A large ship's steering wheel stands on display just outside the office of William C. Norris, founder and chairman of Control Data Corp. It is a trophy presented by his staff after Norris outdueled International Business Machines Corp. in a celebrated antitrust suit a decade ago.

It is also a fitting reminder of Norris' long turn at the helm of Control Data, which makes advanced computers and sells an encylopedic range of financial, educational and data-processing services.

Today, Norris' helmsmanship is under sharp attack as the company reels under the heaviest losses in its 28-year history.

To critics on Wall Street, Control Data's financial crisis is proof that the 74-year-old Norris and his top managers have stayed too long, have held control too tightly and have swung too far off course following Control Data's widely scattered business pursuits. For years, Norris' willingness to buck the prevailing currents in the industry, and his idealistic determination to use computer technology to address "unmet social needs" have bred skepticism and resentment in the investment world, and now he is paying for it.

A recent Fortune magazine survey of some 4,000 corporate executives, Wall Street analysts and outside directors placed Control Data among the five least-admired American companies ranking dead last in terms of long-term investment value.

In a recent interview in his office at Control Data's headquarters, with an icy panorama of Minneapolis countryside behind him, Norris took on his critics.

"I've heard those [criticisms] now for oh, so many years that it's hard to get my attention on it. It was said, 'Geez, the old bastard will never turn control over to somebody else, [or] turn the operations over to somebody else.' I've proven them wrong . . . .

"I've had very little to do with the [daily] operation of the company now for five years. I've focused on the long range. But that really went unnoticed. They say I've still overstayed my time," said Norris, a Nebraska farm boy turned electrical engineer who became one of the pioneers of the computer industry.

"I'll leave one of these days. I'd like to do some other things before I get too old and have to go into a wheelchair. [But] I don't like to turn my back on fights."

The next three rounds in that fight, early next year, are likely to be critical. Control Data is in default on $309 million in short-term bank debt and must work out a repayment plan before it can borrow again. It needs to raise as much as $100 million from the sale of its profitable Ticketron division. And, most important, it must pull together a battered core business -- the manufacture of peripheral computer equipment, particularly the high-performance disk drives that store and funnel information to and from computers.

The casualty reports from the peripheral business, which provided about 30 percent of Control Data's revenue a year ago and elsewhere in the company, have been grim.

When the U.S. computer market stalled out last spring and demand for Control Data's products plummeted, the company was forced to cut deeply into its payroll and close plants. Employment in the peripherals business has fallen from 17,500 worldwide a year ago to near 9,000 now.

In August, the company's credibility with investors was jolted when it had to lower sharply its revenue and profit reports for 1984 and the first half of 1985. The revisions were required by the Securities and Exchange Commission, which objected to some of its accounting methods.

A month later, a $300 million debt securities offering intended to repay the banks had to be canceled because of the worsening peripherals business.

In October came the announcement of a $255 million loss for the third quarter of 1985, a much larger blow than outsiders had expected.

The result is a severe cash squeeze -- "a vicious whirlpool," analyst Thomas J. Crotty of The Gartner Group, said two months ago.

"The company is in a race to cut costs before exhausting its working capital," Jonathan M. Fram of Paine Webber Inc. said last month.

"Fundamentally we have been since September at a standstill with the banks," Control Data President Robert M. Price said in an interview. "We neither borrow nor pay back. As a consequence, over that period of time . . . you have a whole sequence of defaults that are triggered by not paying things when they're due. . . . It's that short-term debt that is the subject of our immediate attention."

Price said the bank negotiations are going "okay," but are complicated by the large number of lenders and the uncertainty of the company's current financial condition. "What that does is cast a cloud over the company," he said. Removing that cloud becomes increasingly important. "The longer it goes on, the more unsettled people get," he said.

As the investment community awaits Control Data's fourth-quarter financial report, the sense of crisis has eased somewhat.

Paul Karras, an analyst with the Minneapolis investment firm Piper Jaffray & Hopwood said he now believes Control Data will be able to reach an agreement with its lenders on short-term debt.

"We view it likely that the negotiations will succeed," he said. But the optimism is qualified.

"Anything is possible -- 1986 is a very gray area" for Control Data, he said.

The cutback in payroll and operations -- equivalent to about $300 million in annual savings -- will add to cash flow next year. And that cash -- plus revenue from the sale of assets -- will have to be used first to pay off the banks, then to continue the investment in Control Data's remaining operations, Price said.

But a recovery in the peripherals remains the most important financial challenge, and there, the issue is still in doubt, says Lawrence Perlman, who took over the peripherals business this year.

Perlman warns that there will be "no winners" in the overheated competition next year in disk drives. The advanced drives Control Data makes -- each capable of storing the equivalent of 20,000 to 566,000 pages of typewritten copy -- are also made by an expanding number of rivals, American and foreign.

