Band of Brothers

A Philadelphia Inquirer help-wanted ad for an office manager led Maria Yeo through the door of Bentley Systems’ original office suite in West Philadelphia’s University City Science Center. It was a fresh day in March 1986, and the first thing she noticed was the view: drywall in every direction. “One thing I remember is that there were no windows in our office,” she would recall. “So in order to be able to see outside, you had to go out in the hallways and walk around to find a window.”

After interviewing with Scott Bentley—who’d chosen the space—she felt immediately that she’d landed in a special place. “I could tell from the beginning that this was going to be different than anywhere else I’d worked. You could just tell how excited they were and how seriously they took it.”

At the beginning, she was one of six or seven colleagues—and had a dubious distinction. “Scott always used to say that I was the adult in the company, because at the time I was older than anyone else,” she laughed. “But he was a very good boss. He was really very patient. He wanted things done right, and he would be willing to explain to you how to do things. And there was a lot of learning to do.”

That was perhaps an understatement because Bentley Systems’ first year in Philadelphia had been a whirlwind.

Scott Bentley had some professional experience in programming and systems development when he joined his brothers. He’d spent the previous four years working for a company called STSC, where he helped develop a system used by the Third Federal Reserve District for allocating the costs of paper check clearing. But Barry and Keith brought him on board primarily to take charge of business functions. “They wanted somebody to ‘deal with lawyers, accountants, and UPS,’” recalled Scott, who also came to oversee marketing and sales.

Tasked with setting up the company’s operations, Scott rented office space at the University City Science Center, an urban research park and entrepreneurship incubator situated a few blocks north of the University of Pennsylvania, from whose Wharton School he’d earned a bachelor’s degree in economics and decision science in 1979. What the digs lacked in the way of natural light they made up for by way of another fortuitous circumstance.

“Directly below us was a company that had an Intergraph VAX,” Keith said. “I’m not sure that was a complete coincidence, but it turned out to be great, because we bought time on their computer, and ran a wire down.” 

It was not, in fact, a coincidence. Having visited Keith and Barry in Pasadena to hear them out about joining the team, Scott recognized the value they’d derived from Caltech’s VAX. So after Keith decided to move operations back East to be closer to Rene, Scott sleuthed out an engineering company in Philadelphia. that had “the smallest Intergraph system you could possibly have,” and negotiated a sublease of space and computing time. The company was called Kuljian, and finding them was surprisingly easy. “Intergraph, God bless them, published the names, phone numbers, and telex numbers of all their customers,” Scott recalled. “And that’s how we got all of our original customers.”

The convenient access to computing power begged for another pair of hands in the coding department. So while Barry wrapped up his doctoral work in Pasadena—where he also oversaw the 1986 acquisition of Dynamic Solutions by the Waters Corporation, a Massachusetts-based laboratory instruments company—he and Keith made a bid to attract their youngest brother.

Ray, who had earned a degree in mechanical engineering at Rensselaer Polytechnic Institute and a master’s in computer engineering at the University of Cincinnati, was working in General Electric’s jet engine division in Ohio. Although he’d had some formal education in computer graphics and the FORTRAN language, his older brothers considered him a novice programmer. But they also knew Ray to be a quick study, and thought he might be able to tackle a new project they were calling PseudoPlot.

“It was similar to PseudoStation in concept,” Barry said, “in that, you’d be using a cheap pen plotter instead of having to use the $40,000 plotters that people used to have to buy.”

Keith and Barry knew how to get their brother’s attention. In lieu of a job interview, they threw down a gauntlet. “As a mutual tryout, we asked him to write a software package to plot designs to an HP pen plotter,” Barry said. “We mailed him the Kernighan and Ritchie C Programming book, an HP pen plotter, and the vague manual for it.”

“It was something I’d never seen before,” remembered Ray, who relished being thrown into deep water without a raft. “But I’m pretty good at figuring stuff out. I mean, you try, it doesn’t work, you try something else. If nothing else, my father taught us to be persistent.”

“He had a few questions,” Barry recalled, “but characteristically, he got about halfway through the book and the manual and simply started pounding code. In the course of a few weekends, he had beaten the plotter into submission and was actually generating some plots.”

“Raymond,” Barry concluded, “is the greatest software starter I’ve ever seen.”

