Computing at the Speed of Light-2

Optical computing – processing information using light instead of the electrons used in ordinary electronic computers – may soon be a reality. In 1990 Alan Huang unveiled the world’s first digital optical processor, the device that would form the guts of any optical computer.

William Ninke, Mr. Huang’s boss at Bell Labs, calls the achievement a milestone. “The Wright brothers only flew so far and they only flew so fast. No one expected them to build the Concorde their first time around. But because they had built on a solid engineering and technological base, as we have, the opportunities for doing more were immense, ” he says.

Though many now are willing to concede that an optical computer is possible – a concession Mr. Huang sees a triumph in itself – only a few believe it can ever be a practical alternative to today’s electronic technology.

“It’s easy to get seduced by the idea of computing with light, ” says Jim McGroddy, IBM’s director of research. “The work is interesting, the physics is interesting, and their devices may be useful for certain purposes, but one shouldn’t kid oneself about this area’s applicability or practicality.”

IBM did extensive research on photonics, as the field is sometimes called, during the 1960s. Its scientists concluded that computing with light uses more energy than computing with electricity, and therefore that photonics would never compete with electronics. Light is good for transmission and connections, Mr. McGroddy says, but not much else.

Even Joseph Goodman, Mr. Huang’s former teacher and a Stanford electrical-engineering professor who is considered the guru of optical computing, says that “the first commercial general-purpose digital optical computer will appear sometime between the year 2000 and infinity.”

Like some others, Professor Goodman suspects AT& T is less interested in general-purpose optical computers than in highly specialized ones that would switch telephone calls and other information through its burgeoning optic-fiber network. Huang and Ninke concede such a use of their photonic technology is likely, but Mr. Huang says his main goal, endorsed repeatedly by superiors, remains development of a general-purpose machine.

Despite the detractors, this approach has many well-regarded believers. “The field has changed, and half of my fellow researchers don’t even know it, ” says John Caulfield, director of the Center for Applied Optics at the University of Alabama in Huntsville. He calls Bell Labs the leader, and its SEEDs ” unquestionably the best optical devices in the world.” Though Professor Caulfield won’t endorse Mr. Huang’s idea of an optical computer, he says the opportunities for parallel processing with light can’t be ignored.

Photonics research is burgeoning. Spending in the US has been estimated at about $100 million annually. Over 50 institutions working in the field are funded by the US military alone. The Pentagon is interested because optical computers would be harder to break into and wouldn’t be disabled by the electromagnetic burst of a nuclear blast. A market-research firm, Frost & Sullivan Inc., has gathered data suggesting the market for all types of optical computing devices reach $1 billion.

OptiComp Corporation, a small firm in Zephyr Cove, Nevada, aims, like Mr. Huang, at a general-purpose optical computer. Its founder, Peter Guilfoyle, demonstrated a special-purpose prototype in 1984. But many researchers say they are leery of Mr. Guilfoyle’s work, which uses a different principle to make light beams switch each other.

Most research has been focused so far on individual elements that would go into a computer, especially those that could be used in tandem with electronics.

Hewlett-Packard Corporation has about 50 photonics researchers mostly looking into interconnection rather than optical computing. “Our work is aimed at projects that can be of practical importance in a reasonable time frame, say three to five years, ” says an official. Ted Laliotis.

Aaron Falk, lead scientist at Boeing Aerospace & Electronics, says he will believe in optical computers only when he sees optical “transistors” that can be made of something as cheap and versatile as silicon. Bell’s SEEDs are made of the much-more-costly gallium arsenide and aluminum gallium arsenide. “There’s no reason to believe that won’t happen sometime soon, ” Mr. Falk adds, “though I sometimes joke that we’ve spent so much time looking for this dream material that it should be called “unobtainium”.

Mr. Huang hates such pronouncements. “The Japanese don’t say such things, ” he says. “The Japanese don’t ask, “Will it be viable in five years? ” They know it’ll be viable sometime, so they go ahead. Doesn’t anyone in this country realize that ten years isn’t a long time? ”

Mitsuhito Sakaguchi, head of NEC’s opto-electronics research laboratories, predicts that optical computing will develop, with a 20-year lag, the same as electronic computing did. Taizo Nishikawa, deputy director of MITI’s Industrial Electronics Division, says, “Optical computing is one of the most important technologies of the near future.”

Japanese electronics conglomerates have an edge because optics involves a merging of previously unrelated technologies. The same Japanese companies make both semiconductors and computers, and the biggest computer companies are also the biggest makers of telecommunications equipment – a field already revolutionized by optics.

Nevertheless, most Japanese experts say the US is still ahead, citing the pattern wherein they do better perfecting a technology rather than inventing it. Japanese companies develop many systems, but the fundamental motive is commercial. In the US the important thing is the frontier spirit.

Ответьте на вопросы к тексту:

1. What is AT& T?

2. Who is Joseph Goodman?

3. What are Mr. Goodman’s expectations?

4. What is light good for, according to Mr. McGroddy?

5. Why is the Pentagon interested in optical computers?

6. What is Mr. Falk’s joke about?

7. Why, according to many researchers, cannot photonics compete with electronics?