Alternate View Column AV-66
Keywords: VR virtulal reality hypertext hardware software computing simulation 3-D graphics
Published in the July-1994 issue of Analog Science Fiction & Fact Magazine;
This column was written and submitted 12/27/93 and is copyrighted ©1993 by John G. Cramer.
All rights reserved. No part may be reproduced in any form without
the explicit permission of the author.
This page now has an access count of:
I live in Seattle, the city which last Fall was host to two major international conferences of interest to science fiction readers: The Annual International IEEE Symposium on Virtual Reality (VRAIS-93) and The 5th ACM Conference on Hypertext (Hypertext-93). I was able to attend both conferences, and I'll use this column to provide an overview of what I learned there.
I. The IEEE Virtual Reality Annual International Symposium (VRAIS-93), Seattle Sheraton Hotel & Towers, September 18-22, 1993.
Let's start with a definition: Virtual Reality (VR) uses a computer interfaced to the eyes and ears (and perhaps touch and other senses) of the user to create a simulated environment that responds to the user's movements and body motions, creating a verisimilitude of reality. I first wrote an AV column on VR in 1990 ["A Visit to Virtual Seattle", Analog 11/90]. Since 1990 the field has advanced in technology and also in public visibility. VR has also become a standard staple of science fiction, from the holodeck of STTNG to the many recent VR stories and novels. Video game manufacturers have made tentative VR experiments with 3-D glasses and "power gloves", and VR games are now a staple of video parlors around the world. VR has also been regularly featured in major newspaper and magazine articles, and "new age" figures have adopted VR as a metaphor for their other-worldly musings and pronouncements.
In a sense, VR is a victim of its own hype, for it has come to be quite firmly fixed in public consciousness while still in its technological infancy. Expectations for VR are likely to greatly exceed the technological reality. Simulation of the complexities of the real world with a computer is not easy. It demands computing resources that are near the upper limits of currently available computing power. Today's VR is still limited, tending to be jerky and of limited resolution. It can also produce disorientation and "simulator sickness" nausea in the user. VR simulations, even when not completely successful, have taught important lessons about connceting computers to human perception.
The discussions and demonstrations at VRAIS-93 show that the field has reached a new maturity, an new professionalism. As one of the speakers noted, VRAIS-93 was the first conference on Virtual Reality at which Shirley McLaine would not have been an appropriate keynote speaker. I will select several aspects of VRAIS-93 for particular attention:
§ VR Hardware - Last year the VPL Corporation, manufacturer of the standard DataGloves and EyePhones of 1st generation VR, declared bankruptcy, leaving behind a large hole in the availability of human interface VR hardware. At VRAIS-93 there was evidence that several manufacturers are now provide alternatives. Several types of high resolution head-mounted displays and "boom-mounted" displays are now in common use. There are also some interesting alternatives to data gloves, including a "space stick" suspended on force fed-back wires, allowing the user to "feel" virtual objects.
From my point of view, the most interesting commercial VR system was Sun Microsystems' Virtual Holographic Workstation Prototype. Sun demonstrated a standard SPARCstation that had been modified by adding a dedicated graphics board designed to support the heavy computational load of VR. The workstation included StereoGraphics 3-D glasses with liquid crystal lenses to "blank" one of the user's eyes while the other eye viewed the computer display. The glasses used a Logitech ultrasonic positioning system to measure the user's head location and orientation and feed these data back to the workstation. In place of a data glove, the Sun VR workstation supplied a StereoGraphics 3-D Siz-Axis Mouse, a mouse with buttons and a pointer that could be used to manipulate objects in the 3 dimensional space viewed by the user. I used the system and was very impressed with its capabilities.
Perhaps the most interesting aspect of the Sun VR workstation was its price. In 1990 when I last wrote about VR, the "entry fee" for a quality VR setup with a full set of VPL human interface gear driven by a couple of Silicon Graphics computers (one for each eye) was about $150,000. The Sun VR workstation, which appears to have superior interface hardware and computing capacity, costs about $15,000. As the entry price comes down, the number of developers of VR will grow very rapidly. A naive extrapolation of the cost curve indicates that quality home-VR is almost upon us.
§ VR Software - I'll focus here on only one of the many VR software systems discussed at VRAIS-93. The MUSE (Multi-dimensional User-oriented Synthetic Environment) system was developed by Dr. Creve Maples and his group at Sandia National Laboratory. MUSE is an open real-time shell designed to provide a hardware-independent programming environment for the development of VR applications. The MUSE demonstrations I saw included a simulated airplane flight over a mountainous pseudo-landscape, a fly-through examination of a medical brain scan, a surrealistic virtual environment created by children in a middle-school class, and an amazing model of the solar system, with the orbits of planets and their moons re-mapped to logarithmic distance scales. It is clear that broad standards and high-level software tools are needed to provide a common basis for ongoing VR development. MUSE promises to offer these in a freely available form.
§ Other VR Themes - The presentations at VRAIS-93 spanned a broad
range of applications of virtual reality. Let me simply mention a few topics
Aviation - VR flight simulation; walk-through inspection of aircraft design.
Chemistry - Complex organic molecule "matching and fitting" with VR
Communications - business and scientific meetings and conferences in VR space
Entertainment - Theatrical layout and staging, experimental networked VR games
Medicine - the virtual cadaver; VR medical image evaluation, remote diagnosis
Physics - Use of Aristotelian vs. Newtonian mechanics in VR simulations
Psychology - Minimum information required for recognition of faces
Teleoperation - disabled access through VR, space station maintenance
That's a brief overview of two days of workshops and three days of presentations. I hope it at least communicates the flavor of VRAIS-93.
