
Apr 8, 2007, 6:17:29 PM4/8/07
to
Jeff Nyman

[...]

> Now, granted, Montfort seems to be talking about interactive fiction as
> "producing" narratives as opposed to "being" a narrative. Perhaps. But to
a
> player, does that distinction matter?

It does matter. If the narrative is *produced* through interaction,
authorial control is usurped by the player. It is my belief that this is not
the case. We simply don't have the technology to make it happen. I believe
the author retains full control over the narrative, even in a game like
_Slouching Towards Bedlam_, with its multiple plotlines. Through interaction
the player merely *unlocks* already existing, pre-fabricated, plotlines. The
word "produce" is abysmally ill-chosen. I've never felt that I've been
"producing" (co-authoring?) anything while playing IF.

I agree with your assessment, Jeff. Monfort's essay is neither a theory nor
a thesis, and its terminology is sloppy and misleading.

[...]

>
> Commands that the player issues to game/work of IF are said to be
> "diegetic", meaning at the level of the characters, as it applies to their
> thoughts and actions. Actually, you could argue they're not, though,
because
> they're issued by the player -- not the player character. If anything,
> they're somewhat metadiegetic (or hypodiegetic), which means that part of
> the diegesis (story) that's embedded in another one (and often understood
as
> a story within a story), as when a diegetic narrator themselves a story.
So
> here the player is "narrating" a story in the sense of generating
> narratives. (In this sense, the game itself -- the programming behind it,
> that is -- would be the extradiegetic level, the level of the narrative's
> telling.)

> Montfort says that "[inputs] such as those that save, restore, quit,
> restart, change the level of detail in the room descriptions, or address
> some entity that is not part of the IF world -- to ask for hints, for
> instance -- are directives. A directive is, in narratological terms,
> extradiegetic." Personally, I think this could still be thought of at the
> metadiegetic level. Of course, in reality, it would probably be simpler to
> just say such things are "non-diegetic."

This is not how "extradiegetic" is used in narratology. "Extradiegetic"
usually refers to characters who exist outside of the text. If you write a
story about George W. Bush being abducted by space aliens you create an
extradiegetic character. This is how Mieke Bal (_Introduction to the Theory
of Narrative_ p. 118) and Gerard Genette (_Narrative Discourse_ pp. 226-37)
use the term. What Montfort is referring to with his "extradiegetic
directives" is the static fiction equivalent of opening a book and turning
its pages. He has misapplied a narratological term to describe the handling
of the physical manifestation of a narrative. Narratology doesn't concern
itself with the physical handling of books because even the silliest
academic isn't silly enough to invent fancy Greek-sounding nomenclature for
the act of turning a page.

This is complicated by the fact that QUIT and SAVE and RESTORE can be made
diegetic. _Slouching Towards Bedlam_ is a case in point. Closing a book can
never be made part of the story, but QUIT can. So Montfort may be fumbling,
but he has fumbled -- unwittingly -- onto an interesting difference between
static and interactive fiction.

signw...@lem.signwriter.name

unread,
Apr 8, 2007, 7:17:50 PM4/8/07
to
Well, that was intimidatingly well-thought-out.... :)

> Now, granted, Montfort seems to be talking about interactive fiction as
> "producing" narratives as opposed to "being" a narrative. Perhaps. But to a
> player, does that distinction matter? Is this part of what he meant when he
> said "deeply understanding how interactive fiction is experienced"? If so,
> I'm curious how many players agree (or care one way or the other).

Hm. Could "producing" vs. "being" a narrative be precisely the
distinction between interactive and static fiction? A work of IF is a
sort-of quantum superposition of possible narratives: any *particular*
narrative has to be produced by the interactions of a player.

People do complain about works of IF being "too linear". Perhaps we
can consider that synonymous with "insufficiently productive"? Whether
that gains us anything, I don't know.

> Commands that the player issues to game/work of IF are said to be
> "diegetic", meaning at the level of the characters, as it applies to their
> thoughts and actions. Actually, you could argue they're not, though, because
> they're issued by the player -- not the player character. If anything,
> they're somewhat metadiegetic (or hypodiegetic), which means that part of
> the diegesis (story) that's embedded in another one (and often understood as
> a story within a story), as when a diegetic narrator themselves a story. So
> here the player is "narrating" a story in the sense of generating
> narratives. (In this sense, the game itself -- the programming behind it,
> that is -- would be the extradiegetic level, the level of the narrative's
> telling.)
>
> Montfort says that "[inputs] such as those that save, restore, quit,
> restart, change the level of detail in the room descriptions, or address
> some entity that is not part of the IF world -- to ask for hints, for
> instance -- are directives. A directive is, in narratological terms,
> extradiegetic." Personally, I think this could still be thought of at the
> metadiegetic level. Of course, in reality, it would probably be simpler to
> just say such things are "non-diegetic."
>
> Confused yet?

Um, yes? :) Actually, no, I think I'm following you....

> All I'm trying to show is that you can argue a lot of
> different ways with these terms, none of them really enlightening in my
> opinion. Further, the goal was to operationally define these concepts. How
> Montfort worded it is that "other theorists" should be able to apply the
> terms in the same way. Well, I've just shown that this is, at the very
> least, problematic with how the "theory" is being put forth.

Well, in this case the "other theorist" argues that Montford is
erroneous in classifying commands that the player issues to game/work
of IF as diegetic, and that they are better considered as
metadiegetic.

Discourse for the win!

> In "Player Characters, Non-Player Characters, and Other Persons", I think
> there's at least a valid distinction here: "The difference is that a
> character in interactive fiction must be an existent who acts within the IF
> world. Being a part of the simulation, rather than being a part of the story
> that the generated narrative tells, is essential for a character in
> interactive fiction." As he later says of such characters, if they can be
> called this: "They have a sort of existence within the IF world, but there
> is no opportunity to interact with them."
>
> Montfort attaches importance to this distinction. He says: "An opportunity
> for the interactor's input to influence the behavior of a person -- not
> simply to cue an appearance -- would seem to be important in designating
> this person an NPC." But why is that important either to the author or
> player? It's not, as far as I can see.

If I've understood this correctly, an "other person" is basically the
same as a character in a work of static fiction, yes? Mentioned
entirely in the text. Interactivity is a prerequisite for Montford to
consider a character an NPC.

Agreed, it may not matter to a player, but an author deliberately
structuring an IF around NPC interaction may have to make creatively-
motivated decisions about NPC implementation, distinct from the design
decisions an author might make about an "other person"'s appearances
in static text, even static text within the same IF.

This one, I think, may have something to it. Maybe not a great deal,
but *something*.

> I also think this "type" of character concept is interesting, however, in
> terms of something he says in the next section (World, Rooms): "Rooms, like
> characters, are simulated and are part of the IF world; they are not just
> mentioned in some of the narrations that are produced." But, in many cases,
> they are just mentioned: again from the player and author point of view. So,
> in both cases, I'm not sure what this has bought us in terms of
> understanding.

I think the point is, again, implementation: you could implement a
particularly boring room as nothing more than a line of static text:

"Walking north from Room #1, you pass through three formal gardens and
the entrance hall before reaching

*Room #2*

It's dark. You can't see anything."

Room #2, even if all you can ultimately do there is leave, is
simulated - *implemented*. It has an existence beyond being mentioned
in the text. Again, even if it makes no difference to the player -
which in this case, I think you can argue it does - it surely does to
the author?

> I think what the article
> has done, taking away some of the jargon, is provide a relatively useful
> start of a model for considering some aspects of text-based interactive
> fiction that could potentially be used to operationally define such fiction
> from other games and other types of fiction.

I'd say that it did so, in part, by *providing* the jargon, but I'll
grant you the point wholesale.

Jacek Pudlo

unread,
Apr 8, 2007, 7:31:02 PM4/8/07
to
<signw...@lem.signwriter.name>

> Well, that was intimidatingly well-thought-out.... :)
>
> > Now, granted, Montfort seems to be talking about interactive fiction as
> > "producing" narratives as opposed to "being" a narrative. Perhaps. But
to a
> > player, does that distinction matter? Is this part of what he meant when
he
> > said "deeply understanding how interactive fiction is experienced"? If
so,
> > I'm curious how many players agree (or care one way or the other).
>
> Hm. Could "producing" vs. "being" a narrative be precisely the
> distinction between interactive and static fiction? A work of IF is a
> sort-of quantum superposition of possible narratives: any *particular*
> narrative has to be produced by the interactions of a player.
>
> People do complain about works of IF being "too linear". Perhaps we
> can consider that synonymous with "insufficiently productive"? Whether
> that gains us anything, I don't know.

Let's bring this down to earth. _Slouching Towards Bedlam_ offers a plot
with five variations. Each plot variation is pre-fabricated -- i.e. it
exists before the player opens the game file. No matter how much the player
fiddles with the game, he will *never* reach a sixth variation. Please
explain what the player is "producing."

signw...@lem.signwriter.name

unread,
Apr 8, 2007, 8:04:12 PM4/8/07
to
On 9 Apr, 00:31, "Jacek Pudlo" <j...@jacek.jacek> wrote:
> <signwri...@lem.signwriter.name>

If I've read Montford correctly, a narrative following one of those
five prefabricated plot variations, the precise text of which includes
"Bedlam"'s responses to everything that particular player does,
whether or not it advances the game along one of those prefabricated
strands.

"Produced", you see? True, it's from prefabricated parts, but any
given playthough can produce a text which differs from any other given
playthrough.

Don't look at me, *I'm* not the theorist here.

steve....@gmail.com

unread,
Apr 8, 2007, 9:13:31 PM4/8/07
to
Yeah, that's a bad paragraph. -- It has the resonance of that moment
when the "master" of disguise loses his false mustache in his glass of
water.

But let's be fair to Nick: this was an earlier essay, written back
when he suffered a confusion common in grad students -- a false belief
that academic-sounding thought is a means to intelligent thought.
(And, yes, a false hope that the disguise is actually working.)

So while we will still look forward to hearing such definitions as "a
structure of rules within which an outcome is sought" from the likes
of such terminally academic-wannabees as Graham Nelson and the dogs
who follow such leads, by this point such grammar and phraseology are
well beneath the likes of Nick.

This is not to say that this essay is some error better forgotten! --
Even this borderline juvenalia is well worth reading. Of course an
attentive reading uncovers the cliche and finds its vacancy, but
behind that cliche one can already see the driving concepts. Even in
this relatively early work, pretension and its desire for concealment
have not destroyed the larger ideas. Nick is an academic in the right
way: his work is begun and I hope will always evolve beyond the base
grammar and mere form of academic thought.

Just take the first objection: "text-accepting, text-generating
computer program" -- yes a false generalization aimed at profundity,
but it's more. (Alright saying that it accepts text, yes that's just a
duh thing to say. But the other part.) Addressing one's will to the
issue of IF's text generation is worth the effort. There's lots of
modes in which it generates text: from auto-generated default
responses to report-grouping to disambiguation questions to state-
variant conversation to merely printing room descriptions. It's
obviously just the beginning of the thought, but it's a good thought
and a good beginning, despite being painfully doctorate.

Adam Thornton

unread,
Apr 9, 2007, 2:52:51 PM4/9/07
to
In article <ZRdSh.38084$E02....@newsb.telia.net>,

Jacek Pudlo <ja...@jacek.jacek> wrote:
>It does matter. If the narrative is *produced* through interaction,
>authorial control is usurped by the player. It is my belief that this is not
>the case. We simply don't have the technology to make it happen.

We certainly *do* have the technology to make it happen, but the problem
is that the thing that results will be no fun at all.

One can certainly imagine an open-ended sim game, in which you wander
around a landscape that reacts to you. Unfortunately, what you end up
with is basically a narrative-interface roguelike. Much as it pains me,
I think I would be forced to agree with Chris Crawford here: "that's not
a story!"

In fact, the Maze section of _Reliques of Tolti-Aph_--if it were shorn
of all its unique rooms and talismanic items, and were the entirety of
the game, rather than being an obstacle in the service of a larger
narrative--would be an excellent example of this. For that matter, so
is playing "Stochastic AD&D" where you turn to the Random Dungeon
Creator in the back of the first edition DMG, and roll up random rooms,
populate them with random monsters, and have random fights with them.

