Ola Fosheim Grøstad <email@example.com>
Ola Fosheim Grøstad <firstname.lastname@example.org>
Tue Aug 7 11:49:56 New Zealand Standard Time 2001
Brian Hook wrote:
> 10000 * 10000 * 4 bytes = 400MB of uncompressed data for a
> relatively small world with no non-terrain content. Some people
> want HUGE worlds, on the order of hundreds of kilometers by
> hundreds of kilometers. And, of course, you want 1000 worlds.
I certainly don't want 1000 worlds the size of earth. (well, several
years ago I probably did want that) In a game I want most people to
enjoy 50%+ of the content and like to see the energy spent on
increasing that quality. I'd rather have one world that is actually
I find AOs world big enough. It will take you hours upon hours at
full speed seeing it all... The trouble is: the fun that provides
is limited. Making it bigger will not improve it at all! Putting
more play related content in the same space would.
Besides, I think you want to grow the world by adding new content.
I think the fact that you add content is just as important as the
fact that there is a lot of content. It's like fashion and
politics. "this is fresh", "this is the latest", "new stuff to
explore", "I am going to be the first one", "they care, this world
sux, but it is improving", "at least they are trying", "something
big is happening!".
> For huge, vast, sweeping content, algorithmic creation is
Yes, that is true, and you had that on the Commodore 64. The
novelty of such worlds was great fun.
> There is no choice in this matter. You can try to give designers
> relatively coarse level tools, but at most they'll get to say
> "make a desert planet".
Ack... No romantic spots, eh?
> The two aren't the same at all. Music is not created through
> natural forces -- it's an inherently man made construct. The
> creation of planets, by and large, is not a human activity =) As
> such, algorithmic generation of terrain is an exercise in modeling
> natural phenomena.
You could say that. Although that won't make your games more
interesting than the physical world.
Anyway, music can be said to be an extracted and intensified version
of what you find in nature. A lot of music explicitly is! So, I am
not convinced of your argument.
Humans have evolved to appreciate the nature as it is. If you want
to address that through algorithmic content, you basically have to
reengineer the human mind. Which I guess is what artists do in an
intuitive fashion. You may think that you model nature, but what
you actually is trying to do is to model the world as perceived by a
human. But that is not sufficient, you also need to provide
parameters for your model which spans a perceptionally uniform
space. Which is one of the basic obstacles with going for non-linear
systems in my opinion. Yeah, there are some interesting spots, one
in a billion. A "find the needle in the haystack" type of problem.
I once thought GA/GP would be the optimal solution, but even then
you still face severe trouble, because you do need an advanced model
of the human in order to select. And even when you succeed it is
all surface and no meaning. So basically you need a very very
complex model which covers both perception and experience in order
to generate interesting content. IMO, this is a field for research,
not a field for development :-).
> But your argument is valid for algorithmic generation of human
> content like buildings, objects, quests, etc.
You want the virtual world to be MORE REAL (intense) than the real
world if you go for modelling. As in art. Experience should be
more intense than walking in the mountains. If it isn't compelling,
why would I play?
Just to illustrate what one are competing with:
In the physical world I have access to lakes, beaches, sailing,
fishing, parks, skijumping/skiing, sports fields, forest,
vikingships, a castle, cafes, botanic garden, all kinds of
libraries,galleries/museums/opera/concert/theatre, various sorts of
electronic entertainment, shopping, cafes and restaurants. All
accessible in 30 minutes, most in 15. And most populated with other
human beings (no crowding) and with a reasonable chance of meeting
someone you know.
(I'm not even going to start listing what I can access in 2 hours
time of travel.)
Another argument against fractal content: Look at the movie
industry. Sure, special effects do sell films, but it takes more
and more and more to even make the audience raise their heads to
have a look. Sfx by themselves does not impress the audience
anymore. They expect them, but are not pleasantly surprised by
them. The audience expects the computer to be able to do anything,
it isn't special, it's its "nature". Watching a real magician pull a
rabbit out of his hat, now that is something to ponder upon! ;)
Of course GEEKS think otherwise, but true non-geeks, like my
grandmother, is not at all impressed by computer graphics. She is
however impressed by accounting programs! (perhaps because one
machine replacing all that tedious brain-straining human effort is
against the nature of the world, as perceived)
> This may be so, but to me it seems the two are largely unrelated.
> There is a need to create large quantities of content. At the
> same time there is a need to create a huge amount of interesting
I'm not convinced of this. There are extremely simple game concepts
with no content to speak of that keep hundreds of thousands of
players busy for months. What they need is each other, feeling that
they have a place to fill, that the world is making progress, and
that they as a group are able to keep their heads above the water.
Not to say that content is not important. Among other things,
content is probably perceived as more tangible than friendships and
thus they are willing to pay for it.
For content I expect that the commercial side will eventually go
with the well known content business
model. Buy/license/authorise/distribute content made by others, then
you don't have to pay for the bad stuff. Used with success in most
segments of the entertainment market. (books, movies, TV, consoles)
It will happen when the market is monopolized, I'm sure. In 20
Ola - http://www.notam.uio.no/~olagr/
MUD-Dev mailing list
MUD-Dev at kanga.nu
More information about the MUD-Dev