John Vacca
A platform-independent standard for VR (virtual reality) called VRML (Virtual Reality Markup Language) could make navigating through on-line museums, libraries, and marketplaces on the Internet as common as interacting with textual information is on the WWW (World Wide Web) today. By the time you read this, VRML pioneer Mark Pesce expects to have the first freeware VRML authoring tools and browsers for Windows, Mac, and Motif platforms available on the Internet (http://www.eit.com/vrml/).
Most of the current efforts--including VRML--revolve around nonimmersive (i.e., special headsets or helmets are not supported) VR.
VRML 1.0 is based on a subset of Silicon Graphics' (Mountain View, CA) object-oriented OpenInventor 2.0 3-D scene and graphics d
escription file format. Developers use VRML 1.0 to create objects that are rendered as an Internet user "walks" through a virtual room. Developers can attach hot links to HTML (Hypertext Markup Language) so users can click on an object and view additional text or images.
VR on the Internet does not require VRML. Using in-line graphics--good old-fashioned Mosaic with pictures in it--you can view graphics in the middle of a Mosaic page. By pushing a button that moves the point of view forward, backward, or sideways, the graphics are updated as you "move" through space. Each selection causes a network request to the Mosaic server that renders a new scene and/or loads a new HTML page.
However, Sandy Ressler, who is project leader of the Open Virtual Reality Testbed at the National Institute of Standards and Technology (Gaithersburg, MD), says the in-line method's drawback is performance. "You don't get the real-time response of VRML," he says. VRML is file-based, and once a 3-D scene is transferred
to your machine, performance is as fast as your local PC can deliver. Also, because VRML is based on OpenInvention, which sits above Silicon Graphics' OpenGL graphics library, OpenGL accelerator cards will accelerate VRML.
The prospect of widely available VRML browsers received a big boost when Silicon Graphics published the source code for parsers of VRML so developers can write VRML viewers. Companies such as Netscape Communications (Mountain View, CA, (415) 528-2619) and
Template Graphics Software
(San Diego, CA, (619) 457-5359, http://www.sd.tgs.com/template) plan to integrate VRML into viewers.
Companies such as TGS are also developing VRML authoring packages. Rik Carey, manager of 3-D graphics software at Silicon Graphics, notes that writing tools that convert a file from a specific modeling application's file format to the VRML format should be painless. "If you can write an authoring tool, you can convert it to VRML," he says. Other toolkit developers are taking a w
ait-and-see attitude.
"My feeling is a lot of very good ideas have gone into VRML, and it may be a standard that follows HTML into the big league," says Kevin Yurica, product manager for Autodesk's (San Rafael, CA, (415) 332-2344) Cyberspace Developer Kit, which is a C++ class library for developing VR applications and simulations for Windows and Windows NT. "CDK is consistent with the direction that VRML is moving toward, but there are some issues that have to be addressed."
One of those issues is the lack of support in VRML 1.0 for assigning interactivity attributes to 3-D objects. For example, with VRML 1.0, you can walk around and look at a 3-D representation of three balls atop a table, but you can't currently push the table over and watch the balls crash into each other. Pesce and others say a future version of VRML will add that support, along with support for letting multiple people interact with each other in a virtual world.
screen_link (54 Kbytes)
VRML browsers from companies such as Template Graphics Software (San Diego, CA) will let Internet users interact with nonimmersive VR applications.
The Net's Next Big Thing: Virtual Reality