We clearly have the technology to do this. We don't do it, not because
it's hard, but because it's really boring. This is of course my old
tired hobbyhorse of simulation versus narrative; rather than rehash it
here, I simply invite the interested reader to take it up with my
simulation of Chris Crawford in the holodeck scene of _Stiffy Makane:
The Undiscovered Country_.

>even the silliest
>academic isn't silly enough to invent fancy Greek-sounding nomenclature for
>the act of turning a page.

I doubt this. I do not, however, have an example at hand.

>This is complicated by the fact that QUIT and SAVE and RESTORE can be made
>diegetic. _Slouching Towards Bedlam_ is a case in point. Closing a book can
>never be made part of the story

...mmmmaybe. I bet Italo Calvino, or someone like him, has *tried*.

Adam

Richard Bos

unread,
Apr 9, 2007, 7:18:04 PM4/9/07
to
ad...@fsf.net (Adam Thornton) wrote:

> Jacek Pudlo <ja...@jacek.jacek> wrote:
> >This is complicated by the fact that QUIT and SAVE and RESTORE can be made
> >diegetic. _Slouching Towards Bedlam_ is a case in point. Closing a book can
> >never be made part of the story
>
> ...mmmmaybe. I bet Italo Calvino, or someone like him, has *tried*.

Doug Hofstadter, Goedel Escher Bach, the dialog where, amongst others,
the (non-)end of (IIRC) the Wohl-Temperierte Clavier.

Richard

steve....@gmail.com

unread,
Apr 10, 2007, 8:33:16 AM4/10/07
to
Adam wrote:
> >Closing a book can
> >never be made part of the story
>
> ...mmmmaybe. I bet Italo Calvino, or someone like him, has *tried*.

Walter Benjamin, for one example, does this sort of thing all the
time; it's a vein of the whole mystic tradition. It's a 19th (and
20th) Century extension of the allegorical topology perfected by
Bunyan.

Jacek Pudlo

unread,
Apr 11, 2007, 5:15:03 AM4/11/07
to
<steve....@gmail.com>

[...]

> There's lots of
> modes in which [a work of IF] generates text: from auto-generated default


> responses to report-grouping to disambiguation questions to state-
> variant conversation

A program generates text by applying syntantical rules to a set of
glossaries. This is most certainly not how text is produced in IF. The vast
majority of the output consists of complete pre-written sentences. Yes,
there are occassions when the program outputs things like "The [objectname]
[is/are] out of reach." but these are just a tiny fraction of the overall
output. The text-generated portion of the output is the one which is
narratively and artistically the least interesting. Do you play IF for the
thrill and pleasure of reading auto-generated text?

> to merely printing room descriptions.

A number of conditions must be met for a room description to be printed, but
that doesn't make it text-generated. The syntax, the prose style, the
metaphors, the similes, the allusions, the hints -- all these things that
you may find in a room description have been "generated" by a human being,
not a computer program. Let us not denigrate _Anchorhead_, or any other work
of IF, by falsely suggesting that the prose is text-generated.

> It's
> obviously just the beginning of the thought, but it's a good thought
> and a good beginning

No.


Jacek Pudlo

unread,
Apr 11, 2007, 5:33:20 AM4/11/07
to
"Adam Thornton"

> Jacek Pudlo <ja...@jacek.jacek> wrote:
>>It does matter. If the narrative is *produced* through interaction,
>>authorial control is usurped by the player. It is my belief that this is
>>not
>>the case. We simply don't have the technology to make it happen.
>
> We certainly *do* have the technology to make it happen, but the problem
> is that the thing that results will be no fun at all.

So *totally* void and unfun, in fact, that it begs the question if it
qualifies as a narrative.

> One can certainly imagine an open-ended sim game, in which you wander
> around a landscape that reacts to you. Unfortunately, what you end up
> with is basically a narrative-interface roguelike. Much as it pains me,
> I think I would be forced to agree with Chris Crawford here: "that's not
> a story!"

Exactly what I've been saying. It's not a bad story -- it's simply not a
story to begin with.

> In fact, the Maze section of _Reliques of Tolti-Aph_--if it were shorn
> of all its unique rooms and talismanic items, and were the entirety of
> the game, rather than being an obstacle in the service of a larger
> narrative--would be an excellent example of this. For that matter, so
> is playing "Stochastic AD&D" where you turn to the Random Dungeon
> Creator in the back of the first edition DMG, and roll up random rooms,
> populate them with random monsters, and have random fights with them.

A simulation is to its story what the Trojan War is to the Illiad. You can
retrospectively narrate a simulation, but the simulation itself is -- by
definition -- not a narrative.

> We clearly have the technology to do this.

We have the technology to generate a simulation. We do not have the
technology to generate a narrative.


steve....@gmail.com

unread,
Apr 11, 2007, 8:55:23 AM4/11/07
to
Yes I think "generation" was an unfortunate choice. But again to be
fair, Nick isn't trying to say that the text is automatically
generated by the machine. (By a little benefit of the doubt and
process of elimination, along with a basic knowledge of how modern IF
systems work, one can deduce that he's probably referring to the
automation of *when* to print *which* prefabricated pieces.)

Nick is also most likely aware of "Integrated Natural Language
Generation Systems" (Kantrowitz and Bates, 1992), in which the authors
discuss GLINDA, a natural language generator for narrative generation
and NPC communication. (He's buddies with the Facade guys, one of whom
worked on the Oz project, which messed around with GLINDA at one
point.) And I'm sure there's got to be a lot of much more recent stuff
worth knowing on this topic. Maybe his language was struggling to be
sufficiently general to include this sort of research.

Jeff Nyman

unread,
Apr 11, 2007, 9:22:21 AM4/11/07
to
<steve....@gmail.com> wrote in message
news:1176296123....@o5g2000hsb.googlegroups.com...

> Yes I think "generation" was an unfortunate choice. But again to be
> fair, Nick isn't trying to say that the text is automatically
> generated by the machine. (By a little benefit of the doubt and
> process of elimination, along with a basic knowledge of how modern IF
> systems work, one can deduce that he's probably referring to the
> automation of *when* to print *which* prefabricated pieces.)

I agree. This is how I read it as well. The essay refers to a genre that
allows "a potential narrative, that is, a system which *produces* narrative
during interaction" [emphasis added] and is based on a "text-accepting,
*text-generating* computer program." [emphasis added]

When taken in context of the whole article, this does seem to speak to a
more general aspect of generating text (which, speaking loosely, it does do,
in response to player actions) and produces narrative (which, again,
speaking loosely, it does do, in response to player actions).

It's only if you read the terms to mean essentially "bringing something
totally new into existence that wasn't there before" that the idea diverges
from text-based interactive fiction systems that the article seems to be
speaking about.

Backing this up somewhat, in another paper of his ("Natural Language
Generation and Narrative Variation in Interactive Fiction,") he says:
"Modern-day IF systems such as Inform and TADS *generate language* simply by
printing orthographic strings when certain events are simulated in the world
or when objects need to be described." [emphasis added] In that same
article, he also says: "... there is no capability for re-ordering events so
that they are narrated in a sequence that is different from the one in which
they occurred, and it is difficult to even select the order in which objects
will be listed and described." In other words, he certainly seems to be
arguing against such systems generating anything really new, at least in
their current state.

I have to admit, I'm interested in this idea of an underlying "formal
system" to various things. (For example, I presented my belief in puzzle
form that a formal system underlies testing at
http://www.genuinetesting.com/s04/godel.html). That said, I tend to be very
skeptical of this sort of thinking in general and really look for it to be
operationally specific and to provide some predictive and/or testable
aspects.

- Jeff


Jeff Nyman

unread,
Apr 11, 2007, 1:01:41 PM4/11/07
to
I'm probably going to show my literary and/or narratological ignorance here,
but quoting from a related article by Nick Montfort, he says this:

"For instance, the system might change from narrating the story
externally (while focalizing on the player character) to narrating
the story from a point of view that is bound to the character,
incorporating that character's thoughts, mannerisms, biases,
and so on as it narrates. This sort of variation has been used to
accomplish many powerful literary effects, but it would require a
much more extensive model of the player character's consciousness,
of common sense reasoning abilities, and of the construction of
narrative language..."

(This is taken from "Natural Language Generation and Narrative Variation in
Interactive Fiction.")

My question is this:

Why can this sort of thing not be "programmed in" or "written in"? For
example, I find it easier for me to consider writing text-based interactive
fiction in the first person and, in that case, I *can* narrate the story
"from a point or view that is bound to the character." I *can* incorporate
"that characters's thoughts, mannerisms, biases" during the narration of the
game.

I have a feeling when the discussion turns to how narratology might relate
to text-based games and how simulation and narration are so different, I'm
missing some key point that would help me understand what the problem
actually is (or is perceived to be).

(I did just pick up Nick's book "Twisty Little Passages" so maybe that will
help me as well.)

- Jeff


signw...@lem.signwriter.name

unread,
Apr 11, 2007, 1:16:08 PM4/11/07
to
On 11 Apr, 13:55, steve.bres...@gmail.com wrote:
> Yes I think "generation" was an unfortunate choice. But again to be
> fair, Nick isn't trying to say that the text is automatically
> generated by the machine. (By a little benefit of the doubt and
> process of elimination, along with a basic knowledge of how modern IF
> systems work, one can deduce that he's probably referring to the
> automation of *when* to print *which* prefabricated pieces.)

In other words, IF doesn't *generate text* so much as each playthrough
of an IF generates *a text*?

Jeff Nyman

unread,
Apr 11, 2007, 1:28:28 PM4/11/07
to
<signw...@lem.signwriter.name> wrote in message
news:1176311768.3...@n59g2000hsh.googlegroups.com...

I would just say it like it is: a text-based IF game outputs largely
pre-written text. That has the benefit, I think, of being accurate without
using words that can be misread too much, such as "generate" which some will
take to mean the creation of something entirely new that was not
pre-existing in some form. I say "largely" in that definition because some
parts of the text may be randomized or use text substitutions or whatever,
but from a narrative standpoint the text as a whole is pre-written and even
the substituted or random parts will generate from an existing stock.

Speaking loosely, you can call this "generation" from the player's point of
view, I suppose. That's why I always like to consider definitions from a
given point of view: in this case, between that of the author/writer and the
reader/player.

This notion is similar to me of the distinction between "creation" and
"invention", where the one is meant to bring forth something wholly new and
the latter is meant to indicate an arrangement or re-arrangement of existing
parts.

- Jeff


Jeff Nyman

unread,
Apr 12, 2007, 12:43:08 PM4/12/07
to
Responding to myself (but in case anyone's curious):

Having gone through Nick's two articles, one thesis, and a bit of his book,
I've come to the conclusion that I like his ideas but I'm not necessarily in
agreement when they advocate a new system, like the one he presents. Rather,
I see existing systems being able to do what he wants with "narrative
variation" but, that said, I see TADS 3 as the most effective contender for
*all* of the ideas he puts out there. (Not the only contender; but the most
effective. It's just opinion, not fact.)

In terms of why I say that with TADS 3, I refer to things like the effective
message text substitution, the regular expression support for parsing, the
subtime module tied to narrative events, the explicit modeling of states,
the usage of transient objects, the ability to easily extend aspects of
person and tense (such as the moldable past tense extension), the complete
ability to override how things like "undo" work, and so forth. I do think
Inform could do some of these things (ultimately in Glulx, that is) but,
interestingly and for various reasons, I think Inform 7's "natural language"
approach may actually hinder this ability rather than help it from an
authoring point of view.

It would be interesting to put these ideas to the test, I suppose. That's
part of the problem with stuff written at an academic level: it often
doesn't get the actual day-to-day practitioners interested in trying out the
ideas. That's a pity because I've come to feel that the core idea of
narrative variation and different narratological techniques does have a lot
of merit. (I could almost imagine a "Narrative Variation" comp of some sort
to put these ideas to the test.)

