X-Newsreader: Geminisoft Pimmy 3.2 Eng - www.geminisoft.com
From: "John Colagioia" <JColagioia@csi.com>
Newsgroups: rec.arts.int-fiction
Subject: Re: NLP Question
Date: Sat, 06 Jul 2002 09:48:21 -0400
References: <ts1V8.1521$x6.950@newsread1.prod.itd.earthlink.net> <3d24d781.19079434@news.east.cox.net> <vE4V8.344$Yx5.51@newsread1.prod.itd.earthlink.net> <5q5V8.424$Yx5.149@newsread1.prod.itd.earthlink.net> <3d250e22.33066338@news.east.cox.net> <YW5V8.489$Yx5.375@newsread1.prod.itd.earthlink.net> <ag3m74$tvi$01$1@news.t-online.com> <2fiV8.863$f5.47482@news> <ag4bp4$in3n5$1@ID-60390.news.dfncis.de> <NtoV8.900$f5.49200@news>
MIME-Version: 1.0
Content-Type: text/plain
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: ool-182f30fa.dyn.optonline.net
X-Original-NNTP-Posting-Host: ool-182f30fa.dyn.optonline.net
Message-ID: <3d26f53c@excalibur.gbmtech.net>
X-Trace: excalibur.gbmtech.net 1025963324 ool-182f30fa.dyn.optonline.net (6 Jul 2002 09:48:44 -0400)
Organization: ProNet USA Inc.
Lines: 183
X-Authenticated-User: jnc
Path: news.duke.edu!newsgate.duke.edu!news-hog.berkeley.edu!ucberkeley!news.maxwell.syr.edu!nntp.abs.net!uunet!dca.uu.net!excalibur.gbmtech.net
Xref: news.duke.edu rec.arts.int-fiction:105783

"DarrenH" <jeremy@overtime.ca> wrote:
[...]
>If I were to imagine myself dropped into a simulated-environment type of
>game, essentially blindfolded but all necessary detail being dictated to me
>by someone beside me, I think one of the first questions I would ask is
>"where the heck am I".  Simple enough I think.  One of the most basic things
>I can imagine asking when I'm lost and (usually) amnesic.

This may be a stupid question, but why does everyone keep
referring to starting the game with amnesia?  In my
(admittedly possibly limited) experience, the better games
tend to (a) come with background material to read, (b)
quickly establish or at least sketch the background in the
opening text, or (c) make this a key point of the story.

I suspect that if you keep having the feeling that you
don't know what's going on, you're either playing less-
complete games, or you're not reading the material.

>The problem is, nobody thinks to include the ability for the game to
>understand that simple question:  Where am I.
>So as a test I loaded up five games: Curses, Spider and Web, Shrapnel,
>Theatre, and The Space Under the Window (I haven't played all of these but
>I'll submit that they're likely all excellent games.  My comments are not
>meant as criticism of the games themselves but to the envelope they're
>delivered in).
>Here's the responses I received to the question "Where am I":

Question:  Was the opening text of these, including the
description of the starting locale, insufficient
information for you to figure this out?  I get "Attic (in
Meldrew Hall)" in "Curses," and "End of Alley" in "Spider
and Web"--both with fairly complete descriptions.

Responding to the on-screen description of that location
with "Where am I" seems a less like you're playing the
game you've loaded and more like you're trying to
contrive evidence, to be honest.

[...]
>What is my point?  That I should have used the word "look" to find out where
>I am.
>I should have known that, shouldn't I?  But why?
>The reason I should have used the word "look" is because that's the way
>games have always done it.

In particular, most of the games mentioned come from the
Infocom standard; the instructions of those games described
the use of "look."

I'll grant you that it is unfair for most games to not come
with similar documentation.  I'd give Curses more leeway on
that point, though, seeing as how, when it was written, it
would have been unlikely that an Infocom non-fan would've
picked it up...

Although, on that note, I miss "look around" as a verb.

>Had the programmers felt a need to code up a
>WhereIs [Noun] verb, a simple verb/noun grammar, it would have fixed the
>problem.  But it's not worth it, because they know we'll use the word
>"look".