II. The 5th ACM Conference on Hypertext (Hypertext-93), Seattle Sheraton Hotel & Towers, November 14-18, 1993.
About a month after VRAIS-93 the same downtown Seattle hotel was the site of Hypertext-93. Briefly, hypertext is the computer-based elaboration of normal (linear) text which permits a network of text passages interconnected by "links". The reader can read the text linearly in the usual way or can choose to pursue references, footnotes, and comments through many levels along arbitrary pathways guided by curiosity or the need for more detailed information. Hypertext brings with it new and unfamiliar ways of dealing with information. Many of our ingrained attitudes toward books, authoring, publishing, and information distribution are going to need to be reconsidered in the context of this new information technology. The growing availability of computers with excellent graphical user interfaces and high-bandwidth network connections has transformed the concept of hypertext from a remote ideal to a revolution-in-progress.
There now exist a number of implementations of (or approximations to) hypertext, and these provided an important focus at the conference. These implementations include Apple's HyperCard, Eastgate Systems' StorySpace, CERN's WorldWideWeb, a new and powerful program called MacWeb written by a French group, and others. All of these systems were demonstrated and discussed at Hypertext-93, and each had its circle of enthusiasts.
Hypertext as a concept can be traced to the "memex" described in the 1945 writings of Vannevar Busch, but the term "hypertext" was originally defined and brought into sharp focus by Ted Nelson in 1974. In the mid-1980s Nelson and others began Project Xanadu, a visionary attempt backed by $5 million in venture capital to achieve pure Nelsonian hypertext on a commercial scale. Unfortunately the project was terminated by the new management of its sponsor, AutoDesk, last year.
At Hypertext-93 Ted Nelson gave the opening address. It was appropriately inspirational, emphasizing the differences between his vision of hypertext and its current implementations (see above). One important difference between current hypertext systems and Nelson's is that immediate and direct economic feedback is a key element of Nelson's hypertext model. The author of a hypertext work (article, review, story, etc.) receives a direct pro-rated royalty payment by electronic funds transfer for each access of every element of his work, perhaps on a byte-by-byte or word-by-word basis. If the author quotes or excerpts another work of hypertext, the royalty payments also chain back to that previous work. (As an author accustomed to months or years of delay between writing and payment, this would certainly be a welcome change.)
A second difference is an emphasis on transclusion and back-referencing. Nelson requires a system in which an author can snip out or "transclude" a fragment of another hypertext work (a figure, a table, a paragraph of text) and include it seamlessly in his own work. Such transclusions must be back and forward referenced, so that the reader can follow the reference "backward" to its source, or follow a particular hypertext work "forward" to all the places where it is referenced or transcluded.
Nelson observed that these features need to be included in a true hypertext system. He mentioned his new project, Xanadu Light, which is intended to be a less ambitious and more realizable but full-featured version of the Xanadu system.
The talk elicited questions indicating that Nelson's vision of hypertext may be at odds with that of some of its current practitioners, who questioned the importance and relevance of some of Nelson's features. Nelson said in response to a question that WorldWideWeb looked very promising as an up-and-coming hypertext system, but it lacked back-referencing capabilities (not completely true). On the other hand, he felt that CD ROMs were "pre-Columbian" in that when you tried to go too far with them, you fell off the edge.
One of Nelson's answers was perhaps telling. When asked what hypertext system he used for his own writing, he said that he preferred to use Microsoft Word in its outlining mode. There were murmurs of disbelief from the audience. Was the Guru of Hypertext recommending MS Word? Incredible! It was as if a leading gourmet chef had admitted a preference for Big Macs.
Hypertext-93 highlighted several communities of interest in the field, so let me list them:
§ Reference Books and Manuals - The immediate growth industry in the hypertext is the generation of linked reference books and manuals (encyclopedias, catalogs, reference manuals, etc.). Much discussion at Hypertext -93 was centered on how best to construct such interlinked documents, and how best to design tools for their generation.
§ Automatic Linking - A related topic concerned the conversion to linked hypertext of previously existing documents. There are now reasonably efficient automatic systems for producing appropriately linked and threaded hypertext from a conventional document or database, but a small percentage of the links will be wrong. Human intervention is still needed.
§ The Hypertext Fiction Movement - The HyperCard and StorySpace systems are being actively used by a few avant-guard authors to produce works of fiction. Such non-linear story-telling provides a unique and interesting set of problems and opportunities. Will you soon be reading this magazine from a computer screen? Not in the immediate future, I'd say.
§ Hypertext on Networks - The Mosaic/WorldWideWeb system [see my AV column in the 3/94 issue of Analog] was the "sleeper" at Hypertext-93. There was no mention of this network-based approach to hypertext in the printed proceedings. On the other hand, about 1/3 of the demonstrations showed WWW systems using Mosaic, and much of the discussion in question sessions and hallway discussions was centered on WWW. It is clear that next year's hypertext conference (to be held in Edinburgh, Scotland) will prominently feature WorldWideWeb.
The processes we perceive in these two conferences are the shaping of technology by visionary ideas and the modification of existing ideas by the constraints, limits, and opportunities of technology. This dance of emerging technology will continue, but some convergence is already apparent. One can already achieve a remarkable sense of reality in exploring virtual computer-generated environments. One can already "surf the net" with Mosaic/WorldWideWeb, touring the world on electronic butterfly wings while sipping the information nectar of hundreds of workstation "servers". The joining of VR and Hypertext into some Gibsonian cyberspace does not seem to be happening. But the convergence of technologies that is forming is very important and will change our lives in unimaginable ways.
This page was created by John G. Cramer on 7/12/96.