AUSTIN, TEX. -- Skeptics seemed to have plenty of reason to scoff in early 1982 when William Norris, then chairman of Control Data Corp., persuaded competing microelectronics firms to team up for cooperative research to help stave off the Japanese threat to U.S. technological supremacy.

Would the research consortium, Microelectronics and Computer Technology Corp. (MCC), violate antitrust laws? Would members squabble among themselves and spy on research programs to which they were not contributing? Could the group attract the best scientists and researchers? Would members wait for the fruits of long-term research or press for immediate results, leaving the most visionary breakthroughs to Japan?

Skepticism intensified when Bobby Ray Inman, the former deputy director of the CIA who was credited with getting the consortium off the ground, announced his resignation in September 1986. It continued into 1987 as three important corporate members decided to drop out and MCC looked for a successor to Inman.

But throughout the uncertainty, research was going strong at the consortium here, and most of the early questions had been answered in MCC's favor. Now, less than a year after the high-profile Inman stepped down, MCC is evolving from an intriguing but unproven idea into a first-class research institution -- and it is showing some surprising muscle within the microelectronics industry. It has done so through a series of significant but not spectacular research achievements.

"The industry needs no-nonsense, brass-tacks technologies and that is what MCC is releasing, good, solid stuff," said Sheridan Tatsuno, senior analyst for Dataquest, a market research firm based in California's Silicon Valley. "Don't underestimate MCC."

Executives of the member companies who pay for MCC (the consortium calls them shareholders) say that another reason for optimism is Inman's replacement, Grant Dove, a former senior executive for Texas Instruments. Dove enjoys a strong reputation in the industry for understanding how to adapt new technologies to commercial products -- so-called technology transfer. That is the area where MCC most needs help, shareholder executives and MCC officials agree.

"I'm pretty bullish now that we can make technology transfer work," Dove said in an interview, having just visited the chief executives and R&D managers of each of MCC's 20 shareholder companies. Dove, who became chairman and chief executive in July, said he received "a vote of confidence" from shareholder executives with whom he met.

His honeymoon period has been sweetened by several positive developments: NCR Corp. last June unveiled the first commercial product based on MCC technology: A software system called Design Advisor that is supposed to speed the design of certain kinds of custom computer chips.

At least a half dozen other companies are using MCC technology in internal processes; Boeing Electronics Co. last spring set up a separate research center with 220 employes to handle the transfer of MCC innovations to Boeing production.

The consortium is getting a quick jump on the superconductivity bandwagon -- the technology that promises revolutionary improvements in the uses of electrical power.

Those deliveries are part of what MCC officials call "intermediate deliverables" -- usable innovations that are small or medium-sized steps in large-scale projects. MCC's research is divided into four main areas that reflect its incremental focus: VLSI-CAD, chip packaging, software and advanced computer architecture.

Shareholders may participate in one or all of the programs and within each program, they choose which of several specific projects they want to support. They receive the technologies that come from the projects they support. Average annual expenditure per company is about $3 million, but varies widely.

The large majority of MCC's research is tied to scheduled deliveries to companies. "They're not funding us to get somebody thinking about something for eight years," remarked John Hanne, who heads the VLSI-CAD program.

MCC is not, however, neglecting more visionary research. A project to give computers encyclopedic knowledge and the "common sense" to use it is one prominent example of MCC's high-risk ventures. The project, called Cyc, is headed by Doug Lenat, a mathematician and computer scientist whose graduate research at Stanford in the mid-1970s revived dormant hopes for creating machines that can learn.

Although his Cyc project is potentially the stuff of Nobel prizes, Lenat characterized it as more of a necessary evil to enable him to continue research that he has put aside because of the lack of precisely the kind of computer capability he is building at MCC.

"If you're going to solve difficult problems you need to have and use lots of knowledge," he said, terming that ability "consensus reality." That consists, he said, of about two or three million pieces of factual information and an equal amount of human-like reasoning ability.

Lenat expects to produce the first usable results of his work in 1990, and to finish the project in 1994. He said that personal computers will probably have Cyc capability by the late 1990s.

Some scientists fear that work such as Lenat's is too much the exception at MCC. They argue that the bottom line is ultimately better served by basic, rather than applied research and warn that trying to combine the two, as MCC is doing, is tricky.

Michael Dertouzos, director of the MIT Lab for Computer Science, who otherwise praised MCC, cautioned, "If you're trying to do both {basic and applied}, you're deluding yourself." MCC executives say the mix works well, although they acknowledge that it was at times difficult to bring the two approaches together. Even critics give the consortium high marks for synergy, coming from having a highly diverse combination of researchers.

In fact, MCC's critics today are fewer and milder than in the past. There seems to be a consensus in the microelectronics field that the consortium has arrived. To the degree that questions persist, the three most important are: How can the process of technology transfer be improved? How will departures of shareholders affect the consortium? And how does the overall quality of the research staff compare with the very best institutions?

NCR was able to quickly turn MCC technology into a commercial product because one of its senior engineers shuttled between NCR and MCC, keeping track of the development of "Proteus" -- the facility that aids the creation of so-called expert systems, such as Design Advisor -- and championing its use back at NCR. MCC executives are encouraging other member companies to try that approach, so far without success.

Some companies work harder than others to adapt the developments coming from MCC. At Digital Equipment Corp., for every dollar the company spends funding MCC research, it spends between one and two additional dollars adapting the technology, according to Thomas Gannon, director of technical planning and development for DEC.

MCC executives acknowledged that neither they nor shareholder companies are happy with the current pace of technology transfer from MCC laboratories to commercial products. Dove suggests that an increase in the number of relatively short-term appointments to MCC of researchers from shareholder companies might improve communication. Each company also has a liaison to represent its interests at MCC, but "not all have been effective," Stotesbery said.

The issue of shareholder departures is worrisome because the consortium has lost six companies (with two replacements) in the last two years and might lose a seventh. Dove is trying to persuade General Electric to stay at MCC, at a reduced level of participation, and GE executives indicated he probably will succeed.

But Lockheed Missiles and Space Co. will pull out of the software program at the end of the year. It may stay peripherally involved as a participant in the small superconductivity project, according to Robert Wertheim, senior vice president for science and engineering at Lockheed.

Lockheed's departure is a particularly troubling one for MCC. Lockheed executives are the first to pull out strictly because they think they can get more research value spending the money internally. "The likelihood of the {software} program paying off versus its cost indicated we'd be better off removing ourselves," Wertheim commented. The other shareholder exits have been due largely to mergers, acquisitions and major shifts in product lines among member companies.

Whatever its reasons, the Unisys departure affects all four program areas. The 1988 budget of $74 million could be reduced if departing members are not soon replaced, MCC officials said, although the consortium has a $10 million cushion for 1988.

Dove is actively seeking new shareholders. He said he is having "detailed discussions" with Intel and Texas Instruments, among others. Stotesbery indicated that an announcement could come by the end of the year.