...being frivolous (less time for serious programming), then surely this will:

http://cs.nyu.edu/~jhan/ftirtouch/multitouchreel.mpg

Maybe, unlike my previous conviction, the video adapter is more than a
buffer and a digital to analog converter!  I retain my position that
today's popular fancy "pixel shader" hacks to simulate reality (as
conceived in some artist's head) case by case are nothing more than
toys, and that any serious attempt should be done by tracing light
rays (in software, or at least in hardware separate from the video
adapter).

But the prospect of real-time image processing of SYMBOLIC or
practical (i.e., photo album manipulation) could be effect a great
boon in rapid, interactive programming!

I envision a stripped down 'Linux' 'microkernel' (no HURD jokes,
please) where discrete, character-based input (i.e., "keyboard")
serves only to bootstrap a general, powerful, visual-interactive
interface, which is both a program interface and a programming
environment (a la Bourne shell, but visual), where prototype
(initially interpreted, in this environment,) programs
semi-automatically 'trickle down' to C or assembly microkernel
servers.

Nvidia's incomprehensible 'binary'-only drivers look pretty tempting
after these thoughts!  (Reality hits when we consider how inflexible
they are, though....)  OGD1 seems to be caught right between two
niches... too powerful for basic output covered by cheapo cards... and
perhaps too weak for 'fancy' effects for use in applications like
this?

I'd love to get a chance to build one of these touchscreens!  For
those interested, it works by emitting infrared light (via LEDs)
through an acrylic screen, in which the light is totally reflected (a
la fiber-optics) -- nothing gets out.  A rear projector creates an
image on the screen for the viewer.  When a finger touches the
surface, though, 'frustrating' the reflective properties of the
acrylic, infrared light is refracted out the back of the acrylic and
picked up by a 640x480 webcam, processed in real-time by a 2 GHz
Pentium 4!  This approach is innovative because it permits multiple
inputs, unlike most touchscreens.  A diagram:

http://cs.nyu.edu/~jhan/ftirsense/ftirschematic.gif

Note that this is not mine!  It is Jefferson Han's.

Perhaps an OGP board could power the interface?  (Not the infrared
signal processor.)

It is large, unfortunately, due to the rear projector, and thus would
need some clever engineering to become manageable and reach the
consumer market.  I think instead consumers will get wireless,
flexible, place-mat-like touchscreen terminals using some
semiconductor technology, but the current device looks great for
researchers and hobbyists!
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to