...go to the options menu from inside the game - select 3D.
(OK - it uses red/cyan or green/magenta glasses...but it's a start!)
> The idea: We have all these people buying 3D monitors. In CSS there's
> z-index for stacking elements. What would it take to render that z-
> index in 3D on those screens?
>
The trouble is that z-index isn't intended for stacking in the 3rd
dimension, it's for "layering" - which might be being used for API-ish
reasons. Certainly if you displayed all websites using this for a
stereoscopic 3D effect, you'd get some extremely bizarre results!
That's not to say that you couldn't create another CSS attribute for
doing that though (z-coordinate or something).
> I did some Googleing on those Nvidia 3D vision kits that people are
> using for games. They use, apparently, quad-buffering in OpenGL to
> achieve the effect. It was suggested to me that maybe WebGL might be
> fruitful. But unfortunately WebGL is a subset of OpenGL that does not
> include support for quad buffering.
>
That would be a very easy thing to add for WebGL...but not so easy for
regular <canvas> and HTML where the 3rd dimension is not normally
considered.
> Has anyone else been looking into this? Or fluent in the
> considerations/technologies involved?
>
I've played a lot with stereoscopic 3D in the past (mostly with VR
headsets like this one:
http://www.link.com/img/avcatt_ah-64d_high_res_300dpi.jpg
http://www.link.com/img/ahmd_high_res.jpg
...which use two separate displays and don't need quad-buffering - but
cost roughly as much as a pretty decent sports car!). The quad-buffer
stuff is simple enough to use in practice.
The architecture of the browser is such that you'd have to render the
two views into separate <canvas> contexts and have the compositor render
everything into the two sides of the quad-buffer.
Personally, I find stereoscopic 3D without either (a) a really high
quality dual-display helmet (of the
"costs-more-than-a-decent-sports-car" variety) or (b) high-precision
head-tracking...rather nasty. It gives me (and, according to some
studies: about 40% of all users) a blinding headache and occasional
nausea. It works much better in movie theatres with CGI movies because
the screen is so gigantic so that (as a percentage of screen size), your
head motion is negligable - and it's far enough from your eyes that
depth-of-focus issues are unimportant. But on even a large-screen TV
10' from my eyes, the distortion is painful and the 3D effect spoiled
because my eyes have to focus at a distance that's substantially
different from the "virtual-distance" to the images I'm looking at.
Another issue for the web in general (although a non-problem for WebGL)
is that 3D photography is extremely difficult to do well because of
depth-of-field issues. That's why live-action 3D movies are problematic
- but CGI 3D movies do so well. When the computer is generating the
images, you can perfectly control depth-of-field - but when you use a
real camera, you can't.
TV sets that require no 3D glasses are even worse to deal with.
These are all rather fundamental problems that cannot easily be
engineered away. For that reason, I doubt that stereoscopic 3D is more
than a passing fad for use in the home and office environment.
There is one use for the technology that goes into 3D games that I do
find interesting. That is the idea of having two player games that play
on a single TV screen without "split-screen" (which has the disasterous
consequence of letting you see what your opponent is doing!). The idea
is that you give one player a set of those LCD shutter glasses with two
"left eye" lenses and the other a set with two "right eye" lenses and
you can render two completely different 2D pictures - one for each
player. That's a relatively simple idea that has a lot of mileage for
game consoles - but sadly, someone already patented it. It would also
allow my wife to watch "chick flicks" while I watch something else - on
the same big-screen TV.
-- Steve