So, in short, all of what Nick seems to be talking about sounds like it
boils down to the need for a new system where simulation and narration can
be totally separated. However, I see it more as (1) broadening the
storytelling abilities and techniques of authors and (2) finding out what
can be utilized in existing systems to allow those techniques to be put to
use. Nick does sort of indicate this when he says "it will be necessary to
show how original IF works can be developed to take advantage of these
capabilities." What are "these capabilities?" Mainly the narrative variation
he's talking about.

- Jeff


Jacek Pudlo

unread,
Apr 16, 2007, 5:01:08 PM4/16/07
to
Jeff Nyman

> I'm probably going to show my literary and/or narratological ignorance
> here, but quoting from a related article by Nick Montfort, he says this:
>
> "For instance, the system might change from narrating the story
> externally (while focalizing on the player character) to narrating
> the story from a point of view that is bound to the character,
> incorporating that character's thoughts, mannerisms, biases,
> and so on as it narrates. This sort of variation has been used to
> accomplish many powerful literary effects, but it would require a
> much more extensive model of the player character's consciousness,
> of common sense reasoning abilities, and of the construction of
> narrative language..."

What does "system" refer to? Some kind of "narrative engine" that
"generates" a narrative in a similar way as a 3D engine generates 3D
graphics? It's hard to determine from the quotation alone, but the part
about the "extensive model of the player character's consciousness"
indicates that Montfort is indeed talking about a "narrative engine,"
something along the lines of the infamous Erasmatron. The problem with a
"narrative engine" is that discussing it is like discussing the moral
implications of time travel. Would it be morally justified to travel back to
1889 and kill the infant Hitler? Would it be a neat thing if you could have
a "narrative engine" "generate" a brilliant story by simply setting a few
flags? Sure! Would it be a swell thing if you could change the mode of
narration in mid-story by simply setting INTERNAL_NARRATION_ON to TRUE or
FALSE? Yessir! Would it be fun to be a billionaire? You bet!

The problem with simulationist IF theory is that the simulationists do not
discuss interactive fiction, but rather their wet dream of interactive
fiction. I mean, are Monfort's musings applicable to IF as we know it today?
Is Montfort actually suggesting that IF should focus on the creation of some
kind of "literary AI" that will be so intelligent and lifelike and creative
that it will, almost unaided, spout bouts of brilliant internal monologue?

> (This is taken from "Natural Language Generation and Narrative Variation
> in Interactive Fiction.")
>
> My question is this:
>
> Why can this sort of thing not be "programmed in" or "written in"? For
> example, I find it easier for me to consider writing text-based
> interactive fiction in the first person and, in that case, I *can* narrate
> the story "from a point or view that is bound to the character." I *can*
> incorporate "that characters's thoughts, mannerisms, biases" during the
> narration of the game.
>
> I have a feeling when the discussion turns to how narratology might relate
> to text-based games and how simulation and narration are so different, I'm
> missing some key point that would help me understand what the problem
> actually is (or is perceived to be).

The problem is that a simulation aspires to mimesis, while a narrative
aspires to diegesis. The graphics in a 3D shooter are supposed to *look*
like the things they represent. A text can never look like the things it
refers to because a text is made of written words and written words can only
mime silent monologues and streams of consciousness, and, to a much lesser
degree of mimesis, dialogue. Texts are referential, simulations are
representative. This is why textual simulation is a self-defeating oxymoron.
I'm not sure if the simulationists have addressed these issues, which lie at
the very core of their philosophy. I'm not even sure if they're aware of
them.


Adam Thornton

unread,
Apr 16, 2007, 6:56:05 PM4/16/07
to
In article <ouRUh.38717$E02....@newsb.telia.net>,

Jacek Pudlo <ja...@jacek.jacek> wrote:
>A text can never look like the things it
>refers to because a text is made of written words

I see your diegesis and raise you Lewis Carrol's "A Mouse's Tale" from
_Alice's Adventures In Wonderland_.

Adam

Jacek Pudlo

unread,
Apr 16, 2007, 7:46:37 PM4/16/07
to
Adam Thornton

A text can never look like the things it refers to, but its typography can
resemble those things. It's a semi-exception to the rule.


John W. Kennedy

unread,
Apr 16, 2007, 8:55:38 PM4/16/07
to

"Easter Wings" and others by George Herbert.

Lord, who created man in wealth and store,
Though foolishly he lost the same,
Decaying more and more,
Till he became
Most poor:
With thee
Oh let me rise
As larks, harmoniously,
And sing this day your victories:
Then shall the fall further the flight in me.

My tender age in sorrow did begin:
And still with sicknesses and shame
You did so punish sin,
That I became
Most thin.
With thee
Let me combine
And feel this day your victory:
For, if I imp my wing on thine
Affliction shall advance the flight in me.

There's a term for it: "carmina figurata".

--
John W. Kennedy
A proud member of the reality-based community.
* TagZilla 0.066 * http://tagzilla.mozdev.org

Adam Thornton

unread,
Apr 16, 2007, 10:10:20 PM4/16/07
to
In article <xVTUh.38725$E02....@newsb.telia.net>,

Jacek Pudlo <ja...@jacek.jacek> wrote:
>A text can never look like the things it refers to, but its typography can
>resemble those things. It's a semi-exception to the rule.

What about hieroglyphics?

What if hieroglyphics really worked the way, say, Athanasius Kircher
thought they did?

Adam

Jeff Nyman

unread,
Apr 17, 2007, 5:15:00 AM4/17/07
to
"Jacek Pudlo" <ja...@jacek.jacek> wrote in message
news:ouRUh.38717$E02....@newsb.telia.net...

>
> What does "system" refer to?

Nick describes the "system" in his articles and, from what I can tell, it's
a modified "IF System", similar it seems to me as what we have now. That was
one of my critiques: I'm not sure we need a new system, based on the ideas
he describes.

> graphics? It's hard to determine from the quotation alone, but the part
> about the "extensive model of the player character's consciousness"
> indicates that Montfort is indeed talking about a "narrative engine,"
> something along the lines of the infamous Erasmatron.

Well, the main thing he wants is a (development) system that allows you to
separate narration and simulation. His paper "Natural Language Generation
and Narrative Variation in Interactive Fiction" gives a graphic describing
this system.

But, like I said, much of what he describes hinges on the notion of
"narrative variation" and I'm contending, based on what I've read of his
work, and based on what I've played around with in terms of our current
development systems, I think this can be achieved with what we have, based
on the examples he gives of what narrative variation means.

That, to me, is the testable null hypothesis of his articles: can we develop
these ideas *without* the addition of a new system? (His new system is the
actualization of his hypothesis, as I see it.)

> The problem with simulationist IF theory is that the simulationists do not
> discuss interactive fiction, but rather their wet dream of interactive
> fiction. I mean, are Monfort's musings applicable to IF as we know it
> today?

In answer to your question, I think so, but I admit that the constant use of
the terms "narrative generation" or terms like it do make it seem unlikely.
When I read his articles, I think the ideas of narrative variation are
simply different ways to present the story to a player/reader (interactor,
as he calls it). I don't think it requires a different development system.
Rather, I think it requires getting people to think about how they write
their narrative portions of the game in a different way.

Note: I say his musings are applicable to, as you say, "IF as we know it
today" *if and only if* his sole goal is to achieve this narrative
variation. If, as I suspect and as you allude to, there is some emphasis on
a "narrative engine" that "generates" wholly new aspects of a
simulated/narrated world, then --- I'm not so sure. I say I'm not so sure
because I haven't seen anything testable stated that would help me
understand what such a system would be.

> Is Montfort actually suggesting that IF should focus on the creation of
> some kind of "literary AI" that will be so intelligent and lifelike and
> creative that it will, almost unaided, spout bouts of brilliant internal
> monologue?

Honestly, I don't think so, at least based on what I've read.

> The problem is that a simulation aspires to mimesis, while a narrative
> aspires to diegesis.

Here you seem to agree with Nick: narration and simulation must be separate
because their aims are wholly different. That's where I agree to a small
extent, but not categorically.

Simulation and narrative, to me, can be self-reinforcing. The simluation is
what allows an *interactive* narrative to be propelled along. That level of
interactivity is different between a novel and a text-based game (work) of
interactive fiction.

For example, I've often said that as a game player I don't care all that
much about mimesis as some abstract concept. I just want a fun game. Yeah, I
want it internally consistent with itself and that's probably some aspect of
mimesis, but I'm not sitting there looking for every aspect of how the game
world is not like the "real world." If I read a fantasy book (or play a
fantasy game), I know magic is going to break mimesis, at least according to
the world I know. But that same magic can be internally self-consistent
within the world that I'm reading about and thus a degree of mimesis is
achieved. Same thing as an author, for me: I want to present a compelling
story or entertaining gameplay. I feel that can be achieved even if some
mimesis is lost.

The diegesis you mention is basically the recital of the events. The
simulation *is* those events in some sense. I don't see how you can have
separation of the two. Part of how a simluation "aspires to mimesis" would
seem to be via the narration of events because it's that narration that
allows the reader/player to determine if the "world" and the "situations"
are internally consistent. The combination, for me, is what then determines
how compelling the story is *and* how entertaining the simulation is.

I know I'm probably looking at this very simplistically from people's point
of view here. But that's sort of my point as well. I'm not sure people are
looking at "narrationist models" and "simulationist models" and all the
distinctions that may sprout therefrom.

I know Nick is trying to determine how much various aspects of narratology
play into creating text-baed interactive fiction and how those aspects can
be "separated" from the simulation itself. That, to me, is where the
breakdown occurs. I do think some aspects of narratology can be applied.
(Narrative variation, for example.) I'm not sure you can separate those
totally from the simulation nor have I been convinced you would want to
because, ultimately, people are playing a game.

- Jeff


Jeff Nyman

unread,
Apr 17, 2007, 5:21:55 AM4/17/07
to
One clarifying element. Earlier, when I said,

> I'm not sure people are looking at "narrationist models"
> and "simulationist models" and all the distinctions that may
> sprout therefrom.

What I meant by "people" here are the average game players and/or book
readers as opposed to "people" who do think about these aspects quite
deeply, either via their academic route or just from a theoretical
perspective.

And what I meant was simply that all the academic musings on the theory
behind narratology and simulationism may not be how the average person
really approaches such games. Nick's contention seems to be that various
aspects of the theory behind these concepts will lead to games that are
better written. (Interestingly, the articles never speak to why these games
would be more fun to play, however, and I think that's a key component for
presumably obvious reasons.)

- Jeff

Conrad

unread,
Apr 17, 2007, 2:59:58 PM4/17/07
to
On Apr 17, 5:21 am, "Jeff Nyman" <jeffny...@gmail.com> wrote:

> And what I meant was simply that all the academic musings on the theory
> behind narratology and simulationism may not be how the average person
> really approaches such games. Nick's contention seems to be that various
> aspects of the theory behind these concepts will lead to games that are
> better written. (Interestingly, the articles never speak to why these games
> would be more fun to play, however, and I think that's a key component for
> presumably obvious reasons.)

Thank you!

Now we're getting somewhere...

So what makes games fun to play?


Conrad.


steve....@gmail.com

unread,
Apr 17, 2007, 4:43:06 PM4/17/07
to
Jeff Nyman writes:
> Simulation and narrative, to me, can be self-reinforcing. The simluation is
> what allows an *interactive* narrative to be propelled along. That level of
> interactivity is different between a novel and a text-based game (work) of
> interactive fiction.

Yes, that's how it's done currently. Works of interactive fiction
normally handle the progression of the story by some version of "lock
and key" (or, more generally, "obstruction and its means of removal").
The stage is set, and the player explores the first phase, but
progress is blocked until some obstacle has been removed or some goal
is accomplished; as problems are resolved, new problems surface, and
so the story develops (although not necessarily in a strictly linear
order), until the final obstacle is overcome and the story concludes.
Here the narrative is totally determined by the simulation.

