The Wizard of Wishbringer

7 views
Skip to first unread message

Kirk Davies

unread,
Oct 5, 1993, 7:09:50 AM10/5/93
to

I've been meaning to type this article in for a long time now.
I thought it especially relevant since the release of "Return to Zork",
so I got my act together, and here it is.

I will be sending it to ftp.gmd.de for archiving.

Copied without permission from
AmigaWorld January/February 1986 pages 70-73

The Wizard of Wishbringer
=========================

A totally objective, highly critical and unbiased interview with
Infocom game designer Brian Moriarty, by Brian Moriarty.


At first I was elated when the editors of AmigaWorld asked me to review
my new Infocom story, Wishbringer. Here was a chance to sidestep the
jaded critics and bland press releases, and tell the world the truth
about the thankless life of a game designer! Eagerly I sat down and
composed a long, flowing tribute to myself, backed up by a detailed
autobiographical sketch, flattering color portraits and lengthy
examples of Wishbringer's deathless prose.

"Too biased," complained the editors after uncrating my manuscript.

"Of course it's biased," I snapped over the phone. "What did you
expect from a designer reviewing his own game?"

After a heated exchange and many threats, I agreed to ditch the review
and allow myself to be interviewed, but only on the condition that I
ask the questions as well as give the answers.

Q: How did you become a game designer at Infocom? Did you join the
company as a programmer in the microcomputer division, hacking in
machine language on Ataris, Commodores and TRS-80 Color Computers,
until one day Marc Blank, vice president and co-author of Zork, touched
you with his magic wand and made you one of the few, the proud, the
implementors?

Brain Moriarty: Yes.

Q: Wishbringer is your first game for Infocom, right? Where did you
get the idea?

BM: The design started with the game package. I was trying to think
of something neat we could include in the box, a magical item that
would tie in well with a fantasy theme. It couldn't cost too much,
maybe a quarter tops, and it had to be easy to mass produce. At first
it was going to be a magic ring. But that's been done so many times
before - Wagner, Tolkien, Donaldson, etcetera - that I decided to make
it a rock instead. The story emerged from that.

Q: Describe the story in excruciating detail.

BM: [Sigh] Oh, all right. You play the part of a mail clerk in a
small seaside village called Festerton. Your mean old boss, Postmaster
Crisp, orders you to deliver a mysterious envelope to the Magick Shoppe
on the far side of town.

When you get to the Shoppe, you meet an old woman who asks you to read
the envelope. It turns out that her pet cat's been kidnapped by
somebody called the Evil One. The ransom is Wishbringer, a magic stone
famous in local legends. Your mission, should you choose to accept it,
is to rescue the cat without getting turned into a furry toilet seat
cover.

When you return to the village, everything is screwed up. All the
familiar landmarks are twisted into sinister new forms. The streets
are patrolled by giant army boots. Trolls, vultures, hellhounds and
grues make your life difficult, and everything's under the all-powerful
eye of the Evil One.

Fortunately, you're not alone. Friendly pelicans, platypuses and
seahorses will help you if you're nice to them. And if you really get
stuck, you can invoke the power of Wishbringer, the Magic Stone of
Dreams.

Q: Infocom is famous for its clever packaging. What do you get when
you buy Wishbringer?

BM: Besides the glow-in-the-dark magic stone, you get a facsimile of
the mysterious special-delivery envelope from the Evil One, a fold-out
color map of Festerton and a booklet, The Legend of Wishbringer, that
explains the origins of the stone and how to use it to make wishes.
Oh, and you get a disk, too.

Q: Wishbringer is billed as an Introductory Level game. Is it really
just for beginners, or can veteran players enjoy it?

BM: Most of the problems in the story have two or more solutions.
The easy way out is to use Wishbringer. If a beginner gets frustrated,
he can whip out the magic stone, mumble a wish and keep on playing.
Experienced players can search for one of the logical solutions - a bit
harder, perhaps, but more satisfying. It's possible to complete the
story without using any of the stone's seven wishes. In fact, that's
the only way to earn the full 100 points.

The puzzles are highly interconnected. Once you start wishing your
problems away, it's very hard to continue playing without relying more
and more on the magic stone. The impotence of idle wishing - that's
the moral of Wishbringer. All really good stories have a moral.

Q: How long did it take you to write this moral tale?

BM: I started coding in September of 1984. In December, I deleted
most of what I'd written and started again. The disks went out for
duplication on May 1st, so I guess it took nine months altogether.
That's fairly typical for an Infocom title.

Q: How is an Infocom story developed, anyway? What kind of computer
do you use?

BM: Glad you asked. Infocom's Z Development System is based on a
DECSystem-20 mainframe, a machine that resembles a fleet of red
refrigerators. All of the game designers are connected to it, so it's
easy for us to share code and ideas and to play each other's games.

The programming language we use was created expressly for writing
interactive fiction. It's called ZIL (for Zork Implementation
Language). ZIL "knows" about concepts like rooms, objects, characters
and the passage of time. It has instructions the designer can use to
manipulate these concepts in very sophisticated ways.

ZIL itself is written in a LISP-like language called MDL, or Muddle,
which was developed at MIT's Laboratory for Computer Science. Because
ZIL and its utilities operate in a high level environment, it's
relatively easy for us to tinker around with things and make
incremental improvements.

