by Katie Hafner
BBN needed the boost. Throughout the 1980s, the company’s fortunes had mostly ebbed. As the Internet had grown more popular, BBN, which time and again had failed to commercialize on its research efforts, had slipped into relative obscurity. In 1993 the company lost $32 million on $233 million in sales. The next year wasn’t much better, with an $8 million loss on lower sales.
The company had missed its greatest opportunity when it failed to enter the market for routers—of which IMPs were the progenitors. BBN failed to see the potential in routers much as AT&T had refused to acknowledge packet-switching. Anyone wanting to connect a local area network—of which there were now hundreds of thousands—to the Internet needed a router. By 1994, the router business was a multibillion-dollar industry. More than a decade earlier a couple of BBN’s own computer guys had tried to push the company into the router business, and they had been brushed off by a marketing vice president.
BBN’s troubles went beyond failed market opportunities. In 1980 the federal government accused the company of conspiring to overcharge the government on its contracts during the period from 1972 to 1978, and of altering time sheets to conceal the overcharges. The practice was discovered when, in the course of a routine audit in the late 1970s, BBN officials were less than candid with a government auditor. (“BBN had gotten very arrogant,” said one long-time employee.) A federal investigation lasted more than two years. Auditors moved into the firm’s Cambridge headquarters. Senior BBN employees were called before a grand jury. None of the IMP Guys was implicated. But in 1980 two of the company’s high-ranking financial officers plea-bargained their way out of a onehundred-count charge. They were given suspended sentences and fined $20,000 each. The company agreed to pay a $700,000 fine.
At the time, BBN depended on government contracts for nearly 80 percent of its revenues. Given the certainty that all government contract awards to BBN would have been suspended during the course of any lengthy legal defense, had no settlement been reached, the charges could have ruined the company. People in the company felt that the government had overreacted to incorrect accounting practices. BBN, they said, had always given the federal government much more than its money’s worth on contract R&D.
The networking group at BBN, only minimally involved in the government investigation, was simultaneously offering proof positive that government-funded science can bear splendid fruit. TheARPANET was Exhibit A. Funded entirely by ARPA, its creators given reasonably free rein, the network was evidence of a once-pervasive American trust in science. The network was built in an era when Washington provided a little guidance and a lot of faith.
By 1994, BBN’s brush with the government auditors was forgotten. Unfortunately, so was the company’s role in building theARPANET . Only those insiders who were acquainted with history associated the company with the newly popular Internet. When Conrades arrived, he decided it was time to polish BBN’s image. And a silveranniversary bash for theARPANETwas the perfect opportunity.
The invitation list for the party was as scrutinized as an invitation list for a White House dinner. Some names were obvious, of course; but scores of people had had a hand in building theARPANET, and even more people, from all over the world, had been involved with the Internet. Heart, Walden, and others submitted suggestions.Vice President Al Gore, an advocate of the information superhighway, was invited. So was Ed Markey, a Democratic congressman from Massachusetts who had also put the Internet on his political agenda; he accepted the invitation. Bill Gates was invited, although the Microsoft chairman had yet to acknowledge the Internet as a useful tool. He declined. Paul Baran, whose role was minimized at BBN, nearly wasn’t invited at all. In time, the list ballooned to five hundred invitees.
Conrades wanted this to be as much a signal for the future as a celebration of the past. He was planning for BBN to expand its somewhat diminished role in Internet-related businesses. BBN already owned and operated NEARnet, the New England regional network. One of his first moves after arriving was to purchase BARRnet, the regional network in the San Francisco Bay Area. And he had his eye on SURAnet, the regional network for the Southeast.
Seeking a lofty theme, the public relations firm that BBN hired to augment its own PR department came up with one: “History of the Future.” It suited Conrades’s plans for BBN perfectly. Conrades also hired a production company to put together an elaborate video presentation that would include interviews with a core group of pioneers—Larry Roberts, Bob Kahn, Steve Crocker, Len Kleinrock, Frank Heart, and Vint Cerf.