However, the question of whether to leave GE was another matter. Ray had just finished a company training program designed to groom promising young talent to climb the management ladder. He’d just married Terry—and they’d just bought a house in Ohio. Meanwhile, Bonnie Bentley shuddered at the thought of yet another of her hatchlings crowding into the same precarious nest. “It was a big risk,” Ray said. “I had a mortgage and everything.” Yet the prospect of working alongside his brothers was tantalizing, and moving to the Philadelphia area would also bring them closer to Terry’s family.

The youngest Bentley brother had a decision to make.

“I thought for about 15 seconds,” Ray recalled, “and I said, ‘Hell yeah, let’s do it!’”

Four of the five Bentley brothers were now all-in. Greg remained on the board of directors but was charting an entrepreneurial path of his own in a different corner of the nascent software industry. In 1983, he and two colleagues had founded a company called Devon Systems International to develop currency-options trading software.

While Keith and Ray kept refining PseudoStation and PseudoPlot, Barry took responsibility for charting a path toward what all the brothers recognized as the inevitable, and far more lucrative, next platform for CAD: personal computers. He evaluated several models and concluded that the IBM PC AT represented a turning point for the possible democratization of CAD. Designed around the 16-bit Intel 80286 microprocessor, it offered a significant jump in computational power from its Intel 8088-based predecessors at a comparable price: roughly $5,000 (equivalent to $14,500 in 2024).¹ Ina1985lettertoKeithandRay describing the results of his benchmark tests, Barry surmised that “the AT with a good hard disk has plenty of balls to do this job.” So they bought a few, and Keith began working on the inaugural version of what would be called MicroStation.

He went about it with characteristic intensity. “Keith was working 100-hour weeks, I’m pretty sure,” Scott remembered. “For about six months he was living in my house—but I never saw him there. He was always either working or sleeping.”

Bentley Systems sold about 350 copies of PseudoStation in its second year, bringing in welcome revenue. However, the company’s focus definitively shifted in January 1986, when they demonstrated a view-only version of MicroStation to DuPont. By that time, DuPont was already working on their own internal project to do a stand-alone CAD system on a microcomputer, but they’d chosen to develop it around IBM’s CS9000, whose high cost ultimately doomed it in a marketplace that was soon blooming with cheaper PC clones. Whatever the case, DuPont saw enough promise in MicroStation to begin using it internally through a deal that helped Bentley Systems finance further development.

With that reference account in their back pocket, things really began to accelerate. Barry moved to Philadelphia in March 1986. In April, having run out of available brothers to hire, they welcomed Steve Knipmeyer into the fold.

Knipmeyer came from DuPont, where he’d charted a similar trajectory to Keith’s. Working in a sister group focused on engineering analysis applications, Steve helped to develop a graphics system to display Intergraph design files on the IBM CS9000. “That’s how I got to know Keith,” he recalled. “We were right next to each other. He made a huge impression on me—as did every brother who subsequently entered my life. Keith was clearly someone with exceptional capability. He was quiet but radiated expert competency.”

Steve’s motivations were also familiar. He’d been in a fast-track position at DuPont designed to propel promising junior employees through the ranks. “It was very secure. I had a bright future,” he said, “but what I really wanted to do was work for a software company. The engineering department at DuPont was a staff function, not a revenue center. At that time I was just completely fascinated and absorbed with small computers. I was getting restless.” So he called Keith. “Bentley Systems was a very tiny startup, but I really didn’t care. I wanted to do software. It was a huge risk, but I couldn’t think of a better person to align my future with.”

To the brothers, the “very first non-Bentley development employee” cut a striking figure. Standing six feet and three inches tall, Knipmeyer was a supremely composed character who formed fast bonds with an agile sense of humor, minting inside jokes that ran from dry quips to boyish braggadocio. “You know, we were a bunch of nerds,” as Barry put it. “Steve could pass for a business guy any day.” Or as Ray liked to jest at the time, with the playful irreverence that was his trademark, “One of the things that Steve does so well is wearing a nine-piece suit.”

“I wouldn’t really have characterized Barry or his brothers as nerds,” Knipmeyer reflected. “But I think what they were driving at is that I had dropped into a young Bentley Systems right out of corporate America. I knew how to wear a suit, and that got me some credibility. “I was comfortable in that world, the corporate world,” he continued, “but very happy to be with guys I felt a common connection with. We’d all grown up in northern Delaware. Our fathers had had careers at DuPont. We had the same values, the same worldview about almost everything.” Working with them was an eye-opening experience.