Q: Infocom games are available on every home computer I can think of.
It must take a lot of programmers to do so many conversions!

BM: Naw. The Z System produces machine-independant code that can be
executed on just about any computer with enough disk space and RAM.
All we have to do is write a single machine-language interpreter for
the computer in question. Once the interpreter is running, all of our
present and future titles become available for that machine.

The Amiga interpreter was relatively painless. We simply downloaded
the 68000 Kernal developed for the Macintosh and Atari ST systems and
changed the I/O to make it work with the Amiga's operating system.

Q: One of the Amiga's big selling features is its graphics. Why don't
Infocom's games use graphics?

BM: Why aren't all books illustrated? [Pausing for effect] Should we
succumb to the temptation to throw in lots of cartoony pictures and
special effects just because the hardware is capable of it? We'd
rather invest our time in writing better stories, more evocative prose,
making the user interface as transparent as possible, and getting rid
of every bug we can find. We think these efforts result in a better
interactive experience than what has been achieved by "graphics
adventures." Our sales suggest that we're right.

That's not to say Infocom will never do graphics. We've been actively
working on some graphics-oriented ideas for a couple of years now. But
if the day comes when we offer a graphics entertainment product, you
can be sure it won't be Zork With Pictures.

Q: What about Cornerstone, Infocom's powerful, yet oh-so-easy-to-use
database system for the IBM PC? Will there be a version for the Amiga?

BM: It's technically possible. Marketingwise, I suppose it depends on
how many machines are bought and what types of people buy them. You
never know.

Q: What about you? Got any more game ideas?

BM: I've started work on a big science-fantasy game that will be
released some time in 1986. The story has an interesting historical
angle. That's all I can say about it now... except that it will
definitely not be for beginners!


Wishbringer author Brian Moriarty, 28, is the newest member of
Infocom's team of interactive fiction authors. He brings to the medium
the stern morality of a rural New England upbringing and a lifelong
passion for the fantastic. Write to him (or he'll write to himself)

c/o Infocom Inc.,
125 Cambridge Park Drive,
Cambridge,
MA 02140.

Pat Shepard

unread,
Oct 5, 1993, 9:00:49 AM10/5/93
to
Too bad they *did* in fact end up doing "Zork with pictures." Not
only did we get Zork worth Zero, we also got the new thing.

David Baggett

unread,
Oct 5, 1993, 11:43:23 AM10/5/93
to
In article <1993Oct5.1...@trl.oz.au> kda...@titan.trl.OZ.AU (Kirk Davies) writes:

(Actually, Brian Moriarty writes:)



>BM: Why aren't all books illustrated? [Pausing for effect] Should we
>succumb to the temptation to throw in lots of cartoony pictures and
>special effects just because the hardware is capable of it?

Evidently, yes.

Dave Baggett
__
d...@ai.mit.edu Boot up, log in, drop out. MIT AI Lab
ADVENTIONS: We make Kuul text adventures! Ask about Unnkulian 1, 2, 0, 1/2
PO Box 851 Columbia, MD 21044 USA / CIS: 76440,2671 / GEnie: ADVENTIONS

Hans Guijt

unread,
Oct 5, 1993, 12:55:01 PM10/5/93
to
In article <1993Oct5.1...@mnemosyne.cs.du.edu> pshe...@nyx.cs.du.edu (Pat Shepard) writes:
>From: pshe...@nyx.cs.du.edu (Pat Shepard)
>Subject: Re: The Wizard of Wishbringer
>Keywords: article Wishbringer Moriarty AmigaWorld
>Date: Tue, 5 Oct 93 13:00:49 GMT

>Too bad they *did* in fact end up doing "Zork with pictures." Not
>only did we get Zork worth Zero, we also got the new thing.

FOO! Zork Zero is great fun! It's not that there is a graphic awaiting
you at every turn! Most of it is plain text, not a picture in sight!


Hans

(gu...@stpc.wi.leidenuniv.nl)

Richard Saito

unread,
Oct 8, 1993, 5:49:12 PM10/8/93
to
In article <28s4mr...@life.ai.mit.edu> d...@ai.mit.edu writes:
>In article <1993Oct5.1...@trl.oz.au> kda...@titan.trl.OZ.AU (Kirk Davies) writes:
>
>(Actually, Brian Moriarty writes:)
>
>>BM: Why aren't all books illustrated? [Pausing for effect] Should we
>>succumb to the temptation to throw in lots of cartoony pictures and
>>special effects just because the hardware is capable of it?
>
>Evidently, yes.
>

Shouldn't the hardware be capable of running bigger and better parsers as well?


--
-Richard

David Baggett

unread,
Oct 9, 1993, 10:39:03 PM10/9/93
to

>Shouldn't the hardware be capable of running bigger and better parsers as
>well?

Sure, and people (one of whom works 5 feet away from me :) have written
some extremely good, fast parsers. The problem is how to integrate better
parsing, and the knowledge representation required to support it, into a
*game*. It already takes many, many hours to write a decent IF game, and
if you make the parser take a larger subset of English, you'll
correspondingly increase the amount of uninteresting work that needs to be
done. (Unless you do text generation, and that technology is far, far
from being production quality.)