The simplest AI-like control over the narrative ("Drama Management")
comes in the form of a daemon which periodically checks what
conditions are fulfilled, and modifies the world appropriately
whenever certain preconditions have been met. Such modification (also
known as "warping") may alter any feature of the game, from
alterations in the map-layout and room descriptions, to the behavior
and location of simple objects, or the mood and motivation of non-
player characters and available conversation branches. This is roughly
the organization of Emily Short's recent work, _Mystery House_,
wherein one core daemon guides the behavior of all the characters. By
no means is this the total realization of the idea of "Drama
Management" -- and it's still strongly simulation based -- but it's
the best example I know of.

Jeff Nyman

unread,
Apr 17, 2007, 5:16:04 PM4/17/07
to

<steve....@gmail.com> wrote in message
news:1176842586.7...@p77g2000hsh.googlegroups.com...

> Jeff Nyman writes:
>> Simulation and narrative, to me, can be self-reinforcing. The simluation
>> is
>> what allows an *interactive* narrative to be propelled along. That level
>> of
>> interactivity is different between a novel and a text-based game (work)
>> of
>> interactive fiction.
>
> Yes, that's how it's done currently. Works of interactive fiction
> normally handle the progression of the story by some version of "lock
> and key" (or, more generally, "obstruction and its means of removal").

And I see that with so-called traditional fiction as well, such as a novel.
I'm reading "The Lions of Lucerne" right now, for example. I'm hanging out
with this protagonist, vicariously living with him, as he has one obstacle
after another ("obstruction") thrown at him and seeing how he confronts or
evades those obstacles ("means of removal").

Here the simulation of the world is passive, it seems to me. I don't
interact with it at all: I'm just carried along by it as long as I continue
to read. (I guess that is a form of interactivity, but it's pretty limited.)
That simulation is presented to me as a more-or-less continuous narrative.

So then I come across text-based interactive fiction and I largely see the
same thing, but, of course, the level of interactivity is different, thus
making the simulation more "active" in a sense.

At this point you may be thinking "What the hell is he talking about?" I'm
just trying to indicate how I'm currently thinking about these things. So
these thoughts are what led me to look more closely at Nick's idea because I
do realize that one thing I like about good novels (whether written by
Stephen King, Brad Thor, Clive Barker, etc, etc) is how the narrative is
presented: how the story is told *separate from the simulation itself*. So
there I sort of agree with what Nick is saying: consideration of separation
of simulation and narration. That said, I don't think this is in some
absolute sense: it's rather just realizing how they reinforce each other.

The other aspect of that is the narrative variation that Nick mentions which
I recognize is also part of what I like. For example, Richard Morgan writes
novels very differently form Neal Stephenson in terms of how they narrate
events. Stephen King utilizes variation in a different way than, say,
Bentley Little does.

So, I guess my main point is: I see Nick's work bringing two important
things to light: (1) separation of narrative from simulation; (2) narrative
variation.

What I don't see is why this requires a new system, which he seems to
advocate.

> Here the narrative is totally determined by the simulation.

Now I wonder if Nick is presenting an idea wherein the simulation determines
the narrative?

But, again, I don't see it like that. I see it that neither determines the
other, per se, at least in some categorical sense. They're both
self-reinforcing -- or so it seems to me. Can you just have a simulation
without a narrative of some sort? I don't see how. Even the most threadbare
narrative is still a narrative. Can you have a narrative without a
simulation? I don't think so, if you broaden out the notion of simulation
and consider it as being along a continuum of passive to active.

All of this is, of course, my opinion. This is what I'm curious about in
terms of how other people see it.

> This is roughly
> the organization of Emily Short's recent work, _Mystery House_,
> wherein one core daemon guides the behavior of all the characters. By
> no means is this the total realization of the idea of "Drama
> Management" -- and it's still strongly simulation based -- but it's
> the best example I know of.

Okay, and maybe here is where I'm encountering my stumbling block. The first
part of your statement here reinforced in my mind the contention that what
Nick talks about can be done now with existing "IF systems" without the
invention of his new one. Your second statement, however, regarding the
"total realization" is what I'm not getting. What would that "total
realization" be like? Right now it seems to me that anything beyond what
we've discussed would be something very much different from the simulation +
narrative either of traditional fiction or text-based interactive fiction.

Hopefully some of my ramblings here make sense.

- Jeff


Jeff Nyman

unread,
Apr 17, 2007, 5:21:11 PM4/17/07
to

"Conrad" <conra...@gmail.com> wrote in message
news:1176836398.1...@e65g2000hsc.googlegroups.com...

>
> So what makes games fun to play?

In this particular case, the question for me is what makes *text-based
interactive fiction* games fun to play. Because that's a key discriminant
here in some ways. (Maybe.) I think the ideas of narrative variation could
make such a text-based game fun to play, over and above such techniques
being used in a graphical-based game due to the nature of how people read
text and how they "interact" with text. In other words, I think the
techniques of narrative variation could be made more manifest in a
text-based game rather than any other. (I'm not sure about that; it's a bit
off-the-cuff on my part.)

Ultimately that's really the hypothesis I sort of see being sought here with
Nick's original work and with the periodic discussions that come up about
narratology or about simulation vs. narration and so forth.

Regarding "fun to play," I also think along the lines of what's "fun to
read" as well. Since reading is such a critical element of text-based
interactive fiction, I see that part as often not being discussed. I see
discussions about how to create better puzzles or whether or not toilets
should be modeled, but rarely discussion about the content of the narrative
that propels the story along.

- Jeff


steve....@gmail.com

unread,
Apr 17, 2007, 5:52:06 PM4/17/07
to
Jeff Nyman writes:
> I see it that neither determines the
> other, per se, at least in some categorical sense. They're both
> self-reinforcing -- or so it seems to me. Can you just have a simulation
> without a narrative of some sort? I don't see how. Even the most threadbare
> narrative is still a narrative. Can you have a narrative without a
> simulation?

Well I was talking about how simulation is leveraged by the author to
handle the pacing and the order of narrative events, the progression
of the story.

I recommend you check out the Oz Project...

http://www.cs.cmu.edu/afs/cs/project/oz/web/

... and especially the "Drama Manager," see _An Oz-Centric Review of
Interactive Drama and Believable Agents_ (Mateas, 1997). Of particular
interest is the Façade "Drama Manager," which Michael Mateas describes
in chapter 8 of _Interactive Drama, Art and Artificial Intelligence_,
Mateas' Ph.D. thesis. These are available from the Oz webpage under
"Publications."

Jacek Pudlo

unread,
Apr 17, 2007, 5:56:53 PM4/17/07
to
<steve....@gmail.com>

Wow! An actual exemple of a simulationist piece, instead of the usual empty
talk! Let's take a look at what textual simulation can accomplish when
infused with a compelling narrative.


PART OF TRANSCRIPT OF _MYSTERY HOUSE_ BY EMILY SHORT

Tom comes in the way you did, carrying some railroad money, a dry chicken
bone and a wrench.

> ASK TOM ABOUT ME

"I don't know or care about yourself," says Tom.

> ASK TOM ABOUT WRENCH

"Now," murmurs Tom. "If I were a necklace of priceless jewels, where would I
be?"

> ASK TOM ABOUT BONE

Tom hasn't left the room.

In the other room, the clock strikes 2:13 PM -- two deep bongs and a tinkle
of minutes.

> ASK TOM ABOUT RAILROAD MONEY

"If I find them, I'll let everyone know."

END OF TRANSCRIPT

In this short transcript spanning just four command lines, there is one
grammar error (an attempt at "natural language generation"?) one
inconsistency, two failed attempts at character interaction that are left
mysteriously unexplained and, from what I can see, very little drama.
Perhaps drama would unfold, had I bothered to continue playing, but I was
put off by the irresponsive NPCs who move around the house like ghosts
chasing Pacman. Actually, the best part of the game takes place before you
enter the house. There's little simulation there, but some good writing
instead.

I suppose the player is left to construct the drama himself, out of random,
semi-coherent snippets of text.


Adam Thornton

unread,
Apr 17, 2007, 6:21:55 PM4/17/07
to
In article <FobVh.38850$E02....@newsb.telia.net>,

Jacek Pudlo <ja...@jacek.jacek> wrote:
>I suppose the player is left to construct the drama himself, out of random,
>semi-coherent snippets of text.

It worked for Samuel Beckett.

Adam

Jeff Nyman

unread,
Apr 18, 2007, 7:06:21 AM4/18/07
to
<steve....@gmail.com> wrote in message

> I recommend you check out the Oz Project...

> http://www.cs.cmu.edu/afs/cs/project/oz/web/

> ... and especially the "Drama Manager,"


Thanks for the references. I took a look and it's definitely interesting
reading. I need to go through all this a lot more, but one thing I can say
is that there is more here in terms of defining what's being talked about. A
few thoughts...

From the Oz Project Overview: "If our example had been a conventional story,
the author alone would decide exactly what happens to the protagonist. In
interactive drama, the interactor is the protagonist and determines the
action."

This depends on how widely you look at it, I would think. In a game,
ultimately *all* things that happen to the protagonist *are* determined by
the author. In that sense, the author does determine *all* action. It may
not be as obvious as in a "conventional story" and, granted, the interaction
goes from passive to active, but since a lot of these papers are talking
about the overall story arc, as opposed to individual elements of action,
this seems a more important distinction than it's often treated.

Again, from the Overview: "Even though the interactor is choosing what to
do, say, and think, there is a destiny, created by the author of the
interactive drama. This destiny is not an exact sequence of actions and
events, but is subtly shaped by the system, which embodies dramatic theory
and principle, in order to create a cathartic experience."

See, this sounds nice --- but it's not really an experiment at all. "Create
a cathartic experience?" I mean, yeah, we know what the words mean, but --
it would seem to be that this idea is already falling into the trap of what
Nick's papers do. Incidentally, the first part of this quote matches what I
said above. To wit, they say "Even though the interactor is choosing what to
do, say, and think, there is a destiny, created by the author of the
interactive drama." Right. It's the author that creates the basis for all
action (the "destiny", as they call it).

The rest is from "An Oz-Centric Review of Interactive Drama and Believable
Agents", which I found to be an interesting paper. The null hypothesis here
for me is that "believable characters" and the "drama management" can be
created with the systems we have. What's lacking is not the system, but how
the games are written.

Note that Nick had his two ideas: (1) separation of simulation vs. narration
and (2) narrative variation. This concept also has two: (1) drama management
and (2) believable characters.

For me a key question, upon reading this paper, was: how much AI does a
current IF system have to support in order to make believable characters?
The problem is that they haven't answered that question operationally
because, of course, that will differ from various people. (Read reviews on
Amazon and you'll see massive disagreements over movies and novels regarding
how believable or not the characters were in the movie or novel.) To their
credit, they do say this: "The success of a believable agent is determined
by audience perception. If the audience finds the agent believable, the
agent is a success."

I know they have this thing called "Hap" and Nick had his system as well.
But I never see these things. So all the academic musings are, to me, just
that: musings with little experimentative value. Except ..... I think
current systems can handle at least some of what they're talking about.
(It's questionable to me that current systems need to handle *all* that
they're talking about.)

But that seems to me to be a good avenue of exploration. I'm curious why all
these academic musings haven't translated down to the community that's most
likely to implement them and distribute them to people who can determine the
experiment's success. (Or maybe this has happened?)

At one point the paper says: "The Oz drama manager controls a story at the
level of plot points. Plot points are 'important moments' in a story."

This reminds me of setting up Story Points in TADS 3.

Then: "Given a particular set of plot points, the space of all possible
stories is the set of permutations of all possible plot points. The vast
majority of these permutations will be garbage - unsatisfying stories which
don't make sense. The author of the story has some particular ordering of
the plot points in mind - this is the story she wants to tell. Rather than
expressing this preferred sequence via structural constraints on the story
world, the author writes an evaluation function that captures her sense of
aesthetics for the story."

See -- but what's missing (as with Nick's paper) is how this would make an
entertaining game. I'm not saying it wouldn't --- but it's an assumption.
Why is not better to write the "particular ordering of the plot points in
mind" rather than this permutation set up? To me, it's like Stephen King
saying he had "particular ordering of the pages in mind" but rather than
present them that way, he's going to rely on the reader's "evaluation
function" to put the pages in an order they want that "captures [their]
sense of aesthetics for the story."