“It was exhilarating. I thought that my circle at previous employers were very capable, but this was a completely world-changing experience
to work alongside those guys. Keith and his brothers are absolute world-class software developers. They occupy a very rarefied niche,” he said. “It was pretty humbling to realize that, wow, you need to recalibrate your entire worldview about where you stand in terms of engineering. And that happened over and over again, through the years, as people came in.”

The next one through the door would be able to savor a window view, because presently the brothers bought an Intergraph VAX of their own and moved Bentley Systems to Lionville, Pennsylvania, establishing new headquarters in a building on Gordon Drive. Despite what Scott called the “insane margins” the company was making, this was still a time when the very idea of a software business was hard for many old-economy traditionalists to swallow. “The bank completely disregarded Keith and Barry,” Scott laughed, “because they didn’t own real estate. But I owned either half or all of a $42,000 twin in Philadelphia. So, therefore, I was creditworthy. So, we borrowed $75,000 and bought that computer.”

The VAX purchase signified how closely the company’s fate was tied to Intergraph. That was in some respects sensible. In 1985 Intergraph had surpassed its “Big 5” competitors to become the second-largest CAD vendor after IBM, with revenues of $526 million, according to CAD historian David Weisberg.2 Its client list included some of the biggest names in the architecture/ engineering/construction (AEC) services sector, and it enjoyed a particularly commanding presence in mapping and highway design. That gave PseudoStation and MicroStation well-paved inroads into the top end of the CAD market.

Yet it also made them vulnerable. Insofar as PseudoStation enhanced the appeal of Intergraph’s hardware, Bentley could be a friend to the bigger company. However, if a prospective customer decided to buy PseudoStation in lieu of buying an Intergraph workstation, Bentley represented a foe. And a 500-pound grizzly like Intergraph hardly lacked the means to see off a fox that got too crafty. Be it through litigation or competition, “we were always sure that Intergraph would somehow squash us,” said Ray. The menacing rumble of the apex predator reached the Bentleys’ ears in the spring of 1986. While attending a local Intergraph user group meeting in Toledo, Ohio, Keith and Barry heard “some pretty scary rumors” that “Intergraph had bought 50 percent of a company that had the same obvious idea that we had, which was that you could run a CAD program on PCs that was compatible with the Intergraph DGN.” Further digging proved the rumors true.

A company called CNR Research had developed a software package called C-CADD, furnishing Intergraph with a chance to kill MicroStation and eat Bentley Systems’ lunch.

“We were worried,” said Barry.

“If these CNR guys were any good,” Ray feared, “there was no way we could compete.” So, he worked the phone to scare up a demo version of C-CADD, and the brothers held their breath as they waited to see what they were up against. “Keith loaded it up on his computer the very night we got it,” Barry recalled, “and it took him about eight minutes to break their half-witted hardware protection scheme.” Then, while his brothers slept, the founder and CEO worked into the wee hours exploring C-CADD’s capabilities. By morning, he was ready to lead Barry and Ray on an unforgettable guided tour. It was like the C.F. Braun episode all over again but with a flipped script. One of the first things they noticed was that none of the text looked right, because the programmers had unaccountably used the wrong font. And the shortcomings snowballed from there. “Some of the more advanced elements didn’t even draw,” Barry recalled. “The user interface was absolutely pathetic. It crashed a bunch of times on us.” The brothers were soon giddy. “We could carve a better program out of a banana,” Barry crowed.

“Nevertheless,” he added, “they really did represent an existential threat to Bentley Systems—because maybe they’d hire some real programmers, and maybe Intergraph would put some guys on it and make it better.” That meant that MicroStation needed a serious upgrade.

“We had done a good job of showing files. We could display them at least 10 times as fast as C-CADD,” Barry said. But they’d only just started coding the capacity to actually author files—“to place lines and circles and arcs and all that kind of stuff.”

“We had a lot of work to do before we were ready to sell a full-fledged package,” he said. But the C-CADD demo had swelled their sails with confidence, energy, and an unmistakable bravado. With the clock ticking and the pressure on, the band of brothers—of which Steve Knipmeyer was now an indisputable part—knew there was only one path forward: “We decided to show Intergraph how real men program.

”For Barry, the next six months were “the most fun I ever had in my life.” Working around the clock, the four developers made dramatic strides. “I can’t even describe how exhilarating it was to be part of a team working with our backs to the wall—in a competition we knew we could win.”