E.g., I already don't want to have to code a clever description for every
noun in a room description, even if most things are not important to the
plot, but to make the game really support the parsing (limited as it is), I
must do that. The fact is that a game is a *game*, and it has a plot (if
it's good) -- and that means that most of the things that the players will
try are irrelevant and unimportant. This is the same thing I harp on when
people espouse the virtues of highly nonlinear IF -- sure, it might be neat
(though it might be incredibly frustrating or boring, too), but YOU code a
whole 75-100 room game like that. As with any long-term project, half the
battle with writing IF games is just stamina.

Imagine the amount of work you'd have to do to really support a parser that
could handle things like

>tightly tie the cord to the stubby end of the stick using a tight slipknot

and

>who was it that just came in, looked me in the eye, and said,
"Howdy, Y'all?"

The original comment, which was Moriarty's "why should we use graphics just
because they're there," is annoying because it just shows that all that
talk about "the best graphics are in your head" was hype, and nothing more.
Though I don't blame anyone for going to a graphics and sound format (I
write graphics games too), it sucks to realize that what most people at the
time thought was a set of "higher goals" in the gaming world was really
just marketing swill. Keep in mind that Infocom did NOT fold because
people stopped buying text-only games -- they crashed and burned because
Cornerstone was such an incredible (and unexpected) failure.

Stu Galley

unread,
Oct 13, 1993, 6:17:37 PM10/13/93
to
In article <297sk7...@life.ai.mit.edu> d...@ai.mit.edu writes:

: In article <1993Oct8.2...@beaver.cs.washington.edu> rcs...@cs.washington.edu (Richard Saito) writes:
:
: >Shouldn't the hardware be capable of running bigger and better parsers as
: >well?
:
: Sure, and people (one of whom works 5 feet away from me :) have written
: some extremely good, fast parsers. The problem is how to integrate better
: parsing, and the knowledge representation required to support it, into a
: *game*.... [deleted for brevity]

Well said, and in fact the parser that we used in Infocom's Zork Zero
and Shogun was bigger and better than in our previous titles, a
complete rewrite based on linguistic principles instead of a clever
hack. If Activision had not shut down Infocom in 1989, we might have
put some very interesting capabilities into Shogun.

: The original comment, which was Moriarty's "why should we use graphics just


: because they're there," is annoying because it just shows that all that
: talk about "the best graphics are in your head" was hype, and nothing more.
: Though I don't blame anyone for going to a graphics and sound format (I
: write graphics games too), it sucks to realize that what most people at the
: time thought was a set of "higher goals" in the gaming world was really

: just marketing swill....

I don't follow your logic here. For one thing, I wouldn't take
Moriarty's words for Infocom's marketing strategy, let alone its
corporate policy. For another, the "graphics entertainment product"
that he wrote about in the middle of 1985 is probably not a graphical
adventure game, but rather the computerized board game Fooblitzky.
Graphics and sound in Infocom adventure games were more or less
dictated by Activision, its parent company, starting in 1988.

--- stu <gal...@think.com>
Infocom fovnder

Neil K. Guy

unread,
Oct 14, 1993, 3:11:24 AM10/14/93
to
gal...@think.com (Stu Galley) writes:

>Well said, and in fact the parser that we used in Infocom's Zork Zero
>and Shogun was bigger and better than in our previous titles, a
>complete rewrite based on linguistic principles instead of a clever
>hack. If Activision had not shut down Infocom in 1989, we might have
>put some very interesting capabilities into Shogun.

Just out of curiosity, how did the last Infocom interpreter differ
from previous interpreters? Did it maintain the general "verb - direct
object - ( preposition - indirect object )" kind of structure or did
it take a different tack altogether?

- Neil K. (n_k...@sfu.ca)

David Baggett

unread,
Oct 14, 1993, 10:43:59 AM10/14/93
to
In article <29huq1...@early-bird.think.com> gal...@think.com (Stu
Galley) writes:
>[I wouldn't take]

>Moriarty's words for Infocom's marketing strategy, let alone its
>corporate policy.

I wouldn't either, except for that fact that most (all?) of the people
involved in the creative side of Infocom who are still working on games are
doing graphics games. Moriarty's work for Lucasarts and the whole Legend
gang (Meretzky especially) are clear evidence that the principles only went
as far as the pocketbook allowed. Again, it's not that I think this is
inherently bad; graphics games are *clearly* more lucrative today and "ya
gotta pay the rent." It's just that we were told again and again by these
people that they were restricting their games to an all-text format for *a
darn good artistic reason* -- not just because that's what was raking in
the bucks. It sounds awfully hollow now.

>For another, the "graphics entertainment product" that he wrote about in
>the middle of 1985 is probably not a graphical adventure game, but rather
>the computerized board game Fooblitzky.

Well I was mainly referring to later things like "Loom." I'm not saying
that he was being two-faced *when he made the comment*. I'm just saying
that *in retrospect* remarks about the artistic advantages of the text
format seem a bit too *convenient* given the current state of affairs
(which is, as we all know, that everyone has abandonded the text format and
gone to what *sells big*, so we get stuff like "Spellcasting 401 -- Babes,
Babes, and More Babes" instead of Planetfall).