Bob Metcalfe, the inventor of Ethernet, was now publisher of a computer trade newspaper calledInfoWorld.He wrote an opinion column on the upcoming event that he titled, “Old Fogies to Duke It Out for Credit at Internet’s 25th Anniversary.” “I’ll be there to see old friends, to renew some old animosities, and to join in the jockeying for credit— of which there is plenty to go around,” Metcalfe wrote. “I’ll begin by making sure partygoers realize that most TCP/IP traffic is carried by Ethernet, which I invented . . . As the party peaks, I’ll see how much credit I can grab from Vint Cerf and Bob Kahn for the invention of internetworking . . . Failing that, I’ll see if I can smile my way into the group photo of the inventors of packet-switching.”
Weeks and days before the event, BBN’s public relations firm placed stories in magazines and newspapers.Newsweekran a lengthy piece on theARPANETpioneers, and so did theBoston Globe.BBN put together a video news clip, which was aired on more than one hundred local newscasts. Ray Tomlinson, whose scrutiny of his keyboard at an opportune moment had produced the @ sign, was celebrated as a folk hero in a story aired on National Public Radio on the evening of the party.
The guests of honor began arriving on Friday, September 9, and gathered for a reception followed by a press conference at the Copley Plaza that afternoon. As a joke, Wes Clark, now a consultant in New York, pinned Larry Roberts’s name badge to his own sport coat. At the press conference, several of theARPANETpioneers, who outnumbered the journalists, delivered speeches. In his speech, Bob Taylor wryly remarked that the people who had been invited had sent their grandfathers instead.
The most notable absence was that of Licklider, who died in 1990, but his wife, Louise, accepted the invitation. Bernie Cosell, the IMP team’s ace debugger, who now lived in ruralVirginia and raised sheep (“Too many people, too few sheep,” read Cosell’s e-mail signature), was unable to come because of the expense. Others refused. Famously averse to parties, Will Crowther declined the invitation; repeated phone calls from fellow IMP Guys could not change his mind.
At a stand-up Mexican buffet following the press conference, everyone mingled. Some people had seen each other a few days earlier, or a few months earlier, but others hadn’t seen each other for years, or even decades. New spouses, old spouses, and premature aging were quietly remarked upon. Larry Roberts, now running a small company that was building a new generation of switch, lived in Woodside, California, a well-to-do community on the San Francisco peninsula. Now fifty-eight, Roberts was on a daily regimen of “smart drugs” (Deprenyl, used to treat Parkinson’s, was one; melatonin was another) to regain the powers of concentration he possessed at twenty-eight. With characteristic intensity, he steeped himself in the subject. He read hundreds of research reports and had even produced an “anti-aging” videotape.
In 1983, Taylor had left Xerox’s Palo Alto Research Center. His departure had sparked a rash of resignations from loyal researchers, who followed him to Digital Equipment Corp., where he set up a research lab just a few miles from Xerox PARC. For years he lived around the corner from Larry Roberts and neither man knew it. One day a piece of mail addressed to Lawrence G. Roberts was misdelivered to Taylor’s house, and the two discovered that they lived a few hundred yards apart.
On day two of BBN’s fete, Saturday morning, came photo sessions, one after another. First was the official group shot. The group was large, about twenty-five people. When Len Kleinrock missed the first official photo shoot, another had to
be arranged. For another pose, the core group of IMP Guys was asked to pose precisely as in the original IMP Guys photo taken in 1969. “Could you guys lose some weight?” Cerf called out as the photographer tried to place everyone in the shot.
Afterward, the scientists were bused two blocks away to the Christian Science Center for aWiredmagazine shoot. Gamely, the nineteen men squeezed themselves onto a short, narrow bridge at the Center’s Mapparium. Being engineers, they couldn’t help but offer some advice to the photographer, who was having a little trouble fitting them all in the shot: Try a different angle. Try a different configuration. Try a different lens. Try a different camera. Later, a few in the group grumbled about hangers-on having shown up for the photo, but for the most part, the general good cheer of the weekend was beginning to infect them all.