That’s when we knew: We can beat these guys.

They extended MicroStation to enable users to create lines, ellipses, arcs, cones, cylinders, and closed shapes, which could be stretched or combined into more complex geometric forms.3 For Intergraph users accustomed to the cumbersome process of drawing those elements—which would only pop into view after multiple points had been clicked on the screen so that it often took three or four tries to get the dimension and placement exactly right—Barry figured out a way to “rubber-band” an element so that its outlines were visible even as it was being formed, changing dynamically as a user moved the cursor. Ray pressed forward on 3D perspectives and functionality, developing a compute-efficient way to multiply geospatial coordinates by homogenous matrices that Barry wrote in assembly language. They called it the Honkermobile, after the wheeled cart that was constantly shuttling between the programmers, loaded with all kinds of stuff.

Ray was in fact gaining a reputation in the realm of 3D graphics—and novel problem solving more generally. Soon after the Space Shuttle Challenger disaster in January 1986, Bentley Systems received an unexpected call from a team working at NASA’s Marshall Space Flight Center. Having heard that PseudoStation could display Intergraph design files on Tektronix terminals, they wondered if it might be possible to overlay 3D models of the launch vehicle on video of the fatal flight, to detect deflections and other possible abnormalities. Keith said they could give it a shot—and handed it off to Ray, who did what he loved best, plunging into virgin territory.

“That was one of the first 3D designs we ever got going,” Ray reflected. It was a wicked challenge, especially given the computing resources of the day, to superimpose a 3D model onto the two-dimensional plane of a video whose angle and scale were constantly changing. “But Raymond managed to get the thing to line up,” said Barry, providing another confirmation of the growing suspicion among experts about what had gone wrong on that cold winter day: “It showed a spume of gas coming out right where the O-ring on the model was located.”

Every week, Keith or Barry or Ray or Steve would add something that made MicroStation “work a little better”—or faster or more efficiently, a crucial consideration for an application that would need to run on 640 kilobytes of RAM.

“By August we were way ahead of them,” Barry recalled. But that didn’t solve the other half of the problem. They didn’t just need MicroStation to be better; they needed Intergraph to give up on C-CADD, which was presumably being improved as well. The strategy they hit on was to give demos
to select PseudoStation customers—knowing that the word would get back to Intergraph about what Bentley Systems was doing.

The speed with which that effort gained momentum became clear one day when an international telex arrived from Bayer AG, the German pharmaceutical giant behind Aspirin. How one of the biggest Intergraph customers in Europe had found out about Bentley Systems was something of a mystery— which meant that the word-of-mouth campaign was clicking. Whatever the case, now Bayer wanted a demo. And they wanted one badly enough, it turned out, to follow up their telex with a DHL package addressed to Scott Bentley containing two business-class airplane tickets to Germany.

I’d been out of the country once in my life, maybe twice,” Scott recalled. “And I knew nothing about MicroStation ... I’d never touched it before.” But there was no question now but to go. So, Keith gave Scott a diskette, showed him how to lead the software through its paces, and the next thing he knew, Scott was in Leverkusen. “I was nothing but a courier,” Scott insisted, but a capable one was precisely what the moment called for. “They were so impressed that they became one of our biggest and earliest MicroStation customers.”

While Scott proceeded on the business- development front, Ray was busy trying to recruit another programmer to the team.

His name was George Dulchinos. They’d met during their first week as students at RPI, where Dulchinos gained a reputation for being the go-to guy for anyone who needed help with a computer project. Though he now had a terrific job and had just gotten engaged to be married, Dulchinos conceded that computer graphics sounded like it might be fun.

“He took a leap of faith,” Ray said. Yet as 1986 entered its final frame, the uncomfortable truth was that his friend was in fact hovering in midair without a net. News of MicroStation had indeed reached Intergraph—which issued a forceful response to its customers. In so many words, Barry recalled, it amounted to: “Cease all contact with Bentley. Use the CNR solution instead.”

At some point, Ray was overcome by a sense of obligation to his friend. “I remember calling him up and saying, ‘George, this thing is not looking that great. First, we might get sued. And now they’ve bought another company to do the same thing...’”

In Ray’s recollection, a steely voice came from the other end of the line: “‘Well, I already quit,’” George told him. “’And I can’t go back.’”