I realize I'm caricaturing a bit but, again, I'm approaching this as a
non-academic who wants to see how these ideas have practical value, both for
authors and players/readers.

The paper says: "This aesthetic is captured by some set of features the
evaluation function looks for in a permutation. Conforming to the shape of
some dramatic arc may be one feature in the function. Given a permutation of
plot points, the evaluation function rates the permutation. Assuming the
author has successfully captured her aesthetic, the original story should be
ranked high by the function."

I think it's an interesting idea -- but I'm questioning how workable of a
one it is.

The null hypothesis for me is that we can simulate something like a drama
manager in current IF Systems. Whether it has to be *exactly* like the drama
manager they talk about is left open for question because it's not shown --
to me -- that this system of theirs is effective or would be entertaining
for players. That's a key point that all these academic things leave out:
would this be fun? Would people want to play/read this?

I'm finding, and I may be wrong here, that like most academic things, it
separates itself from the actual usage of the stuff it promotes. (I see this
all the time in the testing world when academics come up with things; I see
it all the time in physics world when I have to hear about how superstrings
"tell us so much.")

The paper says: "A drama manager can be more or less generative while it
controls a story. To the extent that a drama manager has a fixed description
of a single story (linear) or set of stories (branching), it is not
generative. The possible stories that a user can experience while
interacting with the system are fixed. To the extent that the manager can
create a new story each time a user experiences the system, the story is
generative. Another way of thinking about this is capacity for surprise. To
the extent that a manager can surprise its author with a novel story, the
system is generative."

But to the game player, they don't know if it's due to a clever author or a
"generative drama manager" unless the contention that every single game
could be a new experience, which is clearly what they are talking about. But
is this what players seek? Do people want their favorite book to change each
time its read? Do people want a play they go to having a different story the
next time they see it? No, clearly not. But, granted, we're talking about
games: not plays and books. Yet a lot of the research that goes into these
(and Nick dealt with this too) is supposedly stemming from the research of
dramatic theory and narratology -- elements used in book writing and play
writing -- so I'm curious how what's being described even matches those
media, beyond how it might match game media.

It seems that the contention is often that games are different from
"conventional" story media and so different approaches could and should
work. Yet, these games are being considered in the context of that very same
media, such as in terms dramatic theory, narratology, etc. (This goes back
to how I didn't like how Nick tried to avoid the use of the word "game,"
instead preferring "work" because of the connotations of the terms.)

These are just some initial thoughts. Like I said, I want to go through
these with the same intensity, for lack of a better term, that I went
through Nick's papers.

- Jeff


Jeff Nyman

unread,
Apr 18, 2007, 1:53:32 PM4/18/07
to
"Conrad" <conra...@gmail.com> wrote in message
news:1176836398.1...@e65g2000hsc.googlegroups.com...

> So what makes games fun to play?

After reading the Oz material that Steve brought to my attention, I feel I
can answer this question a bit differently than I did before. Or, at least,
I can build on what I said before. You'll have to bear with me for a bit of
explanation about the ideas and the current systems.

I have to hope that these so-far academic ideas are meant to make games more
enjoyable across a certain spectrum of reading and playing. The contention
seems to be that a new type of IF system is needed. I'm not convinced of
that. That said, I can see why the idea is out there. To wit:

I think TADS 3 can do these things but doesn't encourage them.
I think Inform 7 encourages these things but can't do them.

"Can do it" and "can't do it" are being used very loosely here. I sacrified
accuracy for brevity. I just mean that, in my opinion, one system makes such
narrative concepts much more likely to (ultimately) be capable of being
implemented well than another. That's clearly just an opinion, one that
borders on a hypothesis for me based not only on having played around with
Inform 6/7 and TADS 2/3, but also having played many games of various
perceived caliber written in both systems.

When I say TADS 3 "doesn't encourage it", what I mean is that the focus is
so much on the language itself with all the documentation. That's not only
to be expected, it makes eminent sense. But when I see extensions like the
Story Guide, Rules, Relation, and the various knowledge-based extensions,
like Memory, I think there's a rich treasure trove of material for how to
write good narrative variation and good drama elements in TADS 3 games. All
that combined with the rich feature set of TADS 3, including the NPC
modeling (for the believable actors), makes me then believe that concepts
like narrative variation and drama management can occur quite naturally as
an outgrowth of good writing techniques and good programming techniques.

When I say Inform 7 "encourages it," I mean that only by the fact that it's
somewhat natural language approach could make a game writer think that they
are an author, in the true sense of the word, of a story and, as such,
attempt to write as an author would, in terms of narrative and drama and the
considerations stemming from those concepts. That's the encouragement I
refer to and it's certainly not to be scoffed at. The implementation of
Inform 7, however, is what I think ultimately doesn't encourage this beyond
that surface level.

I think the mechanisms for dealing with techniques that would make narrative
variation and drama management *efficient and effective* are not present in
Inform and I think part of that is due to the nature of the Z-machine itself
and how constraints are imposed. That would seem to suggest that Glulx would
perhaps remove those constraints --- and maybe it does, at least in terms of
certain elements. But I don't see it providing any more mechanisms to the
Z-machine that deal with the actual techniques of writing and the
encouragement of it. That would have to come from the language itself and
extensions written to it.

That latter point is where, for me, Inform 7 falls into the TADS 3 situation
of relying on the programmatic above all else, but I think the programmatic
nature of TADS 3 allows more than the programmatic nature of Inform 7.

I guess what I'm saying is that I see what comes out with Inform (either 6
or 7) tends to be very cookie-cutter like, in that its pretty much always
more of the same. I'm not saying the games themselves, but rather the
*mechanics* of those games that Inform presents. I always feel like I'm
interacting with the same game, just different plot. I've noticed with TADS
3 (and TADS 2, for that matter) that the games themselves don't always seem
so locked into the same mold. Even though I realize in many ways they are,
it doesn't always *feel* like it, and that's part of what matters: how
things are perceived to various types of players. The mechanics of TADS seem
more flexible to me than those of Inform.

So the *mechanics of the system* do seem to play a very large role in the
ability to craft and tell a good story. That's another hypothesis I see
rearing its head. And I guess that's what I see the Oz Project and Nick's
papers striving for: a way to encapsulate techinques into an IF System that
encourage various elements (like narrative variation, and so on) that work
well in other similar, related media (novels, films, plays, etc).

I think part of what they're arguing is that the bar has been set pretty low
for text-based interactive fiction. Not having written any publicly released
games myself, I'm perhaps treading a fine line when I say I can agree with
that, even while recognizing that there are some absolutely fine games out
there. But when I look at what makes those games "absolutely fine" (from my
viewpoint, of course) I realize it has to do more with how the author told
the story: how the narrative developed, how the points of view were used,
how compelling the back story was, how much I felt like I was involved in an
unfolding story, rather than walking along a pre-destined path.

So, finally, in answer to your (Conrad) question: It's this *story-telling
ability combined with contextually-relevant puzzles and situationally useful
items* is what, to me, makes a game fun to play and to read. Part of that
comes from the author and part of it comes from the mechanics of the system
the author must work within.

- Jeff


Adam Thornton

unread,
Apr 18, 2007, 2:29:21 PM4/18/07
to
In article <L4KdnfrKAvRZx7vb...@comcast.com>,

Jeff Nyman <jeff...@gmail.com> wrote:
>I guess what I'm saying is that I see what comes out with Inform (either 6
>or 7) tends to be very cookie-cutter like, in that its pretty much always
>more of the same. I'm not saying the games themselves, but rather the
>*mechanics* of those games that Inform presents. I always feel like I'm
>interacting with the same game, just different plot.

Because "Freefall", "Robotfindskitten", "Curses", "Textfire Golf", and
"The Space Under the Window" are all the same game?

Inform (6, anyway) gives you a *lot* more flexibility to do things that
are not traditional IF than TADS does. This may not be a good thing in
an IF development language, but there it is.

I'll assume you mean "IF games that are actually IF games" when you talk
about Inform games here, but I'm still somewhat puzzled.

>I've noticed with TADS
>3 (and TADS 2, for that matter) that the games themselves don't always seem
>so locked into the same mold. Even though I realize in many ways they are,
>it doesn't always *feel* like it, and that's part of what matters: how
>things are perceived to various types of players. The mechanics of TADS seem
>more flexible to me than those of Inform.

The only thing I can think of that you might accurately mean here is
that either the Inform default library messages have more of an
editorial tone to them than TADS (which I think is definitely the case),
or that TADS authors tend to customize the library messages more than
Inform ones do, which might be the case.

I think it's the first one. Inform--unless you customize the library
messages, tends to snark at you with things like "Is that the best you
can think of?" or "You jump on the spot, fruitlessly," rather than a
simple "You can't do that." Is this tone what you mean by "the same
game"? If it is--would it help if there were a full-on replacement for
the Library messages making them blander and more neutral? I don't
think that would be a hard project, and if you feel it would
significantly enhance your play experience, then it might well be an
avenue worth pursuing.

Of course, the library messages are already always customizable in an
Inform work, but it's also true that few authors take the time to do so.

>So, finally, in answer to your (Conrad) question: It's this *story-telling
>ability combined with contextually-relevant puzzles and situationally useful
>items* is what, to me, makes a game fun to play and to read. Part of that
>comes from the author and part of it comes from the mechanics of the system
>the author must work within.

I still don't really understand what you mean by "mechanics"--can you
elucidate a bit more, if it isn't just "tone of default library
messages" ?

Adam

steve....@gmail.com

unread,
Apr 18, 2007, 2:58:00 PM4/18/07
to
I agree that I6 is clumsy to program in and I7 is horrible to program
in, but I don't agree that I7's naturalesque syntax makes the
programmer think more like a writer, and is therefore more encouraging
of narrative. That's just I7 advertising pablum. You still have to
think like a programmer; you just have to do the extra work of
figuring out how to write the algorithm in (I/E)nglish.

Adam's vague claim that I6 is a more powerful general programming
language as compared to TADS (2?) is probably based on the single fact
that nobody ever bothered to write a Tetris clone in TADS. (And that
he hasn't seen Fundin's space-shootem game.) The claim is certainly
not based on an analysis of the languages.

Adam Thornton

unread,
Apr 18, 2007, 3:27:24 PM4/18/07
to
In article <1176922680....@p77g2000hsh.googlegroups.com>,

<steve....@gmail.com> wrote:
>Adam's vague claim that I6 is a more powerful general programming
>language as compared to TADS (2?) is probably based on the single fact
>that nobody ever bothered to write a Tetris clone in TADS. (And that
>he hasn't seen Fundin's space-shootem game.) The claim is certainly
>not based on an analysis of the languages.

That's true.

Is there a whole meagerie of TADS abuses, as exists for the Z-machine?

Off the top of my head, in the Z-code world, I can think of:

The ones I already mentioned, being Freefall (a Tetris clone), Textfire
Golf (an arcade-golf simulator, of sorts), and robotfindskitten
(sort of like Unix "robots"). In addition to those, I remember a text
editor, a roguelike, and a Scheme interpreter.

Is there a similar tradition of doing things you shouldn't oughta do in
TADS?

In any event, this was my point only insofar as to challenge the
statement that Inform games all felt like playing "the same game." This
was then vaguely laid at the feet of the "mechanics" of the system.

As I stated, I *think* this was an assertion about the default library
messages in each system, but I really don't know, and it is this point
that I really would like to understand.

Adam

Jeff Nyman

unread,
Apr 18, 2007, 3:57:07 PM4/18/07
to
"Adam Thornton" <ad...@fsf.net> wrote in message
news:1l4if4-...@quicksilver.fsf.net...
> In article <L4KdnfrKAvRZx7vb...@comcast.com>,

> Inform (6, anyway) gives you a *lot* more flexibility to do things that
> are not traditional IF than TADS does. This may not be a good thing in
> an IF development language, but there it is.

That may be the case, although the Wasteland RPG that was recently posted
for TADS 3 would seem to be tricky to do in Inform (either 6 or 7). I could
be wrong but I think TADS could do those "abuses" or "alternatives" but no
one has done so to the extent that they have with Inform.

