"And when I found my devil, I found him serious, thorough, profound and solemn. It was the spirit of gravity. Through him all things fall."
Some of the most engaging viewpoints in computer graphics are embedded in "top-10" lists of unsolved challenges. Creating these is a popular pastime started by Ivan Sutherland in 1966 and continued most recently by Jim Blinn in his 1998 SIGGRAPH keynote address in. Blinn also references Paul Heckbert's list of the top 10 unsolved problems of 1987. These are fascinating lists, which serve both as milestones of our industry and archives of our changing focus. I hope that an appropriate visionary continues this tradition sometime in the current decade.
It won't be me. Frankly, I think it's too early for another top-10 list. In fact, at this point, the list of challenges I have contains only two unpretentious items. Both were motivated by a comment I first heard as the SIGGRAPH '99 papers chair, which has been fermenting ever since. The gist of the comment, which I have heard from others during the past few years, can be summed up in the following statement: "All the major problems in computer graphics have been solved, and it is now only a case of filling in the details."
Shocked and dismayed when I first heard this, I couldn't help drawing comparisons with the infamous declaration by Charles Duell, commissioner of the US Office of Patent, who said in 1899, "Everything that can be invented has been invented." It also reminded me of the pronouncement by Ken Olsen, founder of DEC, who said in 1977, "There is no reason that anyone would want a computer in their home."
It is great sport to poke fun at such statements with the wisdom of hindsight, so I will add one more quote that hits closer to home. It came from the chief design engineer of an erstwhile leading computer graphics company. In 1986 he asked me in complete soberness, "You don't really believe that a polygon rendering system could ever replace a good line drawing system, do you?"
The dilemma for me is that some of the people making statements about CG being passé are thoughtful, well-respected researchers for whom I have a high regard; otherwise I could dismiss them as among the doomsayers for whom predicting the end of the CG evolution has become a mantra that sounds cleverly shocking.
At the opposite end of the spectrum, we find researchers placing great store in what appear to be minor advances in CG. And of course, it is natural for them to concentrate their energies with great seriousness on the problem at hand. They become bound to the mountain they happen to be climbing, and their own peak appears as the highest and singularly most important one. With that in mind, here is my list of two challenges to the CG community.
It was at a "Campfire" in Snowbird, Utah, on computer graphics and archeology that I decided in my next life, or perhaps my third life, I would come back as an archeologist. The applications of CG to archeology are limitless and exquisite, such as animating cave paintings by flickering firelight, reconstructing the Temple of Karnak, displaying the Parthenon in its original vibrant colors, or archiving Native American artifacts that are about to be returned to their original and rightful burial grounds.
During the last presentation on archival issues, the archeologist made a somber point: CG tools were totally inadequate for the job! He outlined his desires for a model that would incorporate elasticity, microscopic surface texture, porosity, and the like, and one that could be displayed accurately with a large part of the electromagnetic spectrum, for example, infrared, UV, and X-ray.
I had a sudden glimpse of a task more daunting, but also more beneficial, than any we have seen in the last three decades, during which our focus has been largely on achieving photorealism. Consider, for example, how difficult, but rewarding, it would be to build a graphics model that captures the effects of erosion on the blade of a flint so well that it could be used to estimate its age, and that would show how the flint would accurately appear under UV lighting, offering more clues to its history.
|Computer graphics could be applied to the field in many ways, such as studying how prehistoric people may have created cave art that appears stationary in artificial light (top) but animated by flickering light (below).
Some may contend that archeology is a pretty narrow application, but it isn't to archeologists. Without minimizing the spectacular successes of the animation industry, I find a certain parochialism in the idea that success in CG is measured by how well we do at the movies. It is good to keep in mind that there are applications from archeology to zoology that all have deep and intricate needs.
I once had a prominent psychiatrist and director of a major brain-imaging center tell me that the next greatest breakthroughs in all of medicine will come as a result of melding imaging techniques like CT, MRI and PET with CG. We haven't even started to embrace the various CG applications in our imaginations. We need to take time to look around at the other mountain peaks; the vista is breathtaking.
Whatever numbers one chooses, the ratio is meant to convey the idea that when trying to maintain a continuous rate of technological progress, researchers face a geometrically rising level of difficulty. Thus 85 percent of the development takes 15 percent of the work, and the remaining 15 percent of development takes 85 percent of the work. It is the quantified version of the low-hanging fruit analogy. Those researchers who made such large strides in the past are beginning to feel the difficulty in making the same dramatic steps, at least in the few well-studied areas.
This does not mean that we are getting less for more, however. Keep in mind that distinction comes in the last 15 percent. The sprinter who is 0.1 seconds faster, or the pole-vaulter who goes an inch higher, garners the fame, and all others fade into obscurity. Think of how quickly we lose patience with a technique that is only a few years old. If the avatar runs flat-footed, it is no longer acceptable, even though it amazed the Academy only a short time ago. Excellence is found in those last percentage points of effort—right, coach? We have not surmounted the summit yet. (There is probably another clever ditty in this like 85/15/85... that is, 85 percent of success is the 15 percent where 85 percent of the work is found. And driving that further, it is probably only 15 percent of the people who go there. And so forth.)
In CG, the technique that is a bit more realistic, somewhat more striking, or a little deeper in substance is recognized. If you review enough research papers you see this trend. The authors who consistently publish at SIGGRAPH go beyond the clever idea and have done diligence in all aspects of exposition, analysis, and engaging artifacts. I believe that as a community of researchers in computer graphics we are crossing the 85/15 bar, but only in those areas where we have concentrated our efforts, like animation or CAD/CAM. For me it is a point of maturity, an exciting time of brilliance and merit.
I feel pain for any who may have lost that sense of wonderment and magic in computer graphics. I feel joy for those who discover, or re-discover it. It is unimaginable to me that we are exhausting the new ideas in CG, or that we are even close to finishing the job.
"Up the airy mountain, Down the rushing Glen. We daren't go a-hunting For Fear of wise old men."
,a professor of mathematics and computer science at the Colorado School of Mines, is the SIGGRAPH 2003 Conference chair.
"It is unimaginable that we are exhausting the new ideas in CG."