Since I was session hopping, these notes become even more fragmentary
than the preceding ones. Speakers names are from the schedule; I may
have missed substitutions. By the way, I used the Prospero system to
browse & grab the archives the other night and was pleased by the
convenience (Prospero gives somewhat transparent access to associated
public ftp sites; can't find how I got it any more).

Nicole Stenger, Visiting Scholar at HITLab

Description of "Angels" virtual world for children. Forsees a need
for lots of people with paintbrushes to touch up virtual worlds :-)
Kids walking, running, jumping, "complete lack of discipline."
Small environment to limit polygons; use of symbols for evocation,
compaction. Starting scenario: 4 distinguishable hearts. Touch one
to get to paradise, with a corresponding angel guide. How to represent
evil. slide shows angels as downward-pointing daggerlike bodies.

Doug Crockford, LucasArts

.. Problems with VR: expense, hygiene ("putting rented bowling shoes
on your face .. evil hacker transmits computer bug to people"),
defenseless while in VR, visual resolution & tracking inadequate..
Enfeebling vs. empowering: a world where you have really terrible
eyesight. Emphasis should be on experience delivered, not technology:
clips of Terminator 2 & Backdraft effects shown. I gather LucasArts
is staying with what they already do & not getting into VR.

Tom Piantanida, SRI

Areas of possible improvements for SRI to work on: Position/orientation
sensing: magnetic is too slow; looking at phased array ultrasound. Head
mounted displays: legally blind now unless you have $50K for HMD alone.
Tactile: how fast/close should tactors be; force feedback - 7 horsepower
of impulse to let you slap a file cabinet.

William Bricken, HITLab

Purpose of HITL is technology transfer; generating prototypes to
stimulate industry. The denizens of the lab meet outsiders during the
day, work at night. Development there is all integration. Description
of VEOS public domain operating kernel: every entity in a VR has entire
capabilities of the system: chair can call up statistical analysis to
see how it has been used. C code for VEOS and voice recognition will
be distributed on an as-is snapshot basis. Use of "flashlight" instead
of glove - cheaper, more "correct." Why move things through space?
Make them disappear/reappear elsewhere.