Even after this year's painful cutbacks, production costs must be reduced another 20 percent to 25 percent to return the peripherals business to profitability, Perlman said. It would have been tough enough if carried out over several years. Control Data has had to do it in an excruciating nine months.

Norris faults himself for not forcing Control Data to defend itself sooner against the inroads by U.S. and Japanese rivals in the peripherals business. Ironically, Norris had an early understanding of the threat of foreign high-tech imports, particularly from the Japanese, an alarm he began sounding three years ago. It prompted him to help organize MCC -- the Microelectronics and Computer Technology Corp. in 1982, a unique technology-sharing joint venture by Control Data and a score other U.S. computer firms.

Control Data, like many others in the computer industry, was trapped by its own success. It continued to churn out disk drives and other products as the industry's sales soared in 1983 and 1984. But it was losing ground to competitors in some crucial areas: cost control, product quality and the ability to deliver new products on schedule. Caught up with meeting current delivery schedules, it didn't spot the erosion in these areas until too late, Norris acknowledged.

"We saw that. I talked to our people about that. And the average response was, 'Well, damn it, isn't the old man ever happy?' And the answer is 'no,' but that doesn't really make any difference. . . . Just because I say so doesn't make it so down there. . . .

"And the fact is that peripherals had been making good money. The people then running it just couldn't perceive what happened and simply didn't move fast enough. And I'm chief executive officer and I'll take responsibility for that. However, I'll go on to say that even if we had moved as rapidly as we could have moved, we would not have avoided a very substantial part of the losses we incurred."

"We opened the door to competition and it walked in -- in many cases with superior quality," added Price.

Critics say that failure was caused by Control Data's insufficient concern for short-term financial results, which in turn is a product of an "entrenched" management, in the words of Michael Metz, an analyst with Oppenheimer and Co. Although Norris owns only 1 percent of Control Data's stock, he appears firmly in command of its board.

"If that means a management that will not change and is in single-minded pursuit of its own desire to run the operation . . . that's a bum rap," Price responded in the interview. The cutbacks in operations and personnel are evidence of the change under way, he said.

"But some businesses do take longer to come to fruition, and we do have the tenacity to give things a fair shake. That's not the same thing as an unwillingness to change."

Few managements can match Control Data's tenacious patience with its long-shot ventures. Wall Street analyst Stephen T. McClellan notes that Norris' social commitment has, at various times, taken the company into Indian reservation health maintenance, day-care centers, Romanian manufacturing plants, small-farm financing and executive health clinics. Its computer courses offer education and job training to prisoners, college and high-school students, farmers, industrial and inner city workers. But the financial returns from PLATO, Control Data's long-term, $900 million investment in computerized education, are only now beginning to show up.

"Control Data is really like several government agencies: The Health and Human Services Department, the Education Department, the Small Business Administration, the Bureau of Indian Affairs, the State Department, and the Department of Agriculture -- except that Control Data is trying to make a little money at the same time. Often that's what it fails to do to the degree that it should or could," said McClellan in his book, "The Coming Computer Industry Shakeout."

In Norris' long-term view, that varied soil holds the seeds of the company's future, particularly in its commitment to computer-based education.

"The greatest market in the world is educational training and I think even the most skeptical now would no longer argue that computer-based education isn't going to be the major answer. That's pretty obvious.. . . My god, it's the world, almost. So the challenge is to meet that enormous market need and make money out of it, and we're learning how to do that."

"We are not in the process of changing the focus of Control Data Corp., we are in the process of narrowing the focus," added Price. "We truly are in the business of offering solutions to all kinds of problems and needs, from engineering design to education to mundane things like payrolls.

"We do that to make money. That's why we're in business. Our problem is not so much that those things are misguided. . . . The thing that is true is that we have been trying to do too many things," Price added.

"Very simply, we do not have the resources, in people, talent, time or money to do so many things, so we must do fewer.

"We don't intend to stop trying to use computers to solve problems and that includes things that are popularly referred to as unmet needs in the society. . . . The need is as great as it ever was. So the potential is as great as it ever was."

But Price suggested that Control Data's financial crisis has imposed a harder test on all of the company's ventures.

"It's a question of evaluating, within the financial resources of the company, how many and of what kinds of things we can do, and which the best ones are," Price said.

"I'll tell you, we will do what is necessary, what we think at any given point in time is the best thing strategically in the long term for Control Data," he added. "And there are no sacred cows. Not peripherals. Not social programs. Not computers. Not supercomputers. They are all part of being in business. And they are all evaluated. And whoever comes out last on the totem poll at any given point in time has got something to prove."