T_Vision:
"The Billion-Dollar Code"

From The Playful World: Interactive Toys and the Future of Imagination (Ballantine Books, 2000) p.199 - get it here.

By June of 1995, quite a few people were working in VRML. Rikk Carey headed a crew at Silicon Graphics developing the first VRML Web browser, and Tony Parisi had gone off to found Intervista, and began to work on VRML tools. I was nearly finished with a textbook on VRML, in the hope that I could teach people to create their own VRML worlds. Our efforts would be seen by the public for the first time when the Interactive Media Festival took over Los Angeles’ Variety Arts Center.

When I walked into the Variety Arts Center, I’d like to say that I felt déjà vu – I’d certainly seen these walls before, modeled in VRML. A four-month project to model the building and its exhibits had just been published on the Web – the first VRML world of any significance. As I toured the Festival, I realized that the Variety Arts Center looked quite different in reality than in simulation. Oh well. A lot of folks were already inside, chattering about this exhibit or that. I ran into a friend – a Hollywood movie director – who asked me, “Have you seen T_Vision?” Nope. What is it? “Come with me.” He led me up the stairs – and into another world.

On a five-foot television screen directly in front of me, I saw an image of the Earth in space, floating gently against a sea of black. Beside the screen, a large ball – looking like an overgrown beach ball – sat in a metal frame. Someone went up to the ball and spun it. The Earth moved. It was 1:1. As you spun the ball, the image moved on the screen.

Ok, that’s nice.

Someone else walked up to the ball and pressed some buttons mounted on its frame. Now the image of the Earth began to fill the screen. As it grew closer, and I began to see the individual features of the Earth – continents, then lakes, and then rivers. We got closer and closer and closer, and, the closer we got, the more I saw. Now I could make out cities and towns and industrial belts. We were zooming in on Germany, then on central Germany, then on Berlin, then I could begin to make out the streets, and, closer in, I could see the buildings, and finally, just one building, and then, with a screeching halt, I was poised outside the window of an office building, looking in.

This is very nice.

We rode the trip in reverse; from building to streets to city to landscape to continent and, just as suddenly, a serene Earth floated pleasantly on the screen. I’d ended my journey where I’d begun, in high Earth orbit.

My friend leaned in. “That’s not all. It’s live.”

“What?” I croaked. This was too much.

Someone must have heard us, because another button was pressed, and – voila! – the cloud patterns, live via satellite, appeared over the image of the Earth. I felt like I was in the Space Shuttle, looking down. No, it was more than that. My heart was full. I felt like I was seeing God. It was perfect. I had dreamed that such a thing might be possible, someday, but never in my wildest moments did I imagine that it would be here, now, right in front of me, live and in color. I had to know everything about it, so I walked up to a tall gentleman working that amazing device, and asked him to tell him the story of T_Vision.

In 1988, Joachim Sauter, a professor at Berlin’s exclusive Academy of Fine Arts, founded the Berlin design firm of ART+COM with almost no money – but he had ideas galore. Located in the heart of old West Berlin, ART+COM brought artists and engineers together around a series of projects intended to demonstrate the coming integration of computers, communication, and design. At the time, few imagined that the Internet would sweep the planet, and no one knew that the World Wide Web would soon provide universal access to the global information resources of humanity, so the ultra-staid German business establishment had practically no use for such a renegade group of dreamers. If businesses couldn’t see the value of the coming convergence between design and communications, Sauter decided he’d create a few pieces – artworks – which would at least shock and entertain the public.

One of Sauter’s first installations, titled Zerseher (a pun which roughly translates as “dis-ceiver”) portrayed a famous Renaissance painting, Boy with a child-drawing in his hand, by Francesco Carotto. It’s a very striking painting – some might say strange – as the child pictured wears an almost demonic grin. When you walked up to Sauter’s copy of the painting, that odd grin might cause you might stare at it for a few seconds. If you did, you’d see something very curious happening: wherever you looked at the painting, that part of the image would smear, then disappear. The more you looked at the painting, the more you’d destroy it. Eventually, the destroyed image would be replaced with a pristine copy and the process would begin again.

Sauter used some high-tech magic to accomplish this feat. Behind the projected image, which looked very much like an actual painting, a video camera tracked the faces of people who walked up to the pseudo-portrait. Using some very fancy software, written by an ART+COM colleague, Sauter examined the video to locate human faces. From that bit of information, he could then fix upon the eyes, and learn where people’s gazes were pointed. Then he’d “erase” the painting at that location, continuing the destruction as their gaze moved about the image. The longer you looked at the painting, the more it disappeared.