What those abuses are supposed to prove about an IF System is not clear to
me beyond just that you can do it. So you are correct that I should have
made it clear that I'm talking about the games that are the focus of this
overall discussion: games that have a defined narrative content and a
background dramatic structure, such that all these hypotheses or models
(Nick's and the Oz stuff) could potentially be used to make such games more
entertaining.

> The only thing I can think of that you might accurately mean here is
> that either the Inform default library messages have more of an
> editorial tone to them than TADS (which I think is definitely the case),
> or that TADS authors tend to customize the library messages more than
> Inform ones do, which might be the case.

It's not just the responses but you're right: I'm vague on this. It's a feel
I get when I play the games. I admit that's totally subjective and I should
have made it clear at that point in my text that this was solely my opinion.
For example, playing "Return to Ditch Day" (TADS 3) and "The Plant" (TADS
2), I felt more like I was engaging in a game because of how the game world
seemed to respond. When I learned TADS 2 and TADS 3 by essentially recoding
those games, I noticed how much more I could do to make the narrative spin
off a bit differently. I found I was a little more limited in Inform with
this. I felt like I could create a living, breathing world more effectively
and efficiently in TADS than Inform.

Now, mind, I'm talking here about how the system encourages people to think
about what they are writing rather than all aspects of what the system can
do. I say that because clearly the above could be taken to mean that I'm
just not very good at coding in Inform 6 or Inform 7.

> Is this tone what you mean by "the same
> game"? If it is--would it help if there were a full-on replacement for
> the Library messages making them blander and more neutral? I don't
> think that would be a hard project, and if you feel it would
> significantly enhance your play experience, then it might well be an
> avenue worth pursuing.

I don't know if that's what I meant by "tone" but it's a valid thought. TADS
3 has it's own set of stock responses as well. I guess more what I'm talking
about is the gradual way that you can develop a game in TADS 3 such that the
"tone" gradually emerges out of each individual work. For example, I see a
wholly different tone in "Square Circle" than I do with "Return to Ditch
Day" even though both use largely the same kind of code constructs. I could
apply that as well to games like "Finding Martin" and "The Plant" and a few
others. I just haven't felt that same thing when playing Inform games, but I
admit that part of this is colored by how I have learned both systems, in
terms of what I perceive they can and cannot do.

Yes, I know: I'm still being vague.

But it's around this vagueness that I read Nick's articles and the Oz
articles and started to think about how I would apply concepts like "drama
management" and "believable actors" and "narrative variation" in one of the
existing IF Systems. And I realized that this (to me) seems "easier" in TADS
3 because I believe TADS 3 (again, for me) already "encourages" this kind of
thinking by how it presents its mechanics, whereas I don't feel the same for
Inform.

> I still don't really understand what you mean by "mechanics"--can you
> elucidate a bit more, if it isn't just "tone of default library
> messages" ?

One example would be the robustness of an NPC system, for example. The other
part would be how easy it is for me to not only conditionally modify
descriptions of locations but also narrative elements, not only as whole
sections of text but also as conditional elements within the text itself.
Another part would be the modeling of subjective and objective time in a way
that isn't tied at all to the "turn count" or "time clock" of the system.
Another part would be the way that objects could be stored between game
sessions because I believe that can be a powerful element towards some of
the techniques of narrative variation and drama management. Another part
would be the ease and ability by which you can wholly modify how all aspects
of the library work.

These "mechanics" manifest not only in how the game is created, I believe,
but in how its played because I don't believe you can ever truly achieve
this mimesis that some like to talk about. What I do believe is you can make
the game fun enough and the story compelling enough that people don't even
care about mimesis.

- Jeff


Jeff Nyman

unread,
Apr 18, 2007, 4:04:36 PM4/18/07
to

<steve....@gmail.com> wrote in message
news:1176922680....@p77g2000hsh.googlegroups.com...

>I agree that I6 is clumsy to program in and I7 is horrible to program
> in, but I don't agree that I7's naturalesque syntax makes the
> programmer think more like a writer, and is therefore more encouraging
> of narrative.

Well, keep in mind here's what I said:

"When I say Inform 7 'encourages it,' I mean that only by the fact that it's

somewhat natural language approach *could* make a game writer *think* that

they
are an author, in the true sense of the word, of a story and, as such,

*attempt* to write as an author would, in terms of narrative and drama and

the
considerations stemming from those concepts."

[emphasis added]

In other words, I'm not saying the syntax itself does in fact make this the
case. But a putative game writer/author could certainly *think* that it does
because they might associate it more with direct writing rather than
programming. And that's part of my overall point: the mechanics of a
system -- both how you play things from it and write things in it -- can
determine how people use it and how they even conceive of using it. (In the
graphical-based arena, witness how different users approach tools like AGI
Studio, SCI Studio, and Adventure Game Studio. Then compare that approach
that's taken with systems like Lassie or Wintermute.)

> You still have to
> think like a programmer; you just have to do the extra work of
> figuring out how to write the algorithm in (I/E)nglish.

Absolutely, one hundred percent agreed. I agree you can't escape the fact of
programming any more than you can escape the fact that you're writing what's
ultimately a game. (Which goes to my other point that these various academic
projects seem to want to disassociate the notion of "game" from "work" and I
honestly don't see why.) However, I think the emphasis a system places on
its mechanics and its constructs can determine how people think of it. That,
in turn, can determine the kind of works they see as possible with it.

So this latest round of my probably-not-to-intelligent-sounding missives has
been showing how I view the presentation of the mechanics of two system
(Inform 7 and TADS 3) and how those seem to relate to the notions of
incorporating more aspects of narrative and dramatic theory. Keep in mind
that those notions have seemed to involve the idea that new systems would
have to be created to handle such things. So the other part of my discussion
here has focused on why I don't think it's the case that a new system is
needed.

All of this is, of course, talk. I personally think it would be interesting
to at least formulate ways to experiment with these ideas and test them.

- Jeff


steve....@gmail.com

unread,
Apr 18, 2007, 4:29:15 PM4/18/07
to
Jeff Nyman writes:
> I'm not saying the syntax itself does in fact make this the
> case. But a putative game writer/author could certainly *think* that it does
> because they might associate it more with direct writing rather than
> programming.

Point well taken. I think this is the main thing I7 has going for it.
If someone doesn't know what programming is (and I'm talking a
pre-"programming for dummies" dummy here), they might like I7 because
they think it differs somehow from the absolute foreign-ness of what
vague concept they have of programming. So it might be a nice gateway
for an utter novice. But as soon as one starts using it, one becomes a
programmer, so I don't think the point holds for very long. Unless the
user continues to disbelieve they're programming. I don't know -- I'm
skeptical. But if it does work out like that for some people, great.

I agree by the way that I7 feels like a cookie-cutter. TADS 3 is this
way also, from a certain point of view, insofar as it has a sort-of
default world model and user interface. But it's not cookie-cutter
from a programming point of view like I7 is. The language is pretty
constricted by design, and guess it's not necessarily a bad thing,
although it seems clearly associated with the lack of procedural
programming power. That said, I would guess this is almost entirely
transparent to the player. The main thing the player's going to notice
in the different systems is the intelligence of the interface and the
"style" of the default reports.

Jeff Nyman

unread,
Apr 18, 2007, 4:53:21 PM4/18/07
to
<steve....@gmail.com> wrote in message
news:1176928155.7...@d57g2000hsg.googlegroups.com...

> So it might be a nice gateway for an utter novice.

Agreed. Now, I think it would be an interesting experiment if TADS 3 was, at
some point, given a (pseudo)natural language gloss. I'm not holding my
breath and I'm not saying this is the way TADS 3 should go. In fact, I'm
ultimately arguing that it's the very emphasis on the programmatic nature of
TADS 3 that could make authors focus more on narrative elements because the
distinction between the game mechanics and the narrative and drama that
plays out within them is more obvious.

Two points about what I just said:

(1) That still leaves Inform 6, which clearly doesn't hide its programmatic
nature to the same extent that Inform 7 attempts to. So even if my
"cookie-cutter" feelings are valid for Inform 7 -- just my viewpoint -- why
do I feel that same way with Inform 6?

(2) The only reason I say a pseudo-NL interface to TADS 3 would be
interesting as an experiment is it would be a way to attempt to validate
hypotheses. Again, I'm not holding my breath, however.

> But as soon as one starts using it, one becomes a
> programmer, so I don't think the point holds for very long.

Agreed -- and therein perhaps lies a bit of why Inform 7 and Inform 6
converge to me. They're ultimately the same system once you play something
written in it. Beyond a banner statement, I probably couldn't tell that a
game was written in Inform 6 or Inform 7. That's not a statement against
either system because the same could potentially hold for TADS 2 and TADS 3,
at least to a level of approximation.

But what I'm saying is that for me the "mechanics" of Inform 7 are really
the same as Inform 6 when you get to the aspect of how the game presents
itself. Now how the game presents itself is, of course, a combination of the
possibilities of the system (language + compiler + vm + interpreter) but
also of how the author could use the system.

[Slight side-note: one thing I've noted is that people who play Hugo games
tend to like them quite a bit for the stories of those games. This is
interesting to me because Hugo is clearly not used as much as Inform or
TADS. Equally interesting is that Hugo, according to Kent Tessman's manual,
was built largely around Inform's structure but then modified to handle
perceived problems of how games were written. So I say that's interesting
because one thing Hugo games seem to get high marks for is the overall
"gameplay experience" and I have to believe that some of that is due to the
writing. Is that because of how the mechanics of Hugo present themselves to
the author? I don't know. Having played Hugo games myself, as well as
dabbling in the language, I can attest to these same feelings.]

> I agree by the way that I7 feels like a cookie-cutter. TADS 3 is this
> way also, from a certain point of view, insofar as it has a sort-of
> default world model and user interface. But it's not cookie-cutter
> from a programming point of view like I7 is.

And, see, maybe that's part of what I'm getting at. I keep talking about
"mechanics" and as I tried to indicate in my response to Adam I mean this
not just in terms of how the game presents itself ultimately but how the
game can be constructed via the elements that the system makes available. I
guess maybe what I'm looking at is flexibility in some sense. I see TADS as
being inherently more flexible than Inform. I could certainly be way off
base on that. But that inherent flexibility that I see is what leads me to
believe that TADS could be a more effective system, overall, for the kinds
of ideas surrounding narrative and drama.

> programming power. That said, I would guess this is almost entirely
> transparent to the player. The main thing the player's going to notice
> in the different systems is the intelligence of the interface and the
> "style" of the default reports.

And I think this is what Adam was indicating to me as well. That's the part
I'm questioning. I do think there is another aspect that's not going to be
totally transparent to the player and I think it does hit on these ideas of
Nick and the Oz project. What I mean is that the type of story that the
system allows to be told (via the different ways an author can construct it)
is going to be apparent to the player at some level. Perhaps not in the
sense that they're going to say "Ah, well, clearly this system allows
narrative variation techniques and a fine-grained parsing of input." But in
some sense that "Hey, I really like the games that are produced with this
system."

I know this probably sounds like I'm on some anti-Inform crusade and it's
actually not that. I would say the same things I've said about Inform about
systems like ADRIFT or Quest at this point. But I'll be the first to admit
that I'm having a bit of trouble locking onto exactly what I mean here,
which tells me I'm not thinking all aspects of this through well enough.
(What it probably tells everyone else is that I've clearly got way too much
time on my hands.)

- Jeff


Conrad

unread,
Apr 18, 2007, 4:54:42 PM4/18/07
to
On Apr 18, 1:53 pm, "Jeff Nyman" <jeffny...@gmail.com> wrote:
>
> I think part of what they're arguing is that the bar has been set pretty low
> for text-based interactive fiction. Not having written any publicly released
> games myself, I'm perhaps treading a fine line when I say I can agree with
> that, even while recognizing that there are some absolutely fine games out
> there. But when I look at what makes those games "absolutely fine" (from my
> viewpoint, of course) I realize it has to do more with how the author told
> the story: how the narrative developed, how the points of view were used,
> how compelling the back story was, how much I felt like I was involved in an
> unfolding story, rather than walking along a pre-destined path.

