Subscribe for Free Now!
Practical Applications for VRML 2.0
by Jeff Sonstein
VRML Version 1.0 happened because a few people had a vision and because a
lot of people got involved in making that vision real, as real as things
get on the Net. The vision Mark Pesce, Tony Parisi, and Gavin Bell had
was of networked virtual reality distributed around the globe. Version 1 of VRML
demonstrated that this was a possible goal, using existing and ubiquitous hardware
and current software engineering techniques. Version 2 of VRML demonstrates a huge
potential for industrial, educational, and entertainment applications. Version 1
proved the concept, and Version 2 overcomes many limitations of this original effort.
Version 1 of VRML was static, while Version 2 is dynamic. With
Version 1 I could set up a scene, but once the scene was in place it remained
static; objects did not move, user intervention was required to fly a camera
through a scene, and clicking on something could only load up something else.
The vrmLab scene is written
in VRML Version 1 and opens with a particular camera location. The user must
explicitly tell their VRML browser to move through the scene to a new viewpoint.
With Version 2, I can rewrite the scene so that the camera viewpoint is
automatically animated through a series of positions in the scene like
the way a video or movie camera moves through a series of shots .
VRML Version 1 was purely visual, while Version 2 supports the use of
multimedia presentations. With Version 1 I could just show things to
my audience; a computer monitor could display a still but not a movie. I
could only trick some browsers into playing a sound file if the user actively
clicked on a particular object but could not make this happen automatically.
With Version 2, I can have the computer monitor play an MPEG file, and I can
make a sound file begin playing automatically when someone first enters a scene.
VRML Version 1 was still, while Version 2 supports programming activity
into objects. With Version 1 I could not think about making anything happen
to objects once they were displayed. With Version 2, objects can move around
within a scene, things can change both size and color and also texture-mapping not only while
the scene is running but also in response to user input. The static spider icons
at the Version 1 vrmLab just sit there, while in Version 2 they may be
made to walk around the scene, make gestures and say things in response to
the passage of time or user input.
VRML Version 1 followed basically a top-down approach to conceptualizing
3D. With Version 1 I had to pay a lot of attention to the order in which things
were placed in files to make sure that the effects of changes in a file would be
predictable. With Version 2, objects are more clearly encapsulated, can be
more easily reused without carrying along extra attributes I don't really want
a second copy to have, can be changed dynamically, and can be built to
figure out how to deal with varying geometry at run-time.
VRML Version 1 allowed objects to have basically one possible behavior:
reaction to being clicked on by the loading of another scene. With Version 1, all
I had to work with was the Anchor node. With Version 2, objects
may be constructed which are inside sensor nodes and react to user
input like mouse movement in a variety of ways. With Version 2, I can
even use an invisible TimeSensor node to cause events to happen in a
structured and predictable manner just due to the passage of clock
ticks; the user doesn't have to do anything for an event to happen.
Version 2 allows me to both structure scenes which users can dynamically
rearrange and change and also have a good deal of control over what sorts
of prompts happen how and under what circumstances.
VRML Version 1 grouping nodes contained fixed contents. With Version 1
whatever I put inside a node was what stayed there the entire time a user
visited the scene. With Version 2, the geometry which may be inside the
children field of a grouping node may be changed. Boxes may be removed
from a Transform node and spheres added at run-time, depending upon various
events including user input. Version 2 allows dynamic binding of things
into grouping nodes. Prototyping allows the definition of skeletons to be
filled with geometry and other objects when the user interacts with the scene.
The shifts in orientation and expansion in capabilities from VRML Version
1 to Version 2 are quite powerful, and open VRML up to provide support for
a wealth of new industrial, educational, and entertainment applications.
The application scenarios I talk about here are merely those most obvious
to me and by no means comprise an exhaustive list. I will examine
just four: what the government calls an Integrated Electronic Technical
Manual or IETM, what architects usually call a structured walk-through,
an interactive multimedia presentation, and some online curriculum
The folks at Paper Software put out the WebFX VRML browser and the
Live3D VRML plugin for Netscape Navigator. When they put out Live3D, they
realized that extending the Anchor node (which allows users toclick on something to load another scene ) would greatly increase the utility of VRML anchors.