It was all very subversive, and, as Boy with a child-drawing in his hand is the first painting known to portray a child’s artwork, Sauter thought it a very clever commentary on the infantile state of computer graphics at the beginning of the 1990’s. As with most works of modern art, some people loved it – but many others hated it.

Zerseher got ART+COM some badly needed recognition, and some even-more-badly-needed contracts. Berlin, formerly two separate cities, wanted some help planning the proposed reconstruction of the eastern “downtown” – Potsdammerplatz, the heart of old Berlin. The entire area, mostly filled with ugly Communist-era offices, was to be leveled, and a brand-new city built in its place. But what would that look like? The city planners turned to ART+COM, and asked them to create a virtual tour of Berlin – yesterday, today, and tomorrow.

Using old photographs, movies and maps, the ART+COM crew, led by Sauter, meticulously reconstructed pre-War Berlin, in virtual reality. You didn’t need a head-mounted display to explore this vanished world; ART+COM projects shied away from the gear associated with virtual reality, favoring wall-sized displays. With the exception of the depth perception associated with head-mounted displays (the source of their dangerous effects on the nervous system) the effect is fairly similar. You could stand in front of a wall and stroll through a Berlin that had been bombed out of existence almost fifty years before.

Next, they recreated the present-day Potsdammerplatz, which would soon disappear forever. Then ART+COM worked with the city planners to place models of the proposed buildings into the simulation. Now the bureaucrats could actually see how Berlin would look in ten years’ time, when all the construction would be finished, and a new Berlin would have risen from the ashes of Communism. With the Berlin Wall torn down and removed, the city could adopt a stately flow that harkened back to its pre-War glories, but still pointed toward its future as the capital of a unified German nation. Using the simulation created by ART+COM, the city planners could envision the future, and work toward making a city as beautiful as it had once been.

All the while, the staff at ART+COM were learning how to use their computers for ever-more-impressive tricks. They had a good selection of high-end Silicon Graphics computers for their work – these being de rigueur for virtual reality – and were now learning how to create enormous virtual worlds, with thousands of buildings, covering hundreds of simulated square kilometers. It got them to thinking: if we can create a model of Berlin, can we follow it up with something bigger?

At just this moment, Deutsche Telecom came calling on ART+COM, asking them to develop a project for the ITU Conference in Kyoto, Japan, six months hence. Sauter made an offer they couldn’t refuse, an idea so over-the-top that, if it worked, Deutsche Telecom would have the most impressive demonstration ever given at the ITU. (The folks at ART+COM weren’t completely sure they could make it work, but what Deutsche Telecom didn’t know wouldn’t hurt them.) Those six months passed quickly, in a haze of design and programming, and when the fateful day arrived, nothing in Kyoto was going as planned.

Pavel Meyer, dark-haired, short and compact, was sweating, swearing, and tapping on a keyboard. Next to him, Axel Schmidt, a tall, blond Aryan who rarely ever breaks a sweat, looked worriedly into a computer monitor. Both men, far away from home, were working as if their lives depended upon it. A few feet away, Sauter paced back and forth, making occasional circles around Gerd Grüneis, who looked as though nothing could bother him. But he’d started chain-smoking, lighting a new cigarette as he tapped the previous one out. Two hours to go, and nothing was working.

Meyer and Schmidt were crouched in front of the twin monitors of a million-dollar Silicon Graphics Reality Engine, a top-of-the-line “supercomputer” designed to make short work of even the most complicated computer graphics. That is, if it’s programmed correctly. The bigger a computer is, the more likely it is to be finicky about its programming. Reality Engines were notoriously finicky. They could create breathtaking images, but they had to be coaxed – nursed – into doing what was asked of them. And this particular Reality Engine was being asked to do something only one other computer had ever been able to do.

Meyer was at a loss. The program had been working fine, finished just an hour before he’d boarded a plane to Kyoto, the beautiful and ancient capital of Japan. True, he’d cut it close to the wire (it had only just started working before he had to leave for the airport) but Meyer felt sure that everything would go smoothly when he installed the program upon his arrival.

All around them, the final preparations for the International Telecommunications Plenipotentiary Conference were almost complete. The ITU, a branch of the United Nations, holds a conference every four years, an “Olympic” event which gathers all of the biggest telecommunications companies together under one roof for a week of policy making, parties, and demonstrations of Things to Come. In the early 1990’s, most of the telecommunications companies were governmental agencies, and Deutsche Telecom was no exception. With the reunification of Germany, Deutsche Telecom had become the largest telecommunications company in Europe, and at the ITU conference they intended to flex their muscle with an eye-popping demonstration of the future.