>Graphics and sound in Infocom adventure games were more or less dictated
>by Activision, its parent company, starting in 1988.

If that were the whole story, Moriarty & Co. and Legend would once again be
writing text games, now that they're free from Activision's scummy
influence. But it's not the whole story, and neither is "the art thing"
that Moriarty (and so many other Infocom people and ads) alluded to.

It's really not a big axe I have to grind here -- when it comes down to it,
I don't *care* what the former Infocom people are doing, and their loss of
faith in text games is simply an opportunity for others who haven't
(entirely, at least). But the quote from the interview with Moriarty
seemed particularly ironic here in 1993.

Nathaniel D. Daw

unread,
Oct 14, 1993, 11:56:47 AM10/14/93
to
In article <29jojf...@life.ai.mit.edu>,
David Baggett <d...@ai.mit.edu> wrote:

>Well I was mainly referring to later things like "Loom." I'm not saying
>that he was being two-faced *when he made the comment*. I'm just saying
>that *in retrospect* remarks about the artistic advantages of the text
>format seem a bit too *convenient* given the current state of affairs
>(which is, as we all know, that everyone has abandonded the text format and
>gone to what *sells big*, so we get stuff like "Spellcasting 401 -- Babes,
>Babes, and More Babes" instead of Planetfall).

Actually -- I am in complete agreement with you about Loom and all of the
Kings' Quest variants. They seem to me a large jump from the old Infocom ideal;
and indeed Kings' Quest was around when Infocom was still viable and Infocom
wisely never attempted to borrow that sorry concept. In those games, graphics
_replaces_ text.

But the Spellcasting series, as far as I'm concerned, represents the closest
thing to the old Infocom available today. The graphics are completely
superfluous -- you can turn them off without affecting the game -- and the
real heart of the games are puzzle-oriented word adventures very much along the
lines of Sorcerer and Spellbreaker; or probably more appropriately, Leather
Goddesses of Phobos -- which explains the somewhat sophomoric focus that you
seem uncomfortable with. None of this seems to me contrary with what Infocom
originally stood for.

Sure, these people were saying that you don't need graphics to have a good
game, and now they're making graphical games. I'm the first to agree that
that's hypocritical But at least Meretzky, unlike the rest, is making graphical
games that retain all of the advantages -- and I would argue, the spirit -- of
the text format.

:::::::::::::::::::::::::::::::::::::::::::::::::::::::Cette vie est un hopital
:::::::::::::::::::::::::::::::::::::::::::::::::::ou chaque malade est possede
:::::::::::::::::::::::::::::::::::::::::::::::::::::du desir de changer de lit
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::Baudelaire
::::::::::::::::::::::::::::::::Nathaniel Daw::::::::::::::::::::::::::::::::::
::::::::::::::::::::::::::::::nd...@columbia.edu::::::::::::::::::::::::::::::::

David W. LeCompte

unread,
Oct 14, 1993, 6:05:32 PM10/14/93
to
In article <29jsrv$o...@apakabar.cc.columbia.edu>,

Nathaniel D. Daw <nd...@bonjour.cc.columbia.edu> wrote:
>
>Sure, these people were saying that you don't need graphics to have a good
>game, and now they're making graphical games. I'm the first to agree that
>that's hypocritical But at least Meretzky, unlike the rest, is making graphical
>games that retain all of the advantages -- and I would argue, the spirit -- of
>the text format.

I agree with your perspective on the spellcaster series, from what
I've seen of it.

I disagree that they are/were being hypocritical, however. I believe
that there isn't enough of a market for Infocom's old style games at
the prices that Activision would charge for them. Hence Infocom no
longer can publish text adventures.

Likewise Legend, and any other company; My feeling is that to sell the
number of units that the marketing types want to move, you need to
follow the technology curve. Which means using pictures because you
can.

It's unfortunate, but that's the dynamics of shrinkwrap software.

It's a good thing that lower overhead avenues exist for selling text
adventures, such as shareware. The smaller markets allow for more
specialized-interest products to be produced, even if they don't get
packaged in a beautiful box with expensive art on the front.

Just my 2 cents...

-Dave LeCompte


--
These opinions are my own, and do not necessarily reflect
those of my family, my company, my race, or my species.

Mark 'Mark' Sachs

unread,
Oct 15, 1993, 2:23:06 AM10/15/93
to
In article <CEwqp...@mv.mv.com>, tsma...@mv.mv.com (David W. LeCompte) says:
>I disagree that they are/were being hypocritical, however. I believe
>that there isn't enough of a market for Infocom's old style games at
>the prices that Activision would charge for them. Hence Infocom no
>longer can publish text adventures.

Frankly, I wonder about that.

The _concept_ of text adventures has vanished from the marketplace. I'd
agree that it would be difficult and probably impossible to re-inject it.
But if Infocom hadn't gone released Cornerstone (which, I hear, is what
really did them in) they would have certainly lasted longer. And with
games like Zork Zero pointing to the future, I think they could have
carved out a comfortable niche (supported by loyal fans and generally
anyone who was partial to a good adventure) and survived there basically
forever.

But I guess we'll never know, will we.

Excellent day to work off excess energy. Steal something heavy.
[Your blood pressure just went up.] Mark Sachs, aka mbs...@psuvm.psu.edu
DISCLAIMER: Penn State cares about my money, not my opinions.

