(FORTUNE Small Business) – Edson D. De Castro built a company that was great and huge and wildly successful by every measure, except for one: It disappeared. In fact, it's now hard to find any trace of it. "It is part of the food chain, and it has been swallowed up," concedes the co-founder of Data General Corp., a computer maker that peaked with annual sales of $1.3 billion in 1988. "I'm not really convinced that 100 years from now anyone is going to have any record of its existence."

That's a grim fate for a company that helped usher in the age of the PC. De Castro built a successful product--minicomputers that were cheaper and easier to use than their clunky mainframe predecessors--and for a while that was enough. Annual sales reached $100 million in 1975, seven years after the company's founding in Hudson, Mass. In 1984 it sold $1 billion worth of the powerful machines. In addition, de Castro's defection from computer giant Digital Equipment Corp. inspired others in the area to set off on their own, igniting a statewide entrepreneurial explosion that came to be called the "Massachusetts Miracle." Wang Laboratories, one of the other highfliers, dominated the market for stand-alone word-processing machines, and Digital itself rose to become the nation's second-largest computer company.

But one great idea won't guarantee that a business will achieve immortality on the order of Mattel or Gillette. Sometimes a company's greatness is only momentary. All those "Massachusetts Miracle" companies are gone now, victims of another generation of computer mavericks, which came up with smaller, nonproprietary machines first called microcomputers but better known today as PCs. The once--$3 billion Wang filed for bankruptcy in 1992. Digital dissolved into Compaq in 1998. And Data General, having accumulated $300 million in losses, fired de Castro in 1990. Eight years later what was left of the company was sold off. "I read about it in the newspaper," de Castro recalls. They all could have survived, de Castro says, had they only been smart enough to anticipate and adapt.

But nowadays de Castro is viewed as an entrepreneur who didn't have the skill to adapt to a world overrun by desktop machines. That leaves him with a reputation as an ex-success, otherwise known as a total failure. After all, gurus may celebrate companies that are built to last, or even built to flip. But hardly anybody speaks in favor of companies that serve a critical purpose for a limited time. "Minicomputers changed the whole dynamics of the marketplace," says David Lampe, author of two books about Boston's high-technology sector. "I don't judge Edson de Castro a failure. He's a great success--and you can tell him I said so." I could, but I sense that de Castro wouldn't care one way or another. He prides himself on being an engineer, exhibiting all the characteristic mushiness of the breed. "I'm a realist," he says, "and I think these are just the facts of life in business. I have nothing to cry about."

Today Data General is best remembered because one of its computer-building projects was chronicled in Tracy Kidder's The Soul of a New Machine, a book that won the 1982 Pulitzer Prize--and its business accomplishments are all but forgotten. Even the stoic de Castro admits that this is a disappointment. "Look, this is certainly not the outcome I would have preferred," he says. "You don't aim to have this happen. But one doesn't always have control over such things."

Indeed, the 64-year-old views himself as a victim of circumstances--some of which he'll readily take the blame for creating. There's no disputing that Data General was blown apart by the kind of technological tornado that touches down from nowhere. Or seems to, anyway. "I hadn't quite synthesized it all in my head as well and as early as I should have," he admits. "When things are going well, maybe you're not as introspective as you need to be."

Granted, the massive shift to personal computing was kind of a big trend to miss. But maybe it's unfair to judge a pioneer like de Castro as you would any ordinary entrepreneur, the kind who flits from venture to venture seeking to grow a fatter fortune and a larger legend. He spotted an opportunity in a field he understood, set up shop in a refurbished beauty salon, and ended up building a Fortune 500 company (No. 500, 1979). He didn't spend a minute studying entrepreneurship; apparently he didn't spend much more than that studying business either. He flunked out of Harvard Business School after just one year. "That was pretty traumatic," he recalls. So was his downfall, which perhaps should have been expected. After all, is it really fair to assume that the same person who spots the potential of a new technology should also discover the technology that's going to replace it?

De Castro says he might have been able to re-soul the machine maker had his board known enough to give him a chance to move the company into PCs, software, and workstations. But the board wasn't willing to put up with the years of losses the plan would have incurred in the short term. As he sees it, the group chose instead to die slowly. "The most important mistake I made was in picking the board," he says. "They didn't know anything about the computer industry." Ironically, he had selected them for just that reason, to balance his own knowledge. But after what happened, he says, "I don't have a lot of respect for any of them."

He's also not crazy about the get-rich-quicker mindset so prevalent among entrepreneurs in recent years. "Times are different," he says, with a shrug. "I was a CEO for 22 years, which is probably as long as anybody in this industry." He says he didn't make out "fantastically well" financially, although he still owned about 3% of the company when he left. He hadn't sold it, he says, because "I believed in the company. In the end I was a victim of my own convictions." With his reign abruptly cut short, he didn't simply latch on to another startup either. In the mid-1990s he spent 18 months running a small biotech firm in Boulder, but it "wasn't doing great" before it was sold. He's been director of another company he describes as "muddling along." As he'll admit, he hasn't exactly spent the last dozen years sifting through high-profile opportunities. "I didn't have a lot of headhunters chasing me to take over other companies. I wasn't a GE-trained manager, which seems to be the 'in' thing for people to want to drop into companies," he says. "My value was probably greater as a coach to young entrepreneurs." He has taken that role in the classroom, but also as a board member--and a particularly tough one at that. "That's my problem," he notes. "I always tell it like I see it."

And how should the rest of us see Data General's legacy? Maybe the most fiercely innovative companies simply can't help but flame out. "It would have been nice if his company had endured, but what Edson de Castro did lives and breathes in the fact that computers are even more broadly integrated into our society," says Lampe, whose books include The Massachusetts Miracle (MIT Press, 1988). Data General didn't last nearly as long as the U.S. companies profiled throughout this special issue of FSB. But it did leave a lasting influence. Shouldn't that be enough?