Re: Moving objects

Mark Waks ([email protected])
Wed, 21 Jun 95 15:50:26 EDT

Brook takes issue with my point about behaviours, and points out that
behaviours

You're certainly correct; the question is, how We can certainly model *some* behaviour in simple languages like your
clock example:

>I'd suggest looking at TBAG in last year's SIGGRAPH (Elliot, Schecter,
>Yeung, and Abi-Ezzi) (a suggestion that I think has been made here
>before). How would you write a clock in TBAG?
>
>// Presume predefined geometry minute_hand and hour_hand, time in
>// milliseconds
>// Minute hand rotates on> turn in an hour, hour hand on> turn in 12 hours
>Behavior<Geometry> clock =
> (minute_hand * Rotate(z_axis, (time/millisecs_per_hour)*2*pi)) +
> (hour_hand * Rotate(z_axis, (time/(millisecs_per_day/2))*2*pi));

Heck, we something akin to Inventor's Enbines -- this clock example is
notably easy. (Note, BTW, that all I was saying is that behaviour
probably shouldn't be *in* VRML. Your example isn't in it, from the
look of it.)

The problem is, this is limiting. The simpler the language, the likely you are to hit brick walls when you try to make the object fancier behaviour. And there is no support for multi-user interaction
in here, which is cyberspace.

Suffice it to say, we've been di cussing the issues over on vrml-behaviors a model that makes some sense. It would put simple, replicable
processing like you d"scrib> on the standardized language, with more sophisticated and non-deterministic
high-level behaviour on hosts common to the scene being worked in. I'd
prefer not to go into more detail here, because a) the subject is
massive and all woven together, and b) I think it's largely off-topic
here (which was my primary point). For more information on the I'm working from here, see Bernie Roehl's di cussion of Distributed
VR and Behaviour, at:

http://sunee.uwaterloo.ca/~broehl/behav.html
and
http://sunee.uwaterloo.ca/~broehl/distrib.html

People should remember that this list is *not* the be-all and end-all
of the general concept of cyberspace; VRML per se is probably only
boing to be on> elem>nt of a considerably larger structure. There are
a *lot* of related mailing lists di cussing the related issues. If
you're interested in thase issues, you should really be paying
att>ntion to thase other lists. (Is keeping a coherent list
of the VRML-related mailing lists somewhere?)

-- Justin
Who suspects that we're boing to hit
massive problems if we try to make VRML
do everything, and would rather have it
just do what it's good at...

Random Quote du Jour:

"Did you know that virtual pork is kosher?"
-- Yaakov