Live3D extended VRML to allow an anchor to specify a particular HTML frame
in which something was to be loaded when the anchor was clicked. Suddenly,
the functionality and practical utility of VRML scenes shot up. One early
proof of concept experiment with this extension to VRML Version 1 was
the Catalog Demo
scene at the vrmLab. Paper Software's pioneering extension to VRML
functionality was powerful enough that it has been included in the Version
Visualize yourself working as a 'bot mechanic in Netland, 2010. A defective
search engine has been brought into the shop for repairs. To figure out how to
test and maintain the bot, you load up an online catalog into your Web browser
and select Spider, Mk 2 from a menu. What loads up on your screen is a
Web document with a couple of different frames, one to display a model of the
Mark 2 Spider and one containing text about the Mark 2. Point at the head and click,
and the text frame changes to give you information about the equipment in that
part of the Spider as well as testing and repair data.
Do you think that a fanciful scenario? Try imagining a catalog page listing
all of the cars made by a particular automotive company instead. If you click
on a particular car listed in the catalog:
- a 3D model of the car comes up in one frame, and you can sit behind the wheel
- text information about that car displays in another frame
- an announcer's voice begins talking about the car, and music begins playing
- the camera begins moving around and through the car while the announcer talks
Planet 9 Studios began an interesting experiment last year: displaying a
part of San Francisco known as the SOMA District in a VRML Version 1 model.
Given the limitations of Version 1, they did a spectacular job. VRML Version 2
provides mechanisms like the ElevationGrid node which greatly simplify the task of
getting a description of geographic data into a VRML file. Geodata can be made into
a reasonable sized VRML Version 2 file.
Picture yourself working for a local Chamber of Commerce. With the advent of VRML
Version 2, it is now possible for you to present a reasonably detailed online
model of your town and it's tourist attractions. If someone visiting your city
wants to find out how to get from their hotel to a restaurant with a view of the
Golden Gate Bridge, you can provide them with a scene and you can cause their
camera viewpoint to move automatically through the scene, from one point of
interest to another, with an announcer's accompanying voice-over playing all the
Alternatively, picture yourself as an architect with a proposal for a building
project which you wish to present to a geographically scattered audience. You
could model the building and the terrain context in which it is to be placed,
and you can then create a script for the camera to follow in moving the
viewer up to the front of the building and in the doors and up the stairs
to the Grand Ballroom.
Pioneer Joel and some others at Arizona State University worked on a VRML
Version 1 model of the SuperBowl Stadium last year. With Version 2, you can not
only model the stadium but can also offer interaction with a Web document. This
would allow you to set up an online box office, which could include the viewer
being able to actually move their viewpoint to a particular seat to see what
things will look like. You can offer people the ability to try out perspectives
and viewpoints, and then to interact with an HTML document to order tickets.
A very interesting story was presented in the newspapers when the Dayton Peace
Accord was being worked out. It would seem that the participants were stuck at
one point, haggling over how wide a particular corridor should be to allow
safe passage from one area of Bosnia-Herzegovina to another. To illustrate the
point which was being made about the need to widen that corridor, a model of
the terrain in that area was brought up on a screen and the viewpoint was moved
into the hills surrounding the corridor. This made the point needed, and the
argument was quickly settled.
Alternately, picture yourself as a Professor of Geo-Politics and Geography at
a university or as a conflict resolution expert for an international
agency. You could use VRML Version 2 to construct varying scenarios for the
people you are working with to enable them to visualize the geographical
forces which have driven so much of history.
With the advent of VRML Version 2, 3D on the Internet is now integrated with
Web browsers and dynamic scenes and objects are now possible. These twin changes
open up a wealth of potential uses of VRML for industrial, educational, and
entertainment applications. The great leap forward taken by VRML with this new
version should make it very attractive to organizations wishing to present a
compelling presence on the Internet.
The problems involved in commercial and educational uses of VRML are no longer
technical ones. The limits on authors in combining media and integrating
information are mostly removed, content can now be dynamic and interactive,
and consumer-grade machines may be used for viewing. The only limits on
what it is possible for you to do with VRML are the limits of imagination.
If you can conceptualize what you want to happen as a video production, you
can imagine some of what may happen in a VRML Version 2 scene. If you can
imagine your video becoming something interactive and something which changes
to dynamically match what the user does and wants, then you can imagine
compelling and evocative VRML Version 2 scenes.
Jeff Sonstein is the Networks Administrator for New College
of California, runs the consulting firm of Sonstein and Associates,
and started the vrmLab
in early 1995 as a virtual space for VRML experimentation.
Jeff holds an M.A. in Social-Clinical Psychology, has been
playing with computers since the early 1970s, and currently
lives with his family in San Francisco.
#VRML V2.0 utf8 |
© 1996 Aereal, Inc.