Mathematical Institute, (0865) 2-73525

unread,
Oct 15, 1993, 8:56:30 AM10/15/93
to

I'd like to draw attention to something else Mr Moriarty wrote, in passing:
that the time and effort spent on getting passably-good graphics could be
spent better on the text.

In particular, on thoroughly debugging a game. It's only as a result of
writing "Curses" that I've realised just how much of a creative process this
is. Since the first "working" version, several hundred bugs have been
reported to me - most of them by the original testers, but another eighty
or so by players - and fixing them has usually meant adding code for odd
things to do. Often these bugs are comments to the effect that a
message like

You drop the flaming torch, and start a fire which...

is unfortunate when you drop the torch in, say, the middle of a river.
Fixing the bug means trapping this with a message like, say,

You drop the flaming torch, which snuffs out and rushes away in the current..

and these sort of messages definitely add something to the game, even
though they make no effective difference to play.

Debugging also means providing alternative solutions to problems, giving
amusing messages for common blunders by players... it is not simply a matter
of fixing accidents like

You take the fjhgkjdfhg.


Graham Nelson
Oxford University, UK

Hans Guijt

unread,
Oct 15, 1993, 9:39:07 AM10/15/93
to
In article <93288.022...@psuvm.psu.edu> Mark 'Mark' Sachs <MBS...@psuvm.psu.edu> writes:
>But I guess we'll never know, will we.

But there were many more Text Adventure companies. Where are they now?

Agreed, Infocom were the best. But Magnetic Scrolls wasn't at all bad, and
they aren't here anymore. And what happened to Level-9? And all those others?

Hans

(gu...@stpc.wi.leidenuniv.nl)

David Baggett

unread,
Oct 15, 1993, 12:08:54 PM10/15/93
to
In article <1993Oct15.1...@vax.oxford.ac.uk> nel...@vax.oxford.ac.uk (Mathematical Institute, (0865) 2-73525) writes:
>In particular, on thoroughly debugging a game. It's only as a result of
>writing "Curses" that I've realised just how much of a creative process this
>is.

What he said. This is so true. Of all the kinds of software projects I've
ever worked on, IF games seem to require the most testing. Dijkstra would
of course say that this means we're just not developing them correctly, but
how can I predict that someone's going to think to try to burn down the
*outhouse* (in UU2)? :)

Jorn Barger

unread,
Oct 15, 1993, 12:56:39 PM10/15/93
to
David Baggett writes:
> [...] Of all the kinds of software projects I've

>ever worked on, IF games seem to require the most testing. Dijkstra would
>of course say that this means we're just not developing them correctly, but
>how can I predict that someone's going to think to try to burn down the
>*outhouse* (in UU2)? :)

I'm not clear on this. Is it that (1) when they type "burn outhouse" they
get a boring default message, or (2) they get an *illogical* default
message, or (3) something worse? (It's not like it hangs or bombs,
right? :^)

Is it that they're thinking of solutions that ought to work, but weren't
anticipated, and this spoils the illusion of authorial ominscience?
(Is there any such illusion?)

It sounds like there's a general, 'brute-force' solution-- to examine
every combination of noun and verb (oops, and *context*-- yikes...)
...maybe the long-term solution is to have a software tool that uses
a knowledgebase of common stories (eg, "buildings can be burned down")
to *predict* potential combinations?

jo...@mcs.com

R. N. Dominick

unread,
Oct 15, 1993, 12:45:09 PM10/15/93
to
Hans Guijt (GU...@stpc.wi.LeidenUniv.nl) wrote:

: Agreed, Infocom were the best. But Magnetic Scrolls wasn't at all bad, and

: they aren't here anymore. And what happened to Level-9? And all those
: others?

...Level-9 changed *into* Magnetic Scrolls. Proof? Some of the games MS
later released were direct cribs of old Level-9 games (I got a whole bunch of
'em on the Firebird label for $1.29 a few years back. Anyone ever finish
any of the 'Snowball 9/Return to Eden/Worm in Paradise' games? Kinda
interesting non-topological topology, if you follow me.)

I think that if Legend would publish something other than Spellcasting games,
they'd get a lot less flack. The Heechee games were kind of fun, but none of
their games come close to, say, Spellbreaker...

--

R. Dominick / cinn...@kaiwan.kaiwan.com / einstein @ qedbbs.com
"My thoughts are snakes, they play all the games..." -- hnia

David Baggett

unread,
Oct 15, 1993, 9:52:50 PM10/15/93
to
In article <29mko7$4...@Notwerk.mcs.com> jo...@Notwerk.mcs.com (Jorn Barger) writes:
>I'm not clear on this. Is it that (1) when they type "burn outhouse" they
>get a boring default message, or (2) they get an *illogical* default
>message, or (3) something worse? (It's not like it hangs or bombs,
>right? :^)

Before that possible behavior is pointed out to the author (at least
for a TADS game), you'd likely get

Pouring the kerosene on the outhouse doesn't seem productive.

or

I don't know how to pour the kerosene on the outhouse.

or, even worse

Done.

The game certainly wouldn't bomb; it would just say something stupid.