Yes.  And most programs on the Macintosh follow the same UI
guidelines, because those that don't are unusable to those
who have Mac experience; same with just about every
platform.

See IBM's RealPhone and RealCD (Windows programs, in this
case) for why this is generally a truism.

>Multiply this assumption by all of the assumptions we walk into a game
>carrying (Floyd, bring me the ladder) and you've got a gaming interface that
>hasn't really progressed beyond Hitchhiker's Guide to the Galaxy in a real
>sense.  We're unwittingly allowing ourselves to fall into a pattern of
>coding that hasn't changed in years.

I'm fairly sure it's not unwitting; at least, not for all
programmers.  In the case of many of the games you mention,
it was (presumably) a considered decision to mimic the
"look and feel" of Infocom's line.

>We're allowing the envelope, the gaming platform, to make design decisions
>for us.

To be blunt, that's the job of a tool.  It makes your life
easier, partly because it constrains your choices.  That
is, it multiplies your power (be it physical or mental) by
forcing you to focus your energies.

[...]
>But that isn't the same as innovating the design of the envelope itself,
>despite what can already be done 'under the hood'.  The only course to
>progress (and maybe even a new market) is to adapt, rework, and throw out
>ideas.

Or, ask things like "how can this particular thing be done
in existing systems?"  I have yet to find, after five years
of tinkering (which may soon result in an actual game),
anything that (a) can't be done in Inform which (b) can be
done at all easily in another language, IF or not.

I see how to do sense-passing, though I haven't actually
tried it.  I have most of the parts (I believe) of an NPC
that'll act on his own volition (though not yet in
"interesting" ways).  I have some fairly simulation-heavy
objects (like the old "I need a knife, but I don't have a
knife, so I'll break something fragile and use a shard"
kind of thing), though they need major, major debugging
before being let out into the wild.  None of these are
constrained by the platform--no more than they're
constrained by needing to run on a modern-type CPU after
being compiled.

I think that sounds discouraging.  It shouldn't.  If you
or anyone else *wants* to build their own system from
scratch, have a ball with it.  It's an entertaining
activity (having written a handful of compilers,
interpreters, and virtual machines, myself, I speak from
direct experience), and occasionally, though pretty
rarely, proves useful.

However, three caveats:

1.  It's always more work than it seems.  Much more.

I'm not even talking about debugging.  I'm talking about
research into the current state of the art (to make sure
you're actually doing something new), design, and actual
development.

2.  Your energies would probably be more appreciated in
    raising the existing technology to your standards,
    rather than bashing the "establishment" in favor of
    your own, not-yet-existing system.

Doing this leads to the existing community feeling that
it's under another pointless attack (any community is
like this), and some will try to defend themselves,
leading to the annoying, "mad-scientist-esque"
persecution complexes that some people around here seem
to have.

3.  Unless the resulting system is provably better in
    many respects, don't expect anyone to flock to your
    banner.  In many ways, an "almost perfect" system
    can seem like a greater failure than a sparse proof
    of concept.

I can think of at least five "new systems" off the top
of my head which started with the best of intentions,
but shot themselves in the foot by failing to do one
or more of the following:

- Materialize.  As I said, it's hard.
- Maintain most of the advantages of existing systems,
  including error recovery ("That's not a verb I
  recognize") and portability.
- Create a standard library, so programmers needn't
  write their own "take" verb, for example.
- Be extensible, so programmers *can* write their
  own "rant" verb, if they need it.
- Succeed in its original goal.  A system that's
  "just like TADS," but with one new feature that
  doesn't actually work, isn't compelling.

A "good" system also needs (as I think has been said
many times, around here) a good game to show it off,
spurring the imagination of potential users, and
makes sure there aren't any glaring omissions in the
system.

Would any of the existing system designers like to
tell me that I'm talking through my...uhm...hat?  I'd
*love* to hear that this stuff is much easier than
I'm making it out to be.  I hate being the cynical
one...

[...]

OK, I've (yet again) ranted *more* than enough.
Hopefully, there's something useful, or at least
interesting, to someone, somewhere in this mess...
