scorecardresearch Skip to main content
Innovation economy

Did computers get dumped in a pond at Digital Equipment’s Maynard complex?

It depends on who’s telling the story of how engineers used to deal with setbacks at New England’s biggest tech employer of all time.

A archival view of the Digital Equipment Corporation building in Maynard, and the pond in question.Courtesy of the Maynard Historical Society

We’ve all had moments where we’ve wanted to pitch a piece of uncooperative technology out the window.

That’s why I relished hearing a story, back in 2009, about technicians at Digital Equipment Corporation, the now-extinct computer company, and how they would shove non-working computers out of their building into a mill pond below.

Digital was once the second-biggest computer maker in the world, after IBM. It made pretty sizable machines — even though they were called minicomputers. I could imagine the splash that these fridge and minifridge-size devices would make hitting the surface of the placid pond in Maynard.

This year is the 25th anniversary of Digital’s acquisition by Compaq, which marked the end of the company’s four-decade run as a prime mover in the computing industry, and the biggest tech employer New England has ever seen. It felt like a good time to try to figure out whether minicomputer defenestration was a real thing that happened in Maynard — or just an urban legend that circulated among Digital’s alumni.

One of the first people I reached out to was Glenn Rifkin, who co-wrote a biography of Digital’s founder, Ken Olsen, “The Ultimate Entrepreneur.” Neither Rifkin nor his co-author had heard the story about minicomputers, or parts of minicomputers, taking a swim. Neither had Rifkin’s wife, who worked at the complex in the 1990s.


At the Maynard Historical Society, I connected with another former Digital employee, David Griffin, who wrote via e-mail that the computer-tossing tale “certainly paints a colorful picture.” But he added, “the mill pond has been drained several times and no junk pile was ever seen.” Griffin said that during the period when minicomputers were built in the mill — roughly 1960 to 1980 — “components and computers were extremely expensive and were not thought of as disposable. That’s a more 21st century attitude.”


In a Facebook group created by Digital alumni — at its peak, more than 120,000 people worked for the company — there were broadly four kinds of responses. The first: how did you get into our private alumni group? The second: why don’t you jump in the pond to find out? The third: the famously frugal Olsen would’ve never allowed people to waste resources like that. And the fourth: this did happen, I did it, and I remember seeing memos that were written by executives trying to discourage the practice.

The VAXstation 3500 workstation in 1987.

One of those executives was the former head of research and development at Digital, Gordon Bell. He’s now 88, retired from a storied career in high tech, and splitting his time between San Francisco and Coronado, California.

Bell began his career at Digital in 1960. While the windows at the mill complex could open, and were pretty big, Bell said, it would’ve been impossible to get one of Digital’s PDP-8 minicomputers through them.

But Bell did say that Digital’s minicomputers contained logic modules inside that were about the size of a playing card. In the early 1960s, the components on these modules were wired together by hand, by a cadre of women known as “Gloria’s Girls.” By the mid-1960s, machines took over the wiring.

“If you’d get one that you couldn’t explain that didn’t work, and the guys didn’t want to recycle it or have it repaired, they’d just toss it out,” Bell said. “I’m sure there are a bunch of modules right outside the window.” But Bell never actually saw it happen. And he doesn’t remember writing a memo discouraging the practice — though he says Olsen would’ve frowned on it.


Bell suggested that I call another Digital alum, Grant Saviers, who started working at the company in 1968 and rose to vice president in charge of personal computers and peripherals. (Digital never successfully sold inexpensive personal computers; its core strengths were in higher-priced products for businesses.)

Saviers agreed with Bell that it likely wasn’t a whole computer that got ditched, but rather lots of those logic modules. “I don’t recall ever seeing it, but I had technicians working for me who’d done it or seen it done,” said Saviers, now retired and living in Seattle. “In management, I would’ve said, ‘Don’t do that.’”

Saviers told me that there was a memo written by Olsen, possibly in the early 1960s, “that said if you throw a module out the window, you’ll be fired. Who knows if it stopped it?” At some point, the mill was air-conditioned, so the windows were closed more often, and the modules got very big – “the size of legal paper and larger,” Saviers said, “so it was an expensive proposition” to toss one.

Kenneth Olsen, chairman of Digital Equipment Corp., in front of the company's Maynard headquarters on July 11, 1988. Merlin Archive

Some of those memos are at the Computer History Museum, and others are at Harvard Business School’s Baker Library. Tim Mahoney, a librarian there, was kind enough to do a search for me — but came up empty. One problem: Olsen’s memos from the heyday of Digital Equipment Corp. are not yet digitized.


Rather randomly, I met a woman at a conference in 2021 who told me that her ex-husband had occasionally told the story of throwing a minicomputer out the window.

I tracked down the man in question, Joseph Zeh, and he said that no minicomputers ever made the leap. But he did write back that “if they ever dredged the mill pond or the Assabet River around the mill buildings, they would uncover enough modules, power supplies, I/O devices, etc. to build a hundred computers.”

When I rang him up on the phone, Zeh told me he worked at Digital from 1965 to 1991, and for much of that time, “it was just the greatest company to work for you could imagine.” He added that “in its later years, it got pretty political.” Yes, he actually saw people toss modules out the window. It wasn’t a regular practice, he said, and this was an era, of course, when people didn’t think twice about tossing a Coke bottle out the window of their car.

Zeh worked on assembling and testing the minicomputers before they were shipped to customers. If a module wasn’t working, it would get sent back to the test group. If the test group couldn’t identify the problem or fix it, Zeh said, they’d put a tag on it that said “NPF,” for “no problem found.” Sometimes, the assembly crew would send it back again when it caused snafus a second time. Sometimes, if a module had three or more tags on it, “we’d end up occasionally throwing them out the window, and the problem never came back,” Zeh said.


Did Zeh himself ever throw one out?

“We were like 21 years old,” he said. “Young people who may get frustrated do crazy things.” Then, there was a pause. “I guess I threw one or two in.”

Cal Calamari, who spent much of the 1980s and 1990s working as a manager at Digital, recalled that on his very first visit to the mill in the 1970s, he looked on as an engineer tossed out a module that seemed unfixable. “The problem was, it was winter and the pond was frozen,” Calamari wrote. “He then picked up a heavy power supply and threw it out to break the ice.”

Myth confirmed.

Scott Kirsner can be reached at Follow him @ScottKirsner.