Re: LANG: Object Sharing

D. Owen Rowley (owen@autodesk.com)
Wed, 15 Jun 1994 10:38:14 -0700


> Total device independence is an impossiblity here.

total independence of anykind is immpossible in most any circumstance.

I think its important to understand that there is a desireable design goal
in the idea of device independence, and that solid engineering effort can
deliver the goods in a sufficiently acceptable manner.

> You cannot expect a
> person using a KSR33 to access a VRML system. Same goes for the glass
> teletype equivilent. If the person doesnt have a 3d graphic renderer, then
> they dont get the benefit of VRML. Oh ok, they can read the ascii version
> of the world description. Ho Hum.

> As for graphic systems, I have been using a Rend386 based renderer for a
> year or so. It works ok and Joe Gradecki, et al have built some decent
> modem based shared worlds on top of it.

In teh last year several software rendering solutions that run across the
entire field of processors have surfaced. There are several more cooking away
in labs as we speak.
40K textured polys/sec on a 486 platform is now common, we will be seeing more
application level software with these capabilitys over the next twelve
months.
It is not sufficient thinking to design to todays capabilitys, in an evolving
universe those who are standing still are really moving backwards.
One MUST cast their design goals into the future and make wise choices based
on the best information they have, or as is said in the marketplace.
*trust in Allah, but tie your camel tightly*

> The VRML/VRIF standard does not restrict the platform independence, much.
> It does expect that you have some capability for floating point and convex,
> planar polygon rendering. That much of an assumption you need.

I suspect that the TV's of 1996 will do that and more.

the inaptly named set-top boxes, are really multi-media terminals, the
dedicated rendering engines you seek are these boxes , I'm almost sure
of it.

LUX ./. owen