Path: utzoo!attcan!uunet!ginosko!gem.mps.ohio-state.edu!csd4.csd.uwm.edu!mailrus!ames!uhccux!bmartin
From: bmartin@uhccux.uhcc.hawaii.edu (Brian Martin)
Newsgroups: comp.sys.mac
Subject: Re: My favorite thing at MacWorld: A 6-D Mouse
Summary: Actually about Richard Feldman at NIH
Keywords: 3d
Message-ID: <4614@uhccux.uhcc.hawaii.edu>
Date: 22 Aug 89 09:59:34 GMT
References: <15124@dartvax.Dartmouth.EDU> <3948@phri.UUCP>
Reply-To: bmartin@uhccux.UUCP (Brian Martin)
Organization: University of Hawaii
Lines: 49

In article <3948@phri.UUCP> roy@phri.UUCP (Roy Smith) writes:
>In <15124@dartvax.Dartmouth.EDU> thomas@eleazar.dartmouth.edu
>(Thomas Summerall) writes:
>> [HyperSpace] is a new product [...] like a 6 dimensional mouse used for
>> inputting data from actual 3d models. [...] is available for $7500.
>
>	This may be new as a commercially available product, but the idea
>certainly goes back many years.  I can remember having lunch with Richard
>Feldman (of the NIH computer modeling lab) when he described what he called
>a joystring.  It was a 6 degree of freedom gizmo which you could grab with
>your hands (he envisioned a pair of them being used to manipulate molecules
>for docking studies).
>
>	The handles that you held were supported by 6 strings.  As you
>moved the handles, the various strings were pulled and released.  By doing
>some snazzy calculations, you could translate the linear motion of the 6
>strings into the 6 degree of freedom motion of the handle.  But, the really
>neat part was that it was an output device as well.  Under program control,
>you could adjust the tension in the strings.  The idea was that as the
>potential energy increased as you pushed two molecules together, you would
>feel it get harder and harder to move them.  A proof-of-concept pair was
>built, I believe, but the cost was pretty high, like several $10k.  Not out
>of line with the $7.5k you mention (without programable feedback).

In summer of 1986 I was on a medical school rotation at the NIH and was
fortunate enough to see a demonstration of Richard Feldman's joystring. He
used it to manipulate three dimensional representations of membrane-bound
molecules in real time, to attempt to visualize potential ligand-receptor
binding interactions.

Depth perception was provided by piezo-electric goggles which alternately
darkened the visual field of each eye at something like 30Hz. Each eye
received an image of the molecular model generated from a slightly different
view angle, by synchronizing the display of the left/right eye's view angle
to the transmission on light through the lens covering the left/right eye.
With these goggles on, you felt as if you could actually reach out and
touch the molecular represenations.

The software ran on an Apollo workstation, with rendering calculations
offloaded to an FPS array processor. Although I don't remember the exact
figures, he was able to achieve MFLOPs performance equivalent to or possibly
better than a Cray, at a small fraction of the cost of a Cray.

-- Brian
====
    Brian K. Martin, M.D.
ARPA: uhccux!bmartin@nosc.MIL
UUCP: {uunet,dcdwest,ucbvax}!ucsd!nosc!uhccux!bmartin
INTERNET: bmartin@uhccux.uhcc.hawaii.edu