Message-ID: <3D2AD77C.8070908@csi.com>
Date: Tue, 09 Jul 2002 08:30:52 -0400
From: John Colagioia <JColagioia@csi.com>
Organization: Organization?  Ha!
User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.0rc2) Gecko/20020618 Netscape/7.0b1
X-Accept-Language: en-us, en
MIME-Version: 1.0
Newsgroups: rec.arts.int-fiction
Subject: Re: NLP Question
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> <3d26f53c@excalibur.gbmtech.net> <n9hW8.1091$f5.68744@news>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: ool-182f30fa.dyn.optonline.net
X-Original-NNTP-Posting-Host: ool-182f30fa.dyn.optonline.net
X-Trace: excalibur.gbmtech.net 1026217442 ool-182f30fa.dyn.optonline.net (9 Jul 2002 08:24:02 -0400)
Lines: 249
X-Authenticated-User: jnc
Path: news.duke.edu!newsgate.duke.edu!news-hog.berkeley.edu!ucberkeley!newsfeed.stanford.edu!logbridge.uoregon.edu!uunet!sea.uu.net!ash.uu.net!excalibur.gbmtech.net
Xref: news.duke.edu rec.arts.int-fiction:106025

DarrenH wrote:
 >>In particular, most of the games mentioned come from the
 >>Infocom standard; the instructions of those games described
 >>the use of "look."
 > That's my point.  Infocom as a company was essentially defunct
about 13
 > years ago.  Shouldn't we be trying to plow new turf?

"Not so much," as my sister likes to say.

See, here's the thing.  "Look" makes some sense to some people, and a
lot of sense to people familiar with the genre.  We're all together
on this point, I believe.

You can call "look" anything you want.  "Look around," "l," "examine
the environment," "where am i," "what the @#$! is going on around
here," whatever you like.  And, no matter what you choose, some new
people will find it "instinctive," and others won't get it.  In fact,
it'll be just like it is now, except that the veterans, who you
*know* will try to play your game, will be confused and annoyed.

"So," you say (because you're now required to play Devil's Advocate
at this stage of the discussion), "that just means I just need to add
more synonyms."  And, you're right, to a certain extent.  The more
synonyms, the less likely a novice is to get "stuck."  But, there are
two very large problems:

First, the more synonyms, the larger the programming effort, and the
larger the game.  I tried this, once, plugging in every synonym I
could think of for my Inform game.  When I abandoned it, it had
something like one room, two or three objects, and didn't fit on the
Z-Machine, anymore.  This is only a problem if we decide not to be
snobs and favor the "technical elite" who have higher-end computers,
of course.  "Everyone" has a Cable Modem and a 2GHz microprocessor,
nowadays, right...?

The other problem is far worse.  The more you accept, the less you
reject.  This sounds obvious, trivial, and even *good*.  But it
isn't, once you start thinking about the effects.  Let's say you
include the entire contents of the OED, Webster's International, and
any other dictionary you find floating around.  Good, right?  Any
"real" word the player uses will be recognized (at some level) by the
parser.  Cool!

Except, what's the probability that the player *meant* to type "> GUT
CWM"?  I don't know what the hypothetical player was planning on, but
you can be darn sure that the plan wasn't to remove the internals of
a hollow.  And you can't reject it, because, what if that *was* what
the player wanted, right?  You can't dismiss his synonym...

Heck, add an auto-correction feature, so that no matter what the
player types, *something* will happen...kinda like some UNIX
installations I've encountered, actually...

 >>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.
 > Again, that's my point.

OK, then, let me ask a question:  Should artists who try to create
sculptures, for example, in the style of Rodin "move on"?  After all,
Rodin has been dead far longer than Infocom.

Similarly with those precious few companies that still produce hand-
illuminated manuscripts?  Those are very outdated, after all...

I hope those are rhetorical questions.

It's not, I don't think, an uneducated or ill-considered decision to
emulate a style which is (a) proven, and (b) fondly recalled.  And if
there are limitations, I'm not sure we've reached them, yet.

I'm not saying that's a reason to stagnate, by any means.  That would
be short-sighted and ignorant.  However, if "we" (I say, as if this
is some kind of club with membership rules) spread ourselves too thin
chasing "what could be" without a firm enough grounding in what's
possible now, the odds are not slanted in favor of good results.

[...]
 >>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.
 > I'm not bashing any perceived establishment.  I'm saying that if
one wishes
 > to write the engine for their own IF it shouldn't be as if one is
burning a
 > bible.

And I don't think it *is* viewed that way.  Disagreement is hardly
excommunication, after all.

I'll grant you that it can *look* like that, on rare occasions (and
the shocking occasions are much easier to remember), but keep in mind
the tone of the question.  There's a big difference in subtext
(whether it is intended or not) between "why doesn't anyond do this
'obviously' superior thing?" and "I'm thinking of doing this; does
anyone see any problems up-front?"