I think you're right on here: conversely, for me, it will wreck my
enjoyment of a game when I need to go into "test-taking" mentality:
"How did the person who designed this puzzle intend me to solve it?"

Frankly, I've always considered the emphasis on puzzles to be an
attempt to fill the void left by the difficulty of adapting narrative
to an interactive medium: puzzles are easier to program than a truly
responsive storyline (and more interesting to readerish players than
combat sims).

To my mind -- I don't have a theoretical basis to give this claim
objective credence -- the distinction between story and puzzle is at
least as important as between story and simulation. And, depending on
how one defines one's terms, "puzzle" and "simulation," while not
identical, seem to reside on logical levels quite close to each
other... certainly in comparison to "narrative."


> So, finally, in answer to your (Conrad) question: It's this *story-telling
> ability combined with contextually-relevant puzzles and situationally useful
> items* is what, to me, makes a game fun to play and to read. Part of that
> comes from the author and part of it comes from the mechanics of the system
> the author must work within.

Okay, so:

* story-telling
* contextually relevant puzzles
* useful items

I'm going to see how well I'm reading you here:

-- Story-telling sets the context of the puzzles; puzzles are
interesting insofar as they have to do with the story, as they engage
and are charged by the story; useful items are useful in terms of
solving the puzzles and exist at the level of simulation.

That seems true to me, and I'm guessing it's what you're driving at.


If I can take that a little further, the kinds of game that I enjoy
playing are ones that have a broad responsivity to a diverse range of
player action. "Guess the verb" puzzles, I think, irritate us because
they stonewall us; but we also delight in games that respond to our
whacky ideas.

You've talked about "dramatic management," and I'm not certain exactly
what one of those would look like (but I'll look for one to try out):
to my mind, the next generation of interactive fiction engines will
need more elegant ways of dealing with complex behaviors and handling
states of affairs, (which will be preliminary to good conversational
engines), and these things will be more-or-less worked out in the way
inventory management and locations are now.

I see inform 7 as a step toward that, although how big a step I'm not
currently able to evaluation, not having good facility with the thing
yet. TADS I don't know.

Regardless of engine preference, good IF starts with the storytelling;
and even with current limitations, existing technology is manifestly
adequate to make fun games.


Conrad.


ps - I will make a confession here:

When I mentioned "guess the verb" as stonewalling, in comparison with
a game that responds to whacky ideas, my personal ideal is a game in
which a situation can be handled with a few obvious actions, with
acceptable results:

--but that same situation yields better and better results as the
player becomes more innovative with his actions.

I'd like to see that kind of model displace, almost completely, the
item-based approach, and this is where I need to say it:

I confess, this preference of mine is based on the implication of
where the power lay: because when it's a matter of finding the right
item, the power to remove the obstacle is outside the player: but
when it's a matter of finding the right action, the power is inside.

And although I haven't followed up on the Oz project, I do recall a
lesson from the movie...

Make for a short game, tho.


Conrad.


Jacek Pudlo

unread,
Apr 18, 2007, 5:59:28 PM4/18/07
to
"Adam Thornton"

What if pigs could fly?

Do you see _Pale Fire_ as an evolvement of _The Real Life of Sebastian
Knight_?


Jeff Nyman

unread,
Apr 18, 2007, 6:30:06 PM4/18/07
to
"Conrad" <conra...@gmail.com> wrote in message
news:1176929682.8...@d57g2000hsg.googlegroups.com...

> I think you're right on here: conversely, for me, it will wreck my
> enjoyment of a game when I need to go into "test-taking" mentality:
> "How did the person who designed this puzzle intend me to solve it?"

Yes, I agree. That's for me where the contextually relevant puzzles (or
perhaps situations) come into play. Which leads to your thought:

> Frankly, I've always considered the emphasis on puzzles to be an
> attempt to fill the void left by the difficulty of adapting narrative
> to an interactive medium: puzzles are easier to program than a truly
> responsive storyline (and more interesting to readerish players than
> combat sims).

I agree. I like a good puzzle to a certain extent but how that puzzle is
presented can make all the difference in the world. It's interesting as well
regarding the distinction between narration and simulation because a lot of
people would probably lump "puzzles" into one aspect of simulation: a
simulation of a complicating element of the world model (i.e., locked door
or something). On the other hand, the notion of puzzle can be, I think,
broadened out to the notion of situation. What's the situation the player is
in? How can they use the items around them? What do they perceive as
situationally useful, not just in terms of the "puzzle" (situation) but in
terms of how that fits into the wider concept of the narrative itself.

I think what some of these papers are arguing for is a notion where the
"puzzle" is part of the narration or, at least, part of how the narration
develops. I personally don't think I would want to take it that far but
that's still an aspect of this whole thing that I'm looking into.

> Okay, so:
>
> * story-telling
> * contextually relevant puzzles
> * useful items
>
> I'm going to see how well I'm reading you here:
>
> -- Story-telling sets the context of the puzzles; puzzles are
> interesting insofar as they have to do with the story, as they engage
> and are charged by the story; useful items are useful in terms of
> solving the puzzles and exist at the level of simulation.
>
> That seems true to me, and I'm guessing it's what you're driving at.

Correct. Now, in my mind, how those puzzles/situations become interesting
and work at the level of the story and how they incorporate the items of the
world model around them is pontentially spoken to by these concepts of (1)
narrative variation, (2) believable actors, (3) drama management.

That, at the very least, is the one hypothesis that I would make based on
reading these various papers and based on my own attempts at authoring
text-based IF and playing it. It's that concept that I think is potentially
testable via our current IF systems.

Saying that, of course, required me to consider the various systems and how
those systems present mechanics that seem (only seem, mind you) to speak to
those concepts as I understand them. The problem is that those concepts "as
I understand them" would make for a pretty boring hypothesis and experiment.
So that's part of what I was looking at here: are these ideas potentially
translatable into actionable elements that people could try out? In that
case, those terms need to be defined so we recognize what we're dealing
with, which speaks to your next point:

> You've talked about "dramatic management," and I'm not certain exactly
> what one of those would look like (but I'll look for one to try out):

And note that I'm quoting from the Oz project with the terminology so I'm
not entirely sure what it would look like either. That, to me, is part of
the experiment.

> to my mind, the next generation of interactive fiction engines will
> need more elegant ways of dealing with complex behaviors and handling
> states of affairs, (which will be preliminary to good conversational
> engines), and these things will be more-or-less worked out in the way
> inventory management and locations are now.

Those are good points, perhaps: "dealing with complex behaviors" and
"handling states of affairs." Those sound like things that could be
operationally defined in terms of actionable games (or parts of games) that
could be written in various systems that could perhaps help determine how
much the mechanics of a system matters. This speaks to another of your
points:

> Regardless of engine preference, good IF starts with the storytelling;
> and even with current limitations, existing technology is manifestly
> adequate to make fun games.

I agree: I think. That's the part I'm not sure on. Part of the problem is
that what one person considers good, another considers garbage. But what I'm
also looking at is whether that assumption is true, regarding the IF system
being used. Or is it the case that the mechanics of the system and how it
presents its elements can make a difference? Nick and the Oz project group
seem to think so, going so far as to create new systems (or at least the
conceptual basis for them). My point about Hugo, while maybe not indicative
of anything at all, does still interest me in that one of Hugo's design
goals was to make the writing of the games easier than in Inform. Coupling
that with the seemingly common praise for Hugo games, even some not written
by established authors to my knowledge, has made me consider what role the
storytelling medium presents as part of the experience.

It's an open question to me. It's one I used to dismiss but now I'm not so
sure.

> I see inform 7 as a step toward that, although how big a step I'm not


> currently able to evaluation, not having good facility with the thing
> yet. TADS I don't know.

What aspects of Inform 7 do you feel are a step towards that? I'm curious
because your thoughts in that regard may help me better understand why I'm
feeling TADS 3 is a step towards that.

> --but that same situation yields better and better results as the
> player becomes more innovative with his actions.

Right, almost like an adaptive world model of sorts. It has a "knowledge" of
the contextually relevant situations that the player is facing and the
situationally useful ways that a player could deal with those situations.
Knowing the "useful" ways implies that the system can know the boundary of
the "useless" ways and thus respond to those, but in an adaptive fashion:
meaning, not just the same thing each time, but rather the system evolves
its responses to how the player is dealing with the situation.

(Off-the-cuff: it's almost like the system allows the author to design a
game such that it responds to how much the player apparently "gets its",
meaning how much they clearly understand the nature of the world model and
the nature of the story based on how they are attempting to deal with
situations.)

> I confess, this preference of mine is based on the implication of
> where the power lay: because when it's a matter of finding the right
> item, the power to remove the obstacle is outside the player: but
> when it's a matter of finding the right action, the power is inside.

Exactly. I think this makes a lot of sense. I need to think about how I can
relate what you just said here to what I've been looking at with these
various ideas, but I definitely like what you've said here.

- Jeff


Adam Thornton

unread,
Apr 18, 2007, 6:25:30 PM4/18/07
to
In article <4xwVh.38906$E02....@newsb.telia.net>,

I haven't read the latter. Should I? I loved _Pale Fire_ (big
surprise, right?).

Adam

Adam Thornton

unread,
Apr 18, 2007, 6:37:44 PM4/18/07
to
In article <xbOdnT7TtbMm6rvb...@comcast.com>,

Jeff Nyman <jeff...@gmail.com> wrote:
>But it's around this vagueness that I read Nick's articles and the Oz
>articles and started to think about how I would apply concepts like "drama
>management" and "believable actors" and "narrative variation" in one of the
>existing IF Systems. And I realized that this (to me) seems "easier" in TADS
>3 because I believe TADS 3 (again, for me) already "encourages" this kind of
>thinking by how it presents its mechanics, whereas I don't feel the same for
>Inform.

I think I begin to see what you're driving at:

TADS3 (and TADS2 to a lesser extent) provides you with a very rich set
of things in the library to conceptually model the game world.

Whereas in Inform, if you want that kind of deep modelling, you have to
build it yourself; the language and libraries really encourage you to
build a stage set, not a model world (and I7, with its notions of
"scenes" and "not in play," goes much further towards encouraging you to
structure your game as a drama rather than as a simulation).

Is that what you're getting at?

If so, yes, I agree with you: that is a way in which the languages feel
very different to use.

[...]

>Another part would be the modeling of subjective and objective time in a way
>that isn't tied at all to the "turn count" or "time clock" of the system.
>Another part would be the way that objects could be stored between game
>sessions because I believe that can be a powerful element towards some of
>the techniques of narrative variation and drama management. Another part
>would be the ease and ability by which you can wholly modify how all aspects
>of the library work.

Are the library modification hooks any deeper than Inform's?

I'm less sanguine about allowing serialization of the game state and
letting you read it back in outside the explicit parameters of
save/restore. It's neat, but it gives the virtual machine too wide a
road to the outside world for my tastes. I'm *STILL* angry about that
(DOS, not TADS) Comp game that reset my system clock.

>These "mechanics" manifest not only in how the game is created, I believe,
>but in how its played because I don't believe you can ever truly achieve
>this mimesis that some like to talk about. What I do believe is you can make
>the game fun enough and the story compelling enough that people don't even
>care about mimesis.

And I think I agree with you that Inform and TADS3 take very different
roads toward that end. But then my complaint with TADS3 for the couple
of years before its release was basically that people were spending
inordinate amounts of time and effort polishing the chrome of very
obscure corners of the libraries, rather than declaring it good enough
and moving on. I find that stage sets do what I require them to, and I
shouldn't spend too much time modelling the actual world when I can fake
it. However, you might very well get annoyed at my games, because, in
fact, if you don't play along with the presentation, you find out too
quickly that it *is* a stage set, not a model.

Adam

Jeff Nyman

unread,
Apr 18, 2007, 7:04:46 PM4/18/07
to
"Adam Thornton" <ad...@fsf.net> wrote in message
news:o6jif4-...@quicksilver.fsf.net...

