There’s an office across from a yacht basin in Marina del Rey that would please a lot of efficiency experts. The place is super clean. No desk tops littered with yesterday’s shopping lists,

unreturned phone messages or snapshots of the family. Instead of the babble of voices, there’s just a gentle hum. It’s ideal--if you don’t mind the fact that there aren’t any people. It’s a

room full of computers, and except for a few managers in separate offices, they are the entire work force of an outfit called MOSIS. Those machines communicate with customers, accept their

orders, sort them into production batches and send them off to be produced. No sales and marketing people. Just machines. You get the feeling they were made to hum just to reassure mere

business under the wing of the USC Information Sciences Institute. It is the result of what visitors are assured was a rather simple invention, a means by which computers can be programmed

to run a manufacturing operation so that instead of mass producing an identical item it’s possible to turn out an infinite variety. But that’s getting a little ahead of the story. Integrated

circuits of various designs are the brains of computers and electronic equipment. Designing new circuitry is a key part of research work that leads to product innovation and improvement.

Scores of these designs are etched into a single silicon wafer about the size of a large chocolate chip cookie. (Originally they were made by a metal oxide process, explaining where MOSIS

got its name.) The wafers are expensive to produce, and the only hope for profit is to make them in large quantity. This is fine for some users, and major companies with lots of resources

can afford the cost. But suppose all you want to do is test some new circuit designs? Producing a test quantity can be outrageously expensive, and getting on somebody’s production schedule

can take many months. Along comes MOSIS, a computer system that takes individual requests for new circuit designs, puts them together with other requests and produces a tape that

manufacturers of those cookie-size wafers can use to run their computerized production operations. The manufacturers still have cost-efficient mass production, but the customers can order

just a small quantity of each individual test circuit. Suddenly what might have cost $40,000 or more can be done for $2,000 a test, and the order lead time is reduced to a manageable eight

weeks. Result: More testing and more potential innovation. MOSIS was launched in 1980 with funds from the Defense Advanced Research Projects Agency, not one of your better-known government

entities but one that fosters a lot of work at the frontiers of technology. In addition to military and aerospace applications, 60 major universities use MOSIS in engineering research work.

The plan now is to broaden the base of universities to 1,000, permitting undergraduate as well as graduate engineers the chance to play with new circuit designs. Expanded use of MOSIS by

private industry is expected to come next. All of this has implications for the nation’s economy. Many experts believe that to maintain its manufacturing might against low-cost, mass

producers in other parts of the world, the United States must step up the pace of innovation and develop inexpensive means to produce top-quality, custom-made products. With systems like

MOSIS, production lines could churn out masses of, say, cars but with infinite variations. What MOSIS also does is permit small firms, previously unable to compete with the giants and their

huge research and development budgets, to pursue R&D; at relatively low cost. Not bad for a room without people. It’s just an awful waste locating a bunch of machines across from a yacht

basin. MORE TO READ