IMAGINE a high-speed electronic network that links 365,000 small companies, which can use it to form instantaneous partnerships to seize quickly shifting market opportunities.
Imagine further that the network is developed by the Defense Advanced Research Projects Agency (DARPA), which has incubated some of the highest technology of recent decades. Only now it will just be ARPA, developing civilian as well as military technology.
This is one example Clinton aides offer of the more- activist role for the government that the president outlined this week for keeping American industry globally competitive.
As the century opened, the government put American farmers on the cutting edge of agricultural knowledge with land-grant colleges, experiment stations, and extension service agents in every county.
As it ends, the Clinton administration has a similarly activist vision - though on a smaller scale - for supporting American industry with high-technology research and electronic infrastructure.
President Clinton does not want the government to pick winners and losers among products and companies, says White House science adviser John Gibbons, who is leading the technology policy. Instead, Mr. Clinton wants government to choose basic industries with the potential to support high-wage jobs into the next century and support their technology at their pre- commercial stage.
He proposes doing it on the cheap. His technology policy would add only $17 billion across four years to the $290 billion the government would already be spending on research and development.
The major move Clinton proposes is a shift in federal agencies that now work on defense and energy to technology with commercial uses.
The country has over 700 national laboratories, for example, that Clinton foresees supporting industry. Their traditional mission is to serve government customers - mostly defense-related.
Most of the key questions about the Clinton version of industrial policy remain unanswered. The biggest concern of both critics and supporters of this kind of government activism is the intrusion of politics into decisions about spending money.
The administration can also succumb to the temptation to fund big, attention-getting projects for publicity, says Martin Neil Baily, an economist at the Brookings Institution.
"The logic of the technology" should determine which projects are funded, he says, not the congressional districts in which they are located.
The Clinton administration has not spelled out how it will choose which ventures to fund or whether such decisions will be insulated from special-interest and congressional pork-barrel politics. Yet the White House has already decided on some key technologies to pursue. It will continue to fund the space station and the superconducting supercollider.
Some other examples:
* The administration is keenly interested in developing an electronic infrastructure based on information super highways - wires that can carry enough data that a user could download the entire Encyclopedia Britannica in less than a second.
* It will help to develop an electronic car that takes the automobile industry beyond the internal combustion engine.
* It plans a heavy emphasis on biotechnology as a major long-term growth industry that the United States would like to lead.
These choices put the government well beyond backing generic research and into making judgments about technologies, says Claude Barfield, a technology policy specialist at the American Enterprise Institute. "Government is not well equipped" to make those choices, he says.
He prefers to see government encourage investment without bias, such as with research and experimentation tax credits, which are included in the Clinton plan.
Government has long had a central role in driving technical innovation in the US economy. The aircraft and computer industries were developed as defense projects. The first computers were built during World War II by military scientists. Even International Business Machines Corporation in the 1940s and 1950s sold largely to the government. The economic benefits were profound, but incidental.
Clinton announced his technology policy Feb. 22 in Mountain View, Calif., at Silicon Graphics Inc., a fast-growing Fortune 500 company that got started under a defense contract.
Clinton wants to make the government seeding of technological innovation more deliberate and economically efficient.
"These are not large measures," says Dr. Baily of the Clinton proposals.
"The technology may be about as much as he could do, given the budget deficit - although he could have chopped the supercollider and the space station," Dr. Baily says.