The first question (very close to your original phrasing) carries a
subtext of "you people should be doing this," while the second takes
the implied burden (which you didn't mean to imply) off the reader's
shoulder, since it's squarely stated to be *your* project, and simply
invites discussion on the topic.

And, actually, I'm partially at fault for this.  Let me amend my
original response to you, by asking what I should have asked when
responding to your question:  "What problem do you see in modern IF,
that answering the question 'where am i' would solve?"  Or, any of
your other proposed solutions, for that matter.

See, just like it's really easy to shout "this could/should be
better" when confronted with "the establishment," it's also easy to
jump on proposed solutions, assuming that they're just random whims
("I think this would be much better...in green"), forgetting that
there's a problem behind them that might have a *better* solution.

The problem right now is that I don't see the problem; I don't think
many other people around here do, either.  In all seriousness, try to
take some time to outline the problem that *you* see, ignoring what
either of us thinks is "the state of the art," and ignoring any
potential solutions for a moment.

 > Inform is one of the great tools we have for IF; it's not
impossible to
 > improve on it, but I doubt most people would bother trying.

It depends on what you mean.  The language, proper, probably isn't
worth changing to any great extent (though I've heard of people doing
it), because it exposes just about everything that's in the
Z-Machine.  Even so, arguably, Inform could benefit from some added
features, in all likelihood.

The library, though...In addition to dozens (hundreds, yet?) of
library additions, there's Platypus, an entirely rewritten library,
and I can think of a handful of people (including myself) who regard
the library as something that can--and should--be modified whenever
it suits the game.

 > It allows a
 > true programming environment, where IF writers can create their
own commands
 > and alter existing ones.
 > Now the minute you allow change to the standard libraries you're
changing
 > Inform.  Why is that somehow not considered heresy and my
suggestion is?

Put simply--and I'll probably repeat this somewhere, later, so please
excuse where it happens--it's a matter of cooperation.  I'd much
rather, because I think you're capable of it, play in the same
sandbox I'm in, so that you can build on my experiences, and I can
reciprocate when the time comes.  "Me" being any hypothetical person
using Inform or TADS, really.

You build your own system, and I might play the few games that come
out of it.  If I like them, I'll even say, "I wonder how to do this
in Inform."  But it'll be mostly convergent evolution and wasted,
redundant effort on one of our parts.

To put this in the simplest possible terms, I'd much rather see the
bright people (who are complaining) around here focus their efforts
on finding the weak points and overcoming them, rather than claiming
that such weak points exist, pulling up stake, and going their own,
isolated direction.

Stand on the shoulders of giants, as they say.  It's better for
pretty much everybody.

 >>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'm not asking writers to flock to my system.  I was talking about
creating
 > a proprietary system that, to the player, produces an end result
as good as
 > or better than an Inform interface, except that it's not open-source.

A purist might take offense at the term "open-source," which has an
annoyingly strict meaning.  That the source code is freely available
does not make a project open-source.

I don't care about such things, though.

 > Creating another open-source standard wouldn't do anybody any good.

I fail to see how any other standard would do well (two major
standards tends to be the maximum number a typical community can
bear, which explains the relative disinterest in Alan, Hugo, PAWS,
ADRIFT, and the twenty or so other systems lurking around).

 > No
 > popular programs in any other genre are open-source (to my
knowledge), such
 > should be the case for IF, in my opinion.
 > Linux can't really be brought up as an example to contradict my
point here.
 > Linux isn't on it's way to taking over desktops by any stretch of the
 > imagination (sorry if you're on a Linux box).

Technically, the Macintosh makes for a good counterexample.  I'm not
familiar with the actual scheme at Apple, but the OS/X kernel is the
open-source Mach, and the remainder of the non-UI operating system is
FreeBSD, which is almost the Open Source Granddaddy.

If Carbon and Cocoa are what I think they are, almost the entire
system may very well be open source, or at least an instance of a
collection of existing open source projects.

Not that I'm a FSF fanatic--far from it, in fact.  But I think my
point is that openness of construction isn't necessarily related to
popularity or usefulness.

Having said that, I *do* believe that the existing IF languages could
probably benefit from a competing system.  Most of the advances in C,
for example (for better or for worse) have come from alternate
compilers adding their own extensions, which were then copies by
others.  Neither TADS nor Inform has this.

Am I planning to do this?  Maybe, actually.  I have a project or two
brewing that may tie into it in a few months.

[...]
 >>- Maintain most of the advantages of existing systems,
 >>  including error recovery ("That's not a verb I
 >>  recognize") and portability.
 > I see that message in Inform games all too often.

But the point is, it's a *good* message, regardless of whether or not
it bores you.  It's concise, and carries exactly the content that a
player needs to know when typing something unusual--to wit, "try a
different verb."

The message can also be changed with very little effort, which is a
very good thing, even though too few authors actually go to the
trouble.

Contrast this with other systems where the equivalent default is
something like "Eh?" or "What are you talking about?"  Many systems
have a vague response that's either insulting, or doesn't convey
enough information about what might be wrong.  And others have a
message that works in some games (because of the tone), but it can't
be changed.

[...]