>Is it that they're thinking of solutions that ought to work, but weren't
>anticipated, and this spoils the illusion of authorial ominscience?

Right.

>(Is there any such illusion?)

I always *hope* to create that illusion. This is of course modulo the
assumption that the player is somewhat clued in to the genre and won't be
let down when things like

>tell me who that guy who just came in the window is

don't work...

>It sounds like there's a general, 'brute-force' solution-- to examine
>every combination of noun and verb (oops, and *context*-- yikes...)
>...maybe the long-term solution is to have a software tool that uses
>a knowledgebase of common stories (eg, "buildings can be burned down")
>to *predict* potential combinations?

The problem with the brute force (e.g., naive physics) approaches is that
they tend to force you down text generation avenues, and prose quality of
generated text isn't good enough for a mainstream game yet.

I view these things as works of fiction foremost and simulations second.
Text generation is incompatible with this ideology unless it's really
spectacular, and we're far from having even passable text generation right
now.

Jorn Barger

unread,
Oct 16, 1993, 12:26:01 AM10/16/93
to
David Baggett <d...@ai.mit.edu> wrote:
>The problem with the brute force (e.g., naive physics) approaches is that
>they tend to force you down text generation avenues, and prose quality of
>generated text isn't good enough for a mainstream game yet.

Oops, I probably have a peculiar way of using the term 'brute force'--
my image was of someone going thru the combinations one by one, on
paper, and making the literary, mental effort to think how that bare
combination might be infused with meaning and life and poetry...

(A few times when I've strained to write poetry with great care,
it's seemed like *really* hard work-- like I end up ravenous with
hunger!)

jorn


David Baggett

unread,
Oct 16, 1993, 1:40:47 PM10/16/93
to
In article <29nt4p$4...@genesis.MCS.COM> jo...@genesis.MCS.COM (Jorn Barger) writes:
>Oops, I probably have a peculiar way of using the term 'brute force'--
>my image was of someone going thru the combinations one by one, on
>paper, and making the literary, mental effort to think how that bare
>combination might be infused with meaning and life and poetry...

That would be an awful lot of cominations to do on paper. My current
project has over 800 objects (in the TADS programming sense) and probably
more than 100 verbs.

It might be useful, however, for authors to make a pass through their games
once they're done, looking at important decorations and items while
simulatanesouly eyeballing a verb list. Many of the more obvious
interactions might pop right out. Then again, maybe not. :)

Jorn Barger

unread,
Oct 16, 1993, 11:27:03 PM10/16/93
to
David Baggett writes:
>It might be useful, however, for authors to make a pass through their games
>once they're done, looking at important decorations and items while
>simultaneously eyeballing a verb list. Many of the more obvious

>interactions might pop right out. Then again, maybe not. :)

A verb list, and also a list of carryable items, right?

Mathematical Institute, (0865) 2-73525

unread,
Oct 17, 1993, 11:00:30 AM10/17/93
to
In article <29mhum...@life.ai.mit.edu>, d...@case.ai.mit.edu (David Baggett) writes:
> In article <1993Oct15.1...@vax.oxford.ac.uk> nel...@vax.oxford.ac.uk (Mathematical Institute, (0865) 2-73525) writes:
>>In particular, on thoroughly debugging a game. It's only as a result of
>>writing "Curses" that I've realised just how much of a creative process this
>>is.
>
> What he said. This is so true. Of all the kinds of software projects I've
> ever worked on, IF games seem to require the most testing. Dijkstra would
> of course say that this means we're just not developing them correctly, but
> how can I predict that someone's going to think to try to burn down the
> *outhouse* (in UU2)? :)
>
> Dave Baggett

Quite. I have just coded "not being able to smell anything while wearing the
gas mask". In six months of play, I think only one player has ever tried to
do this. But say every player has, oh, a 1/1000 chance of trying it. There
must be getting on for 1000 other dumb and unlikely things to try, so every
player will find at least one of your clever bits of coding.

It's also quite appealing to think that some of the responses my game can
produce have still been seen by nobody. With the original mainframe Adventure,
there did come a point when you thought you'd seen every message it could
produce, and you understood every rule it applied.

Since we were talking about the decline of text adventures, I think another
reason is that they were getting unmanageably large. The Zork 1-size of game
was dictated by the size of a Tandy disc drive, as I understand it. Now
a modern computer laughs at the size of an Infocom game (Zork 1 can run
something like 15 times concurrently in my machine). So since we have much
more room, why not make games 15 times huger?

The answer is that complexity for the player and so the designer increases by
a lot more than a factor of 15. Even if Infocom hadn't gone bust, could they
really have sustained making games of the complexity of Trinity and
A Mind Forever Voyaging for very much longer? And they were only twice as
large. Even if the product was twice as good, I imagine the sales were about
the same, so they must have made a lot more out of, say, Enchanter.

Personally, I think the middle-period Infocom size of game (say, 70 locations
and 70 portable objects) is about right. More does not mean better. And that
fits into maybe 200K of machine at the outside.

Against all this, graphics
(a) eat up memory like there's no tomorrow, so you feel you're really using
the machine,
(b) can be put in just by hiring an artist, reducing the need to find an
author who can actually write English as well as code,
(c) reduce the amount of interaction possible with the game, greatly
simplifying it and cutting down overheads in production.