The multiple paternity claims to the Internet (not only had each man been there at the start but each had made a contribution that he considered immeasurable) came out most noticeably that afternoon during a group interview with the Associated Press. The interview was done over a speakerphone in a suite at the hotel. Kahn, Heart, Engelbart, and Kleinrock sat hunched over the phone as the AP reporter asked questions. Before long the interview transformed into a study in credit management. Taylor arrived late, but not too late to engage in something of a dustup with Bob Kahn, who warned the AP reporter to be certain to distinguish between the early days of theARPANETand the Internet, and that it was the invention of TCP/IP that marked the true beginnings of internetworking. Not true, said Taylor. The Internet’s roots most certainly lay with theARPANET. The group around the telephone grew uncomfortable. “How about women?” asked the reporter, perhaps to break the silence. “Are there any female pioneers?” More silence.
The weekend was as noteworthy for who wasn’t present as for who was. Tim BernersLee, the inventor of the World Wide Web, had just moved to Boston from Geneva to join MIT’s Laboratory for Computer Science. He wasn’t invited, nor was Marc Andreessen, the co-programmer of Mosaic, who had just left Illinois to develop a commercial version of his Web browser. Granted, they hadn’t played roles in the birth of either theARPANETor the Internet (Andreessen wasn’t even born until 1972, after the firstARPANETnodes were installed) and couldn’t technically be counted as founders. But they were behind the two inventions that were already giving the Net its biggest reach into everyday life.
Three years earlier, the NSF had lifted restrictions against commercial use of the Internet, and now you could get rich not just by inventing a gateway to the Net but by taking business itself onto the Net. When reporters asked them to comment on this, some of the originalARPANETbuilders said they found the new commercialization of the Net lamentable. Others welcomed it.
Few of the pioneers had become wealthy. Metcalfe’s invention of Ethernet had made him a multimillionaire. In his nearly thirty years as a computer science professor at UCLA, Kleinrock had guided an army of Ph.D. students, many of whom went on to become luminaries in the field of computer networking. Still teaching at UCLA, Kleinrock ran a successful seminar business on the side. But at the other end of the spectrum was Jon Postel, the unsung hero of networking. He observed the weekend of celebration quietly, much as he had worked for years as keeper of the RFCs and final arbiter in technical matters when consensus couldn’t be reached. Postel believed that decisions he had made in the course of his work over the years had been for the good of the community, and that starting a company to profit from those activities would have amounted to a violation of public trust.
Most of the IMP Guys had ended up in BBN’s senior management. Walden later served as Heart’s boss, and Barker had gone on to run one of BBN’s divisions. The most conspicuous exception to this was Crowther, who had remained a programmer. For years Heart had been Crowther’s champion, lobbying for the company to let Crowther just be Crowther and think up ingenious ideas in his own dreamy way. In the years following the IMP project, Crowther pursued some unusual ideas about natural language processing, and worked extensively on high-speed packet-switching technology.
Severo Ornstein had left BBN in the 1970s for Xerox PARC, and while there he started Computer Professionals for Social Responsibility. When he retired from Xerox, he and his wife moved into one of the remotest corners of the San Francisco Bay Area. For years Ornstein stayed off the Net, and for years he eschewed e-mail.
Of everyone, Vint Cerf was perhaps the most celebrated this weekend. He was the person most of the press turned to for quotes on the Internet’s origins. In early 1994 he had left Kahn’s Corporation for National Research Initiatives to return to MCI as a senior vice president and help build the company’s Internet businesses. His reputation was well known throughout the company. At an MCI operations center in North Carolina someone had hung a sign: “Vint Cerf is the Father of the Internet, but we’re the mothers that have to make it work!”
As Saturday’s dinner approached, there was a great last-minute rush to make certain that the script for the evening would strike just the right tone. Conrades was to be emcee for the main event. No one should get short shrift at the expense of anyone else. It was a nearly impossible task. At the last minute, seating assignments were shuffled yet again. When the dinner finally commenced, some 250 people packed the Grand Ballroom. Lobbying for a telecommunications bill that was pending, Congressman Markey gave a humorous speech. Cerf presented two awards, and Kahn was singled out for his lifetime of achievement. Louise Licklider, frail and elderly, stood up to receive an extended round of applause on behalf of her late husband.
