On 1/18/2012 10:19 AM, Devon D Sparks wrote:
There's a trend in architecture schools to offload the form-finding "creative
burden" to computers with the use of shape grammars. Though they're a driving force
in many departments, some will admit behind closed doors that they're also a bit of a red
herring, and that years in the spotlight have yet to bear fruit. My own observations are
that, rather than easing the burden, shape grammars have shifted the focus of labor:
students trade their Olfa knives for a keyboard and mouse, and spend hours debugging
Rhino scripts instead of erasing lines. Because most grammars are agnostic to physical
law, they also generate needlessly inefficient, material-laden architecture, which
rightfully sends the building scientists into the streets screaming blasphemy.
agreed.
although in gaming, the engineering aspect isn't so big of a deal, it is
difficult to get an automatically generated to be much beyond that of
"fairly lame".
it is a little easier with natural terrain, since terrain lends itself
to a number of strategies:
randomized mid-point subdivision;
other fractal-based strategies;
Perlin noise;
...
generating indoor spaces generally devolves to making a grid, and then
algorithmically placing walls/items/... this is common in many
"roguelike" games (Diablo / Diablo 2, Torchlight, ...).
my personal experience was that the results are "not particularly
interesting".
I suspect that the scene is much less interesting from a first-person
perspective than from an isometric one:
the layout itself is a major source of visual interest "as seen from above".
one does not see the layout first-person, only a bunch of similar
looking walls with a vaguely confusing/maze-like feel (it is infact
almost more visually interesting to see a pixelated rendition of what
the generator spit out than to wander around in the generated
environment itself).
generating "good" (and visually interesting) indoor spaces is seemingly
a harder problem than that of generating natural-seeming terrain.
a lot depends on the type of game theme though:
if it is some sort of LOTR style fantasy setting, one can probably get
by fairly well using primarily auto-generated terrain and the
occasional/simplistic building.
even for first-person, a Diablo/... style world generator would probably
still technically work, even despite the results being "not particularly
interesting".
sadly, different area-generation strategies don't necessarily combine well.
doing terrain-generation or doing like Diablo is not all that difficult,
but combining them is harder (except maybe if the terrain is also
tile-based, which is possible).
I've found that I'm most productive in creative endeavors when my goals are specific,
resources are constrained, tools are comprehensible and transparent, and my attention is
focused. I particularly love the sense of immersion that comes when sketching a scene,
writing an essay, repairing a small engine or designing a program (I think it's what
Csikszentmihalyi termed "flow"). I'd be lost if I had to design an entire
virtual world, as its far beyond the limits of my imagination, and dissatisfied if I
off-loaded the work to a machine, because I'd always know it to be a knock-off of the
real thing. Given a lifetime, I might be able to pull off a reasonable virtual vegetable
garden.
yeah.
my imagination is spread fairly thin here:
trying to deal with all of the technical issues, in addition to all of
the creative ones.
if one spreads their thinking over a large number of areas, one starts
drawing a lot of blanks "just what the hell am I going to do here?...",
whereas, if faced individually, such matters seem to be easier.
like, it is much easier to write ideas for the plot into text files when
not worrying about, say, how this will be expressed in-game, and easier
to think about making a particular piece of game-artwork when not
worrying about how it relates to anything else (plot or story, how it
will be used, ...).
so, "divide and conquer"...
then things relate in either synergy or disharmony, and one can decide
what to keep and what to discard as a more incremental process (even if,
granted, this is hardly a "beeline to completion" or "beeline to
success"...).
if/when things will be "complete", or even necessarily what form they
will take, is far from certain, but even as such "what direction things
are going" is usually fairly obvious.
I am not about to simply drop everything and start over with a new
concept, even if some people often suggest this... maybe because they
think an FPS game with a plot revolving about giant alien space squids
and bio-mecha and large-robot cyborgs engaging in ground-battles with
time-loops/... in the mix is stupid, but whatever...
it is often cheaper to continue in the same general direction and make
occasional "course corrections" than to try to change or abandon
everything because some "new concept of the day" demands it.
It's much more fun to go out into the real world, ask questions of it, and use
tools like pencils, paint, objects or mathematics to help find meaningful
answers. One example comes from learning to draw: I remember being fascinated
by the ideas behind perspective drawing, and was humbled that such simple
principles could have been hidden in plain sight for so long! After playing
around with vanishing points, it seemed that there must be some very
fundamental relationships between the points on the horizons and lines on the
page. This gave way to an exploration of projective geometry, which I was
fascinated to discover is an immensely powerful way of describing relationships
-- from mechanical linkages to structural loads and conic sections. From here
the lines on the page could be mapped to equations of lines, and from equations
of lines to linear algebra. Finding these relationships in ordinary things was
a great excitement, and though I've never used the knowledge to build
a
ny large CAD tool, my small experiments on paper and in silico have given me
a new perspective that I'll happily hold for the rest of my life. To that end,
I'd never want a computer to create a new world to live in, but instead be an
aid to understanding the one right in front of me.
ok.
"creating a new world to live in" is sort of the goal of game creation
though, but it is difficult to pull off well.
Finally, a few books worth mentioning:
Cliff Reiters "Fractals, Visualization and J", which chronicles an exploration of many
neat ideas: from chaotic attractors, to celluar automata, fractal terrain generation and projective
transformations. It uses J as its teaching language, but the code reads like "executable
mathematics", and could be put into another form without too much hassle. Reasonably priced
print copies are hard to find, but Lulu.com sells the eBook for less than the price of some
sandwiches.
yes, ok.
And though I'm always skeptical of attempts to mathematize art and design,
three books worth mentioning are:
Point and Line to Plane : Kandinsky
dunno about the book contents, but the title seems to describe a very
simple concept (fairly central to many CSG operations and similar).
Notes on the Synthesis of Form : Christopher Alexander
On Growth and Form : Thompson
yep.
_______________________________________________
fonc mailing list
fonc@vpri.org
http://vpri.org/mailman/listinfo/fonc
_______________________________________________
fonc mailing list
fonc@vpri.org
http://vpri.org/mailman/listinfo/fonc
