Entrepreneurs who build successful businesses from scratch have something in common with boxing champions, a colleague observes: They tend to stay in the ring too long.

And like aging boxers, entrepreneurs who hang on too long risk being remembered for their embarrassing losing bouts, not their days of glory.

That fate has befallen William C. Norris, one of the pioneers of the computer industry who founded Control Data Corp. 28 years ago and built it into a $5 billion high-technology and financial services company.

Norris retired as chairman of Control Data two weeks ago, leaving his company in the midst of its worst crisis ever. It is in default on more than $300 million in short-term loans and is trying to negotiate a new agreement with its bank lenders. Meanwhile, it cannot borrow from them. It is facing a loss of perhaps $400 million for 1985, according to investment analysts.

One of its key businesses -- the manufacture of large computer memory devices and other peripheral computer equipment -- is locked in a vicious war of attrition with Silicon Valley and Japanese competitors, with no profits for Control Data in sight this year. Nearly half of the 18,000 employes in the peripherals subsidiary have lost their jobs in the past year, a blow that Norris calls the most painful decision he ever had to make.

It is a bitter ending to the unique career of one of the most successful, inventive and iconoclastic business leaders in this industry or any other. Norris not only marched to a different drum, he created the company that built the drum -- and pounded out the beat for nearly three decades.

It isn't possible for outsiders to pin down how much of Control Data's problems are Norris' fault. He turned over operational control more than five years ago to top subordinates, headed by Robert Price, who succeeded Norris as chairman. Norris, however, retained a firm grip on the key strategic issues.

A large part of Control Data's plight is due to the slump that has battered the entire U.S. computer industry. But much of it was self-inflicted: The company slipped up on quality and delivery performance; it was late to recognize threats to its computer services business; and, at Norris' urging, Control Data spread its attention over so many diverse ventures that it could not concentrate on the most critical issues, securities analysts say.

The company's agenda ran from providing computerized job training to convicts to trying to build the fastest supercomputers in the world. "The thing that is true is that we have been trying to do too many things," Price said in an interview last year.

Control Data, as the saying goes, had so many irons in the fire it almost put the fire out.

Norris' critics blame this on his maverick determination to use the power of the computer to help solve "society's unmet needs." A few weeks before he announced his resignation, Norris talked with a visionary's enthusiasm about the computer's potential for improving education at all levels, a quest that has absorbed a huge amount of Control Data's cash, with scanty returns so far.

If critics are right, the range of Norris' interests was a key to Control Data's lack of corporate focus. James O'Toole, a University of Southern California management professor and author of the book "Vanguard Management," is one of those critics.

"If Bill Norris is interested in something, everybody at Control Data knows that's where the fun is, and the excitement and rewards . . . If you have one person who's so powerful and he has a limited span of attention, the corporation gets jerked from one thing to another," O'Toole said in an interview. Even though Norris had turned operational control over, he still set the tone and priorities for the company, and it suffered as a result, O'Toole says. "It hurts me greatly, because what Norris stands for is the way we should go . . . Their philosophy was so good," O'Toole says.

There is a common thread in the histories of Control Data under Norris, Polaroid under Edwin Land and Mars under Forrest Mars, O'Toole says: "A paradox of greatness is that charismatic leaders relieve their followers of the necessity of developing their own leadership skills."

The accuracy of O'Toole's criticism won't be clear for several years, when Price's performance as Norris' successor is recorded. But Norris is probably stuck with the label of a leader who outstayed his time, says William Hamilton, an analyst with the investment firm of Piper Jaffray & Hopwood.

To appreciate Norris, however, it is necessary to understand his unique insights into the global competition in high-technology markets.

A lot of the things that Control Data dabbled in through its many technology-sharing ventures with other companies were based on Norris' perspective about the nature of that high-technology competition with Japan and other rivals, Hamilton notes.

Norris recognized long before many of his peers how cutthroat that competition would be. He understood that, for a company of Control Data's size, the only way to fight effectively was to take a Japanese approach, teaming with other companies to pool technology that none could afford to develop singly. Norris will be faulted for waiting too long to center Control Data's efforts on the most important of these ventures.

But he also will be remembered as a pioneer of that cooperative approach to solving some of the most daunting problems facing scientists and society. Norris created Microelectronics and Computer Technology Corp. in Austin, Tex., a joint research venture by a score of leading U.S. high-tech companies, including Control Data, that was aimed at some of the key frontiers of computer technology.

Another offspring is the Midwest Technology Development Institute, a venture based in St. Paul, Minn., and backed by nine state governments. The institute is designed to pool research resources of universities and companies in the Midwest on projects of particular value to that region.

Phillips Bradford, director of advanced technology for the State of Kansas and an MTDI director, says the goals include making family farming more profitable by putting small-scale food-processing technology into farmers' hands. Another project involves research in three states on the development of new materials valuable to Kansas' small-aircraft industry.

These ventures are Norris at his best, and are an indelible part of his legacy as well.