With less than an hour to go, even Grüneis started to show a bit of nervousness. Something wasn’t right. It had been working – Meyer had left Berlin with a fully-functional version of their ITU Deutsche Telecom demonstration. But Meyer had been testing his software on a different machine, seven thousand miles away. Perhaps this Reality Engine had a subtle difference in its configuration, something that might take Meyer and Schmidt hours to find, if they could find it at all. Deutsche Telecom had set the machine up for ART+COM as directed, but sophisticated computers often take weeks to “tune”. Each one has its own personality, its own settings and peculiarities. That was likely the problem here. Or so Grüneis hoped.

Sauter took to tapping his fingers on the gigantic beach ball interface to T_Vision. Neither he nor Grüneis could do more than look on as the two programmers searched for a solution to their problem. In order to make this project work at all, Schmidt had to write graphics code that squeezed every ounce of performance from the Reality Engine, and Meyer had written an ultra-fast database which contained billions of characters of data – a very tall order for the early 1990’s. In every way, they’d pushed the machine to its limits.

Now Sauter had to contend with Deutsche Telecom executives, who had suddenly begun to look very nervous. The doors are opening in half an hour, one said. Sauter could do little more than shrug and apologize. We’re working on it. Just a little more time. The executive didn’t look any less worried after Sauter’s reassurances.

With less than ten minutes to go, Meyer found the configuration error, corrected it and – voila!T_Vision was up and running. Perfectly. For the next five days, conferees at the ITU Plenipotentiary Conference would crowd the Deutsche Telecom booth, marveling at this demonstration of German ingenuity. ART+COM had hit the ball out of the park, a gold metal performance at the Telecom Olympics, and had given birth to a technology which will form a critical part of the 21st century.

A month later, in Los Angeles, T_Vision made its American debut at the Interactive Media Festival. That’s where I met Joachim Sauter – modest, but glorying in the reception the work was receiving. “This is incredible.” I offered, “Just incredible!” He received my praises thankfully, and, that evening, I met him and Gerd Grüneis at the bar. We drank beer and talked about T_Vision. I was amazed by it. They told me that their model of the Earth was composed of around five billion characters of data, that it could display everything on Earth’s surface down to a few kilometers, and – in a few sections – down to street level. I learned that T_Vision was designed to be networked, that each T_Vision system (there were three in the world – Berlin, Kyoto, and Los Angeles) shares information with the other T_Vision systems, to create a more complete view of the whole from its parts.

That was an idea I could get behind. It sounded very much like my own work. It made me want to put a question to them: Have you heard of VRML? Yes, they’d seen it. Very nice. Do you think T_Vision could work in VRML? They laughed. We’re taking a Reality Engine to its limits with T_Vision. Unlikely that VRML could handle the job. Hmm. Well, maybe someday? Someday, sure.

That was enough for me.

In early 1996, I spent a week huddled over my computer. I unplugged the phone and didn’t even bother to check my ever-increasing flow of electronic mail. I searched across the Web for real-time images of the Earth from space, and I found them at a curious site run by John Walker, a software multimillionaire who had earned his fortune as co-founder of Autodesk, an early developer of Computer-Aided Design (CAD) software. Walker had recently retired to Switzerland at the ripe old age of 45, and, from his comfortable chateaux, created an extensive website to showcase his interests. Among these, he included both astronomy and Earth visualization.

Nestled deep within Walker’s site, I found that he took the photographs generated by weather satellites, prettied them up, and then provided them on his website. They were accurate to within an hour. I set to work, and, a few days later, I’d finished my own homage to T_Vision, which I named WebEarth.

WebEarth is a three-dimensional, live model of Earth, just like T_Vision. Unlike T_Vision, it doesn’t have the extensive detail that requires a high-performance computer. Anyone with a computer and a modem can access WebEarth from anywhere. I wanted it to become universally available, so I made sure that it wouldn’t overwhelm any computer which gazed upon it.

When I’d completed my work, clicked on a link in my Web browser, and actually saw the Earth – live – floating in the inky black of cyberspace, I broke down. (WebEarth has had this effect on several people I know. Something about the image is starling.) It was beautiful – perhaps not quite as beautiful as T_Vision, but beautiful enough. And anyone could get to it, anywhere. You can too: just go to http://hyperreal.org/~mpesce/we5/we.html and see for yourself. It’s the Earth, live and in color, and reason enough, I believe, to see the Web in 3D.

Back in Los Angeles, T_Vision garnered the $5000 judge’s prize at the Interactive Media Festival, the first of many awards it would win in later years, as computers would grow smaller, faster, and ever more capable with 3D graphics. A million-dollar Reality Engine of 1995 would shrink to become the $300 video game system of the year 2000, and once that happened, T_Vision could enter the home...