This is not directed against the views or intentions of Bill or Carl
or anyone in particular, but is just an observation:
Let's keep in mind what we're trying to do. Are we writing a piece of
Interactive _Fiction_, or are we aiming at creating a realistic simulation
of reality (people of the "simulationist" school would of course
answer yes to both questions)?
If we're writing interactive _fiction_, then the important thing
is not that liquids behave just as in real life, but that they fill a
purpose in the narrative. We may have a puzzle that involve liquids, or
we may want to build a magic system on mixing potions, or maybe
our hero is an analytical chemist. All of these situations would call
for more or less realistic handling of liquids in the game.
However, we should not get carried away by the possibilities of
simulation. Suppose, for example, that we're writing something like
the oil puzzle in the original _Adventure_. Then we'll need to be able
to fill a bottle with oil and pour the oil on things. If there are
other liquids in the game (and adventure games always tend to include
one or two rooms that have water in them) we'll need to be able to
fill the bottles with those liquids as well. But do we really need to
be able to mix liquids (OK, so oil and water don't mix in real life -
let's suppose we have some beer as well) or to accurately simulate the
fact that oil will float on water? Wouldn't it be better to spend
that effort on improvingplayability in other aspects?
So, I guess my point is this: before spending a lot of time and effort
on implementing realistic liquids in TADS or Inform, one should ask
oneself the question whether having that would really improve one's
games. The answer may be yes - and in that case one should of course
go ahead. In most cases, I suspect the answer will be no.
I'm faced with a similar dilemma regarding clothes. I would like to
implement a "realistic" clothing class that would keep track of
things like that you don't want to wear two ties, or a skirt
under a pair of tights (while the other way round would be OK).
However, such a system would be quite complicated. So far, I haven't
really seen the need for it. Of course, it would be nice to have, but
I have the feeling that _if_ I were to implement it in a game, then
people would spend a disproportionate amount of time manipulating
their clothing, just to fiond out that it didn't really matter to the
plot.
On the other hand, I have some ideas for puzzles involving clothing.
WHen I get around to implementing those, I might want
reasonably realistic behaviour for the clothes. For example, consider
the following situation:
> i
You are carrying a five-cent postage stamp, a copy of "ALice in Wonderland",
a sixpack of beer, a pair of khaki trousers, and a pair of jeans (being worn).
> s
The doorman says something about your jeans being "leisure wear" and
not acceptable inside the restaurants, and blocks your way.
> take of jeans and wear trousers
OK.
> s
The doorman lets you past, saluting smartly.
This would probably be acceptable to most users. However, it does raise
some questions: what about your other clothes? Are you wearing just a pair
of jeans and nothing else? Won't the doorman have any views on your shirt?
And what happens if you take off your jeans and forget to put on the
trousers? Then the doorman will probably say something about your not
being decently dressed - but then, somehow, the illusion of the only
important items of clothing being your pants breaks down, at least IMHO.
The user might start wondering: "What about underwear? Isn't he wearing any
shirt?" and so on.
So one might consider making a ralistic implementation of clothse. But
then one has other problems: Basically every stitch of clothing worn
by the player has to be implemented. How does one stop such a system
form being overly complicated, and from taking up an excessively large
portion of the player's attention?
If the _only_ purpose of clothing in the game is the doorman puzzle,
then at least I would hesitate to introduce a complicated "clothing
simulator". Maybe I would be tempted/inspired to add a whole lot
of clothing-based puzzles, just to justify such an introduction? That
may, or may not, be a good idea...
Magnus Olsson (mol@df.lth.se) / yacc computer club, Lund, Sweden
Work: Innovativ Vision AB, Linkoping (magnus.olsson@ivab.se)
Old adresses (may still work): magnus@thep.lu.se, thepmo@selund.bitnet
PGP key available via finger (to df.lth.se) or on request.