Some criteria for VRML

Dave Raggett (
Mon, 13 Jun 94 12:44:27 BST

I am off today to INET'94 to give my talk on VRML, and thought it worthwhile
to write down a few comments on what I see as requirements that VRML should
meet, if it is to be widely useful:

To gain broad acceptance we will need freeware browsers for Unix, PC
and macs. I intend to develop a portable vrml browser for Unix, based
on the public domain PEX 3D graphics extensions in X11.

VRML describes scenes as compositions of objects. VR places a premium
on efficient rendering, and so I believe that objects will need to be
specified for rendering at different distance scales, rather than as a
single 3D geometric shape. The methods of interaction supported by each
object will vary according to how far away you are from it. Thus as you
walk up to a wall, the detail on the clock appears and you begin to hear
it ticking. Similarly, the keys on a computer keyboard become visible
and allow you to type on the virtual keyboard. Ditto for a virtual white

Objects can be described as a composition of sub-objects. This allows you
to reuse a given element multiple times. PEX allows you to do specify the
scene in this way and then simply change your viewing position as you
move through the scene. The PEX server then does all the work to render
the scene without further intervention by the application.

Another theme is to generate structures according to rules, e.g. a tree as a
set of branches with leaves, Generative models would give VRML great power,
e.g. to populate an office with cubicles, or to generate a forest of trees,
or a landscape. One way to think of this is as a genetic code together with a
means of introducing variations. Arbitrary complexity can be generated
*deterministically*, e.g. by using a random number generator seeded with a
predefined value.

Deterministic complexity will give us rich VR scenes that take very little
information to transmit, but appear the same to each user regardless of
which browser is used. This is important when we want to meet in shared
VR scenes (virtual presence).

I will try and put flesh on these ideas when I get back from Prague.
In the meantime, please look at my paper (via

Best wishes,

Dave Raggett

----------------------------------------------------------------------------- Hewlett Packard Laboratories email: Filton Road tel: +44 272 228046 Stoke Gifford fax: +44 272 228003 Bristol BS12 6QZ United Kingdom