> I think I begin to see what you're driving at:
>

> [snipped]


>
> Is that what you're getting at?

Yes, exactly. That combined with various experiments I've tried in both
systems to model things, like for example, switching person and tenses
mid-game. I know both systems can do it, but I found that how TADS's
language is constructed, it encouraged this possibility as a game style
more. Another example is a time travel concept I was considering for a game.
In order to truly model the paradox situations in a way that made them seem
paradoxical, I needed something like TADS' transient ability. Further, for
that same idea and a related one given how a past tense story was told, I
needed to be able to totally overhaul how save, restore, and undo were
responded to. Another idea was a game based totally on subjective time,
where the events happened relative to this subjective time sense, which in
turn was relative to what situations the player was encountering. I found
TADS' subtime ideas (used in "Return to Ditch Day", but to limited effect)
works well for this kind of idea.

Now, granted, all of what I just said here might not sound like it would
make any sort of good or compelling story. That's as may be. I'm not sure
yet, myself. But what I found is that how TADS is constructed -- not just
its library, but its design philosophy -- allowed me to consider how I would
do all of these things. I tried those things in Inform 6 and 7 and while I
could do some of them, it was quite a bit more cumbersome and often not
entirely possible, at least to the same extent.

Maybe this speaks to your next question:

> Are the library modification hooks any deeper than Inform's?

To be honest, I'm not sure. I'm an expert in either system but I've spent a
lot of time playing with them. Granted, I haven't written a full game that
I've released yet but I've spent a great deal of time building up snippets
of "situations" and ideas that I then tried to implement in Inform and TADS.

So maybe it is a case that literally everything I'm talking about could be
done in Inform. I haven't seen that and when I've asked questions about some
of these things over the course of time, I've gotten good answers and things
to try. But I never found its usage in Inform as "compellingly possible" (is
that a phrase?) as in TADS.

> and moving on. I find that stage sets do what I require them to, and I
> shouldn't spend too much time modelling the actual world when I can fake
> it. However, you might very well get annoyed at my games, because, in
> fact, if you don't play along with the presentation, you find out too
> quickly that it *is* a stage set, not a model.

That makes sense, in terms of the distinction between stage set and model.
And maybe that is a key point of how I'm looking at this, in terms of
comparing what I see is possible with TADS and what I see is possible with
Inform. Clearly what I'm looking for would only matter to someone who wants
to write a game with those features. And even then that still leaves the
wide-open question of whether those things would be fun to play.

So part of what I found was that I find there is merit to the ideas of
narrative variation, believable actors, and drama management. Based on the
admittedly scant evidence for these things and based on my knowledge of the
craft of writing conventional fiction, I at least see it as a workable
hypothesis that such concepts could lead to entertaining games, perhaps an
evolution in how such games are thought about and presented.

- Jeff


Adam Thornton

unread,
Apr 18, 2007, 7:12:24 PM4/18/07
to
In article <_fWdnfssZPkkPrvb...@comcast.com>,

Jeff Nyman <jeff...@gmail.com> wrote:
>So part of what I found was that I find there is merit to the ideas of
>narrative variation, believable actors, and drama management. Based on the
>admittedly scant evidence for these things and based on my knowledge of the
>craft of writing conventional fiction, I at least see it as a workable
>hypothesis that such concepts could lead to entertaining games, perhaps an
>evolution in how such games are thought about and presented.

You might want to study Level 9's _Knight Orc_. It was a frustrating
game which had a bunch of NPCs, with goals of their own, who interacted
with each other and the player in sometimes-entertaining ways. I really
can't think of anything since that's done a better job of autonomous (if
dumb and greedy) NPCs.

If something like that were done right, with the amount of computing
resources we have to burn two decades later, it could be really cool.

Alas, my NPCs are Tables Of Conversations and occasionally a simple
follow-you-around daemon. They will fool no one in a Turing test,
though sometimes they are good for a chuckle.

Adam

James Mitchelhill

unread,
Apr 18, 2007, 8:14:32 PM4/18/07
to
On Wed, 18 Apr 2007 14:27:24 -0500, Adam Thornton wrote:

> In article <1176922680....@p77g2000hsh.googlegroups.com>,
> <steve....@gmail.com> wrote:
>>Adam's vague claim that I6 is a more powerful general programming
>>language as compared to TADS (2?) is probably based on the single fact
>>that nobody ever bothered to write a Tetris clone in TADS. (And that
>>he hasn't seen Fundin's space-shootem game.) The claim is certainly
>>not based on an analysis of the languages.
>
> That's true.
>
> Is there a whole meagerie of TADS abuses, as exists for the Z-machine?
>
> Off the top of my head, in the Z-code world, I can think of:
>
> The ones I already mentioned, being Freefall (a Tetris clone), Textfire
> Golf (an arcade-golf simulator, of sorts), and robotfindskitten
> (sort of like Unix "robots"). In addition to those, I remember a text
> editor, a roguelike, and a Scheme interpreter.
>
> Is there a similar tradition of doing things you shouldn't oughta do in
> TADS?

There's my ascii video demonstration, written in TADS 2:
<http://wurb.com/if/game/2367>.

--
James Mitchelhill
ja...@disorderfeed.net
http://disorderfeed.net

Conrad

unread,
Apr 18, 2007, 8:36:16 PM4/18/07
to
On Apr 18, 6:30 pm, "Jeff Nyman" <jeffny...@gmail.com> wrote:
> "Conrad" <conradc...@gmail.com> wrote in message

>
> news:1176929682.8...@d57g2000hsg.googlegroups.com...
>
> > I think you're right on here: conversely, for me, it will wreck my
> > enjoyment of a game when I need to go into "test-taking" mentality:
> > "How did the person who designed this puzzle intend me to solve it?"
>
> Yes, I agree. That's for me where the contextually relevant puzzles (or
> perhaps situations) come into play. Which leads to your thought:
>
> > Frankly, I've always considered the emphasis on puzzles to be an
> > attempt to fill the void left by the difficulty of adapting narrative
> > to an interactive medium: puzzles are easier to program than a truly
> > responsive storyline (and more interesting to readerish players than
> > combat sims).
>
> I agree. I like a good puzzle to a certain extent but how that puzzle is
> presented can make all the difference in the world. It's interesting as well
> regarding the distinction between narration and simulation because a lot of
> people would probably lump "puzzles" into one aspect of simulation: a
> simulation of a complicating element of the world model (i.e., locked door
> or something). On the other hand, the notion of puzzle can be, I think,
> broadened out to the notion of situation. What's the situation the player is
> in? How can they use the items around them? What do they perceive as
> situationally useful, not just in terms of the "puzzle" (situation) but in
> terms of how that fits into the wider concept of the narrative itself.

Absolutely! IF puzzles usually are at the level of simulation, but
there's no need that they must be. Puzzles, as you mention, might be
narrational in nature: "How do I reach my desired ending?" is, in the
appropriate game, a puzzle, but it operates on the level of narrative
rather than simulation.

Even recognizing the existence of a puzzle might be a puzzle: and
that steps about as far out of the narrative frame as I can imagine.

I'd say what's actually going on in IF games, if we were to look for
pre-computational analogues, is actually not very close to games.
Games, like chess, poker, or football, are usually rule-based contests
between two or more people (also there are solitary win-or-lose
contests), and the game is generally played many times.

IF games are not well descibed as "puzzles," either: classic examples
of puzzles, like crosswords, ciphers, sudoki, and so on, generally
follow a set and small number of rules, the forms are learned and many
of that form are pursued.

An IF game is not like either of these: you usually only play it
once, or a few times; you may not even see it through; and, crucially,
the rules of the game are generally not entirely known to you, and
discovering them is a major goal of play (and where the fun is).

So, IF games are not much like puzzles or games: in these ways
mentioned, they are like riddles, especially in that understanding the
rules is a major goal.

(Plus, of course, they are like fiction...)

> I think what some of these papers are arguing for is a notion where the
> "puzzle" is part of the narration or, at least, part of how the narration
> develops. I personally don't think I would want to take it that far but
> that's still an aspect of this whole thing that I'm looking into.

Again, I think you're exactly right there: both in expanding the
definition of puzzle and in your caution of including the traditional
simulationist types (since it's better to expand the range of
responsiveness than to replace one kind of limitation with another).

> > Okay, so:
>
> > * story-telling
> > * contextually relevant puzzles
> > * useful items
>
> > I'm going to see how well I'm reading you here:
>
> > -- Story-telling sets the context of the puzzles; puzzles are
> > interesting insofar as they have to do with the story, as they engage
> > and are charged by the story; useful items are useful in terms of
> > solving the puzzles and exist at the level of simulation.
>
> > That seems true to me, and I'm guessing it's what you're driving at.
>
> Correct. Now, in my mind, how those puzzles/situations become interesting
> and work at the level of the story and how they incorporate the items of the
> world model around them is pontentially spoken to by these concepts of
> (1) narrative variation, (2) believable actors, (3) drama management.
>
> That, at the very least, is the one hypothesis that I would make based on
> reading these various papers and based on my own attempts at authoring
> text-based IF and playing it. It's that concept that I think is potentially
> testable via our current IF systems.

Yeah, seems right. It's been mentioned that "believable actors" is a
subjective term, in that people are always arguing over what fictional
characters are believable and what aren't.

Ofc, in an important sense we're talking about art, and most of our
terms will be subjective: but we can talk about willing suspension of
disbelief here: in an important sense, players will forgive non-
representational output if they buy in at the emotional level: look
at the Sims and Tamagatchis.

Drama management... I've got to mull this over.

Well, there's no arguing taste. But even people who disagree on
*which* games are good seem to have in common that *some* are.


> But what I'm
> also looking at is whether that assumption is true, regarding the IF system
> being used. Or is it the case that the mechanics of the system and how it
> presents its elements can make a difference?

I have no doubt that a system can make a huge difference: the more
control an artist has over the medium, the finer-tuned a product he
can make. My point (probably unnecessary) was not to wait for the
perfect medium.

> Nick and the Oz project group
> seem to think so, going so far as to create new systems (or at least the
> conceptual basis for them). My point about Hugo, while maybe not indicative
> of anything at all, does still interest me in that one of Hugo's design
> goals was to make the writing of the games easier than in Inform. Coupling
> that with the seemingly common praise for Hugo games, even some not written
> by established authors to my knowledge, has made me consider what role the
> storytelling medium presents as part of the experience.

I haven't looked at Hugo; I'll check it out.


> It's an open question to me. It's one I used to dismiss but now I'm not so
> sure.
>
> > I see inform 7 as a step toward that, although how big a step I'm not
> > currently able to evaluation, not having good facility with the thing
> > yet. TADS I don't know.
>
> What aspects of Inform 7 do you feel are a step towards that? I'm curious
> because your thoughts in that regard may help me better understand why I'm
> feeling TADS 3 is a step towards that.

Again, I may be off here: but I have a little experience with Prolog,
and it seems that inform 7 is basically an inference engine in drag.
Inference engines are the core of knowledge bases: a certain kind of
AI.

I haven't been able to get a straight answer out of the Inform 7 crew
on this point, for whatever reason, but that seems to be the
understructure of the I7 whirligig.

I don't want to make fabulous claims here: it is not clear to me to
what extent I7 operates in this way, if it can handle recursion,
whether it interprets the I7 code inferentially, but the resulting
game is necessarily procedural, and so forth. I'm guessing from some
of her work that Emily Short has at least read up on AI: she tips her
hand when, for example, she talks about applying a path-finding
algorithm to npc's goal-directed traversal of conversation trees in
her glass slipper game.


Anyway, more practically:

= [I predicted] the next generation of interactive fiction engines
will
= need more elegant ways of dealing with complex behaviors and
handling
= states of affairs, (which will be preliminary to good
conversational
= engines), and these things will be more-or-less worked out in the
way
= inventory management and locations are now.

(What I'm getting at is that coversation is dependant on modeling the
complex behaviors and the states of affairs.)

I7 seems to have ways of breaking *things* up into parts (but no such
ability at this time for actions); If you look at the "relations" data
type in i7, it provides the ability to group *things* (including
people, but aga