The celebration held a special poignancy for Heart, now sixty-five, who had recently retired as president of BBN’s Systems and Technology Division. He had been at BBN for twenty-eight years. Steve Levy, the BBN chairman, called Heart to the podium, and Heart gave a speech in which he pinpointed the reasons theARPANET project had succeeded. “The project was an example of what can be accomplished quickly, with a really strong sophisticated leadership, adequate resources, and an avoidance of the many kinds of bureaucratic foolishness that can affect so many projects.” Roberts had seen to that. Heart ended on a high note. “Only a small fraction of the technically trained population get a shot at riding a technological rocket, and then get to see that revolution change the world.” The networking revolution, Heart said, would rank among a small number of the most important technological changes of the century. And that night, in that room, it looked like he might be right.
The next morning, one by one, theARPANETpioneers said their goodbyes and checked out of the hotel. Everyone was still riding high, at least a little bit. It hadn’t been a bad couple of days.
Chapter Notes
Chapter One
The description of the formation of the Advanced Research Projects Agency and its Information Processing Techniques Office is derived from personal interviews, two books written by Eisenhower’s science advisor James Killian (see bibliography), magazine articles, and from an excellent and thorough history of the agency commissioned by DARPA and written by Richard J. Barber Associates in 1975. The description of Licklider’s early years is based on talks given in his honor, on interviews with Louise Licklider and Bill McGill, and on an obituary written by Karl D. Kryter. The description of Licklider’s introduction to computers is based on personal interviews with Wes Clark and Jack Ruina, and on Licklider’s interview with the Charles Babbage Institute, as well as the Barber Associates report.
Chapter Two
The description of Paul Baran’s work on distributed communications is based on personal interviews with Baran, as well as various interviews conducted by the Babbage Institute. The description of Donald Davies’s early work on packet-switching is based on interviews and correspondence with Donald Davies, and on Martin Campbell-Kelly’s articles and interviews. Arthur Norberg and Judy O’Neill’s awesome report, “A History of the Information Processing Techniques Office of the Defense Advanced Research Projects Agency” also guided us through biographical material and through the
early years of IPTO. The report served as the basis for their bookTransforming Computer Technology: Information Processing for the Pentagon,1962–1986(Johns Hopkins University Press, 1996).
Paul Baran and Professor Manley Irwin supplied the information concerning AT&T’s lawsuits against parties thought to be depriving AT&T of revenues.
Doug Engelbart’s patent for his mouse came on November 17, 1970. It is patent No. 3,541,541.
Chapter Three
The history of Bolt Beranek and Newman was based on personal interviews with Dick Bolt and Leo Beranek. The description of Lincoln Laboratory was based on interviews with Wes Clark, Frank Heart, Larry Roberts, and Len Kleinrock. The description of events surrounding the request for proposals for the Interface Message Processor was based on interviews with Jerome Elkind, Frank Heart, Dave Walden, and Severo Ornstein, as well as various interviews conducted by the Charles Babbage Institute.
Chapter Four
The description of the building of the Interface Message Processor was based on interviews with Dave Walden, Ben Barker, Severo Ornstein, Bob Kahn, Frank Heart, Alex McKenzie, and Will Crowther. We also relied on ARPA’s request for proposals and on BBN’s proposal.
Chapter Five
The description of UCLA’s preparations for and receipt of the IMP Number One is based on interviews with Len Kleinrock, Steve Crocker, Mike Wingfield, andVint Cerf. The description of the Network Working Group’s early work in layered protocols was based on interviews with Steve Crocker, Jon Postel, andVint Cerf. The description of the first log-in session between UCLA and SRI was based on interviews with Len Kleinrock and Charley Kline. The description of the gridlock tests conducted in early 1970 was based on interviews with Bob Kahn, Dave Walden, andVint Cerf.