Re: VRML usage

Michael Sweet ([email protected])
Wed, 19 Apr 1995 22:16:23 -0400

On Apr 19, 6:01pm, Laird A Popkin wrote:
> Subject: Re: VRML usage
> On Wed, 19 Apr 1995, Jeff Sonstein wrote:
>
> > >[snip ...]
> > >
> > >>Is VRML defined to be Client/Server, Peer-to-Peer, or Single System?
> > >
> > >IMHO I would say client / server ..
> >
> > I must admit, I cannot see any way to coord browser-browser interaction
> > (several browsers who are "in" the same space and at least minimally aware
> > of each other) w/o some form of server, and when I toss in *all* browsers
> > being able to interact with the common objects and aware of the
interactions
> > of *others* with those common objects... oh my.
>
> This problem has been solved, by the Defense Simulation Internet (DIS)
> for the past ten (+?) years. All clients are also servers for
> themselves. Thus, if I see you, my computer is a server to yours for my
> appearance, and vice versa. Establishing the connections between objects
> that see each other is tricky (DIS broadcasts, which doesn't scale beyond
> thousands of objects) but isn't unsolveable. That is, I know of several
> companies that claim to be able to handle it...
>
>-- End of excerpt from Laird A Popkin

DIS only provides *minimal* support for interactive simulation. Entities
(participants) have specific types that are essentially hardcoded. You can't
say "entity 5 is a bird", you have to say "entity 5 is type 534256". For DIS
to work, all parties must cooperate WRT exercise, entity, and location. There
is also no common scene database format defined for DIS, tho there are many
proposals being considered.

If VRML ever goes dynamically interactive, we're probably stuck with the
server (maybe a cluster if things get *really* complicated) feeding updates
to the VRML browser on a regular basis, and controlling what kinds of actions
can be taken.

There may be some Good Things we can inherit from DIS, such as the dead-
reckoning facilities, but the fundamental design of DIS is flawed in a
non-coordinated environment (which isn't what DIS was designed for).

-- 
______________________________________________________________________________


 Mike Sweet                  2D & 3D Software for      Easy Software Products
 (301) 994-0377             Silicon Graphics, Inc.      20778 Wolftrap Street
 [email protected]                 Workstations       Lexington Park, MD  20653
                             http://www.easysw.com
______________________________________________________________________________