[leveller] Re: Digital Surf
- From: Ray Gardener <rayg@xxxxxxxxxxxxxxxxxx>
- To: leveller@xxxxxxxxxxxxx
- Date: Sat, 29 Sep 2007 07:52:51 -0700
Carl Lingard wrote:
Great news! Leveller has often been my preferred way to visualise
terrain models or wips. It seems to be the way other terrain
modellers are going too.
Relatedly , what would be super useful is a way of moving between
terrain models and the final renderer in a much more managed way.
We've talked about this on ME-DEM. we need more streamlined ways to
produce textures more directly from the terrain- obviously flow maps
are the basis of this at present. A heightmap is only the terrain
half-finished- and if it's a very large heightmap of varying
geographies (and even latitudes), that's going to be one large
texture requiring a lot of work and/ or thought. The renderers are
going to have procedurals, but, unless you can find a way to drive
those procedurals, across renderers, then our best met is to stick
with maps right?
I don't see maps going away any time soon. Compression mathematics limit
how much interesting content a procedural can express.
"Infinity of patterns" for a function is misleading too -- sure, the
Mandelbrot set, e.g., contains an infinite number of images, but this is
a mere subset of the larger infinity of all possible images. If you
picture the set of all possible images as a plane, then Mandelbrot is a
line on this plane; infinite along one axis but narrowly finite in
another. You can add more functions but all you do is plot more of these
thin lines across the plane. It's too meager a diet for the average person.
Well it's the second life thing isn't it? :) I do go for this and I
know others who do too. The main reason I buy computer games is to
immerse myself in the enviroment (and to see the state of play of
technology)- I often don't complete the games these days. Looking at
the Typhoon engine for eg, that really is very cool and extremely
pleasant to amble about in because it favours simulation rather than
gameplay. Having looked into a few mmorpgs, I know that a lot of
people talk about building houses in the back of beyond. But then
we're back to your stated reservation.
The obvious issue is boredom -- what happens after the novelty of
walking around wears off. This essentially means incorporation of
dynamic elements, and ultimately sharing the view with other people.
There are of course those who find it entertaining for long periods of
time. Model railroads, for example -- if you're into it, you're probably
really into it, and if you're not, you find it puzzling how anyone could
be. But people who are into it will explain that it's not just the play,
it's the constructing. This is probably why SL works fairly well, it
engages people on multiple key levels: building, sharing, viewing,
playing, the first two of which keeps the world always novel. Any one of
these attributes would suffice for some people, but their combination
gives SL the audience it has. In a sense, SL uses people as a form of AI
procedural.
I'm also looking at Second Life scripting, to help spruce up SL
sims.
What kind of scripts would you write?
I'm still getting a feel for it. It's hard to say, but mainly things
that make the system more realistic. If I see one more bird or fish
swimming in a purely circular path, I'll vomit. :)
Have you been following the recent convos on ME-DEM Ray ?
http://www.me-dem.org/component/option,com_fireboard/Itemid,43/func,view/id,2490/catid,4/
We're looking into a real time viewer. Why not get involved more
directly with what we're planing to do? I know that RedRobes is keen
to develop on an open source terrain viewer and in fact integrate his
geomorpher with texture output.
Interesting. It's not really enough of a domain overlap for me to get
into rendering, and rendering is really only part of the issue. You need
to figure out what it is you want to do with your content and how you
want people to interact with it. If you want some kind of shared
community similar to a multiplayer game or SL, then you need to plan for
a server infrastructure, how the bills will be paid, etc.
Google is rumored to be coming up with an SL clone, something about
adding avatars to Google Earth. You may want to wait to see how that
turns out and if you can borrow from it.
One project I always have in the back of my mind is an educational
history tool that lets people visualize the past, for some interesting
period of related events such as World War II. I would mod WhirlWind and
add 3D animations showing (in as much detail as one can) all the known
troop and materiel movements in that timeframe. Then people could really
have a much deeper, personal understanding of what it was like. That
amount of content would need many people to make, so it'd be like
Wikipedia, but vetted since obviously historical accuracy is crucial.
There'd be LOD management as well so, for a large area, one person could
paint a movie clip texture showing overall troop movements using fat red
arrows or whatever, and later when more people had the details they
could fill in the higher LOD with individual soldiers, vehicles, etc.
The highlight, of course, is modeling the two atomic bombs going off at
the end. :) But if you do it right, the entertainment value is eclipsed
by the sense of tragedy -- you'd also see the devestation and all these
people running around and dying off.
The system of course can scale up to include more time periods and
detail. Ultimately humans will have a living 3D record of all of their
history, and what we're doing now is laying the groundwork for that.
I'd probably place my avatar near Cape Canaveral in July 1969. It's as
close as I can ever get to actually having been there. :)
Ray
Other related posts:
