Newsgroups: rec.arts.int-fiction
Path: news.duke.edu!newsgate.duke.edu!news-hog.berkeley.edu!ucberkeley!news.maxwell.syr.edu!newsfeed.skycache.com!Cidera!news-reader.ntrnet.net!uunet!ash.uu.net!world!buzzard
From: buzzard@world.std.com (Sean T Barrett)
Subject: Re: Tone in IF
Message-ID: <GDH02s.2MD@world.std.com>
Date: Thu, 17 May 2001 08:11:16 GMT
References: <gCSkwIA3HDA7EwAX@ntlworld1.com> <emshort-1705010309420001@user-2inik11.dialup.mindspring.com>
Organization: The World Public Access UNIX, Brookline, MA
Lines: 55
Xref: news.duke.edu rec.arts.int-fiction:87057

Emily Short <emshort@mindspring.com> wrote:
>characters about Y things, there is a range of options defined along two
>axes, X and Y.  This is a basically tractable set of possibilities.  If
>the player can also ask all those questions in Z ways, the puzzle space
>opens out hideously into a third dimension, even if Z is a small number
>(like 3).  It becomes increasingly difficult to be sure one has covered
>all the bases.

I think there is more to it than combinatorial explosion. I believe
Zarf has mentioned this in the context of adverbs (a third axis)
and in the case of 'use object x on object y' instead of
'verb x object y' in graphical adventures.

The reality is that in adventure games we *do* use some
3-dimensional things (there are a number of verbs that
take two objects).  So I think the total number of combinations
*does* matter; the 3-dimensional space with four choices
on each axis may well be as tractable as the 2-dimensional
space with eight on each.

Also, I think this matters less when the task is clued
(you're not randomly searching the space in the first place)
or when the behaviors are simulated (put X in Y, or the
many dimensions of change created by Metamorphoses' toys,
for which one could imagine a single machine with a single
longer command producing the same effect); in the former
case because, of course, you're told how to navigate the
space, and in the latter case because it's predictable--you
know how the third dimension affects the outcome.

So I think the biggest issue is when you're given a random
space to explore, e.g. in many conversations.  As a rule,
as a designer, if you make something seem to be unimportant,
people will think it unimportant; so if you make *certain*
conversations depend on what mood you're in, but the vast
majority don't, you've just made things really hard by
introducing a 3rd dimension that people won't explore for
long.

And that is I think exactly the problem with adverbs as
well--that they don't *usually* matter. The introduction
of an extra axis (whether second or third or fourth) which
is "stylistic" without really creating a new action means
it's (a) unlikely to have an effect most of the time and
(b) people aren't going to want to try multiple variations
of the same command with different styles, since the "same"
thing will keep happening over and over.

SeanB
[While I seem here to be critical of the 2-axis theory, I
think it's a perfectly reasonable design rule-of-thumb;
one axis will almost always be too few to be interesting
(though the authors of the LOTECH comp entries would be
quite right to disagree) and there are clearly diminishing
returns on 3+.]