Graham Nelson
Oxford, UK

Nathaniel D. Daw

unread,
Oct 17, 1993, 1:44:30 PM10/17/93
to
In article <1993Oct17.1...@vax.oxford.ac.uk>,

Mathematical Institute, (0865) 2-73525 <nel...@vax.oxford.ac.uk> wrote:
>The answer is that complexity for the player and so the designer increases by
>a lot more than a factor of 15. Even if Infocom hadn't gone bust, could they
>really have sustained making games of the complexity of Trinity and
>A Mind Forever Voyaging for very much longer? And they were only twice as
>large.

One solution to this increasing complexity is to make the game more "linear,"
which we normally think of in a bad sense, but can be a good thing if not taken
too far. In AMFV, for instance, there were several distinct episodes that had
to be completed in sequence. Each episode in itself was non-linear; that is,
you didn't feel like you were locked into a single path through the game, but
on the other hand neither you nor the designers had to face everything
at once. The Legend folks seem to have followed this same concept to make games
that seem, even considering the graphics, much bigger than, say, Trinity -- in
Spellcasting 101, you have to go to a series of islands, in no particular
order, each of which is something like a little distinct game. And in
Spellcasting 201, you get new quests each day, which open up different areas
of the game, etc. Making a large game out of a bunch of more manageable smaller
games makes sense to me.

nd

Jorn Barger

unread,
Oct 17, 1993, 2:39:50 PM10/17/93
to
Graham Nelson writes:
>It's also quite appealing to think that some of the responses my game can
>produce have still been seen by nobody.

Is this a desirable thing, from the player's point of view? In my
"Issues in IF" survey (available by ftp, see sig below) I wrote:
> The 'finished-it' paradox: The first time you reach 'The End' you will
>not have finished reading all the author's text. Are you then expected to
>go back and exhaust every alternate path, to 'get your money's worth' or to
>express your pleasure in her craft? Can an author keep this interesting?
>(With book-tech, you can just reread it cover to cover, although you'll
>miss some of the subtler meanings this way.)

How do *players* feel about the idea that there's jokes in a game that
they may never discover? (I'm thinking of how, when I really like
an author, I make a huge effort to read *everything* they've written.)

Graham again:


>a modern computer laughs at the size of an Infocom game (Zork 1 can run
>something like 15 times concurrently in my machine). So since we have much
>more room, why not make games 15 times huger?

There's a wonderfully logical resolution of this complexity problem:
make the *interpreter* 15 times huger, by including a detailed knowledge-
base of typical realworld interactions! Encyclopedia TADS-anica, if
you will... This would have to play an active role during game development,
making intelligent, *automated* suggestions about what combinations might
need special attention at each point.

=----------=- ,!. --=----=----=----=----=----=----=----=----=----=----=
Jorn Barger j't Anon-ftp to genesis.mcs.com in mcsnet.users/jorn for:
<:^)^:< K=-=:: -=-> Finnegans Wake, artificial intelligence, Ascii-TV,
.::.:.::.. "=i.: [-' fractal-thicket indexing, semantic-topology theory,
jo...@mcs.com /;:":.\ DecentWrite, MiniTech, nant/nart, flame theory &c!
=----------= ;}' '(, -=----=----=----=----=----=----=----=----=----=----=

David Baggett

unread,
Oct 17, 1993, 2:20:25 PM10/17/93
to

I actually meant that you'd be looking through the *entire* code (which
includes carryable items) while looking at the list of verbs. Carryable
items are necessary but not sufficient here -- you need to look at fixed
items as well. The outhouse in UU2, for example, is not a carryable item
-- it's a decoration. But to catch the bug I mentioned before I'd have had
to have looked at the outhouse code while looking through the verb list.

Here's the scenario of how it would work:

1. I look through all the UU2 code, object by object.
2. I look at the outhouse
3. I look at the verb list
4. I notice "burn"
5. I think "burn + outhouse" ... Hmmmm.
6. I add code to handle this case.

Again, I have no idea whether this is actually workable in practice. You
could very well end up sitting there slaving away tediously to no avail
whatsoever. We should look at attempts to mechanize creativity with a
great deal of suspicion, after all...

Philip Stephens

unread,
Oct 18, 1993, 2:13:30 AM10/18/93
to
Jorn Barger writes:

>The 'finished-it' paradox: The first time you reach 'The End' you will
>not have finished reading all the author's text. Are you then expected
>to go back and exhaust every alternate path, to 'get your money's worth'
>or to express your pleasure in her craft?

It's part of the nature of Interactive Fiction that a given player will
not try every possible action, and hence won't get to see every clever
comment and witty phrase. I don't think this is a problem, really.
Different players will get something different out of the experience of
playing the same game, and that is good.

