Re: Mitsubishi Demo in IEEE Computer

Mike Roberts (miker@nashua.progress.COM)
Fri, 23 Dec 94 12:23:51 EST


On Dec 23, 8:24am, John W. Barrus wrote:

> By the way, remember that there are two aspects to distribution of
> rendering info. One aspect is transferring the model geometry and colors
> (materials) to each machine with a window into the environment. The second
> part is transmitting updated information and changes in location and/or
> geometry.

Wups; I guess I should have made it clear that I was thinking about what you
term the second phase. I think we both agree that the first phase (VRML-like
distribution of a world db) is relativly easy in comparison to the second phase.

> The graphical database must be local or it couldn't be drawn.

I am not so sure that this necessarily has to be the case; there exist
multiprocessor systems with low enough latency and enought bandwidth to have the
rendering db distributed amongst processors; when a processor needs data "owned"
by its neighbour, it would ask for it, it would then be transfered over,
remainin while needed, then be disposed of later on. This kind of approach is
common in visualisation/simulation stuff running on multi-processor arrays where
the model is too big to be replicated on each processor; I can imagine it
working for rendering multiple views into the same world on a network with lots
of bandwidth and little latency. Obviously even concidering such an approach for
net.vr would be very silly (and it may have no validity for real rendering in
any case and just be a rather silly idea from someone who has never written a
real renderer) !

> We never planned on scaling that particular system and I don't believe that
> it would scale. In fact, to have scalability, you have to give up some
> consistency somewhere, and Anim maintains consistency between the worlds.

> Do you always see the same thing as the person sitting next to you? How
> about the person 3 offices away? We discussed this a lot when designing
> the new system.

No; I don't think you necessarily do (have to see the same thing). I completely
agree that you have to give up (consistency) in order to receive (scalability).
I think the only way to find out where the acceptable amount of inconsistency
lies is to write the application(s) and see. Presumably the level of acceptable
disparity will differ for differ for different applications.

> More details to come (please let us make sure it works well first!)

Can't wait. Thanks ! I will refrain from asking any more questions about your
new system until it is working ...

-- Mike