However, in regards to the size and complexity of IF games: I'd prefer to
play a game with few locations and objects, but with attention paid to the
fine details of the interaction of the objects with their enviroment, than
to play a game with hundreds of rooms and items, but which has only generic
"You can't do that" responses for most of the actions you try to perform.
For it's the atmosphere that makes a game, not it's size.
For example, _Planetfall_ has a reasonably large map, but most of it is
empty of interesting things to look at, and the game itself involves very
few puzzles overall. Add to that the fact that half of the objects are
decoys that don't actually do anything, and you have an example of an IF
game that doesn't provide much in the way of real enjoyment. (Apologies to
those who liked Planetfall!)
On the other hand, _Suspended_ was a game that had a very small map, but
lots of interesting things to look at, and a lot of detail about the
surroundings and objects. There were often no less than five different
descriptions of the one object, and many of the responses were quite humorous. (Apologies to those who hated Suspended!) :-)

Personally, I find IF games which have large sprawling maps to be a bit
of a pain, because if there's one thing I hate, it's having to single-step
through a dozen rooms to get to where I want to go. I suppose that can be
solved by having an intelligent game that can plot a path to rooms you've
visited before when you type "go to such-and-such a room", but I prefer games
that limit you to a smaller section of the map until you've solved a couple
of puzzles, then don't expect you to return to the completed sections. _Zork_
is not a bad example of this, actually: at any one time, you had two or three
puzzles to solve in order to unlock other portions of the underground empire,
but you generally didn't have to do a lot of walking around.
Furthermore, games that expect you to draw up complicated maps (or redraw
them when you find yourself running out of space) are also a bit tedious,
in my view. I'd prefer a simple map laid out on a grid, and have the authors
spend most of their time on "set dressing, props, and costume design". There
*is* a certain joy in plotting out a map as you explore a game's boundaries,
but I can't stand rooms that exit from the south and enter the next room from
the east! For it's those kinds of tricks that have forced me in the past to
spend 75% of my playing time moving back and forth through rooms until I got
the exits all worked out right! It's okay to do this every once in a while
in obvious places (such as a maze whose twisty little passages look all alike),
but for the remainder of the landscape I'd prefer simple compass directions
and equal-sized rooms (or rooms made out of multiple locations).

Actually, for many years I've had an IF game on the drawing board that
addresses all of these objections of mine, so I guess one of these days I
ought to get around to writing it :-)


>There's a wonderfully logical resolution of this complexity problem:
>make the *interpreter* 15 times huger, by including a detailed knowledge-
>base of typical realworld interactions! Encyclopedia TADS-anica, if
>you will... This would have to play an active role during game development,
>making intelligent, *automated* suggestions about what combinations might
>need special attention at each point.
>
>=----------=- ,!. --=----=----=----=----=----=----=----=----=----=----=
>Jorn Barger j't Anon-ftp to genesis.mcs.com in mcsnet.users/jorn for:
> <:^)^:< K=-=:: -=-> Finnegans Wake, artificial intelligence, Ascii-TV,
> .::.:.::.. "=i.: [-' fractal-thicket indexing, semantic-topology theory,
>jo...@mcs.com /;:":.\ DecentWrite, MiniTech, nant/nart, flame theory &c!
>=----------= ;}' '(, -=----=----=----=----=----=----=----=----=----=----=
>

--
| Philip Stephens, Systems Programmer. | "Many views yield the truth. |
| Address: 43 Malcolm Road, Braeside, | Therefore, be not alone." |
| Victoria, 3195, AUSTRALIA. | |
| Internet: phi...@labtam.labtam.oz.au | -- Prime Song of the Viggies |

Mathematical Institute, (0865) 2-73525

unread,
Oct 18, 1993, 8:07:30 AM10/18/93
to
In article <29s3hm$e...@Notwerk.mcs.com>, jo...@Notwerk.mcs.com (Jorn Barger) writes:
> Graham Nelson writes:
>>It's also quite appealing to think that some of the responses my game can
>>produce have still been seen by nobody.
>
> Is this a desirable thing, from the player's point of view? In my
> "Issues in IF" survey (available by ftp, see sig below) I wrote:
>> The 'finished-it' paradox: The first time you reach 'The End' you will
>>not have finished reading all the author's text. Are you then expected to
>>go back and exhaust every alternate path, to 'get your money's worth' or to
>>express your pleasure in her craft? Can an author keep this interesting?

I agree up to a point. Without giving too much away, if you win "Curses"
you're given some suggestions as to other things you might have done, so
if you want you can go and look. But isn't this like making a film, in a
way? Everybody who goes to see "Batman Returns", say, will get the plot
(such as it is). How many will notice a particularly nice sculptured door
which is only in shot on the far left of the screen for a couple of seconds?
Answer, few people, but that doesn't mean it shouldn't be there.

>
> Graham again:
>>a modern computer laughs at the size of an Infocom game (Zork 1 can run
>>something like 15 times concurrently in my machine). So since we have much
>>more room, why not make games 15 times huger?
>
> There's a wonderfully logical resolution of this complexity problem:
> make the *interpreter* 15 times huger, by including a detailed knowledge-
> base of typical realworld interactions! Encyclopedia TADS-anica, if
> you will... This would have to play an active role during game development,
> making intelligent, *automated* suggestions about what combinations might
> need special attention at each point.
>

Mmm... again, this is the simulation problem. If a game totally simulates
physics, then you aren't surprised to find you can make a balanced pyramid
on the table out of all the things you're carrying. If you find you can do
it in an IF game, you know at once that some problem will be related.

I actually think interpreters are about as good as they need be. Parsers
understand pretty readily anything that's reasonable, these days.

Graham Nelson

Reply all
Reply to author
Forward
0 new messages