Interview with Bob Taylor, 1999
Another interview I did for the 1999 MIT Press edition of my 1985 book, Tools for Thought.
Interview with Bob Taylor, August 2, 1999
By Howard Rheingold
Bob Taylor, who died in 2017, probably had direct authority over, or was directly involved in more world-changing inventions than anyone since John von Neumann. He and the late JCR Licklider wrote presciently about the future of networked computers as early as 1968, and initiated the project that became the ARPAnet when Taylor was a program manager for ARPA. Later, at Xerox PARC, Taylor led the team that created the first graphical user interface, point and click word processor, the Ethernet local area network, the laser printer, desktop publishing, and internetworking. Although books and articles have since recognized that Bill Gates took the idea that turned into Windows from Apple, whose founder Steve Jobs took from PARC the idea of a graphical user interface that turned into the Macintosh. To hear the story of ARPA and PARC directly from Taylor, watch this interview with John Markoff at the University of Texas.
In 1984, just before the publication of Tools for Thought,Taylor and a number of key researchers from his PARC team moved out of Xerox entirely and built the “Systems Research Center” (SRC) at Digital Equipment comany (DEC). The SRC team beat the first truly powerful web search engine, AltaVista. They created a networking scheme for realtime video and audio over a global network, created Modula 3, a language that became the precursor to Java.
“In the long run, DEC made sort of the same mistakes IBM made in the sense that IBM ignored the minicomputer that DEC took advantage of,” says Taylor today. “DEC later ignored the personal computer that was manifested by Apple, Sun, and others. DEC did take advantage of Ethernet but didn't go on to do the other pieces of internetworking properly. They were still wrapped up in time-sharing and didn't see that personal computing was going to surpass time-sharing, the big business of the 1970s.. DEC ran out of gas and Compaq bought it.” Taylor retired in 1996, and still stays in touch with much of the remarkable research community he led nearly thirty years.
From Taylor’s “Computer Systems Laboratory” at Xerox PARC came Bob Metcalfe, the founder of 3COM. John Warnock founded Adobe. Charles Simonyi, who took the “Bravo” word processing software he created for the PARC Alto and transformed it into Microsoft’s “Word” product is a billionaire today, or close to it. Alan Kay is a Disney Fellow. “Some of my former colleagues work SRC, some at Microsoft. One is dean of school of computer science at CMU. Two founded Adobe. There's a great community out there.I am proud to be a part of it”
When I asked Taylor if history has overlooked anything significant, and he responded with a provocative assertion: “Nobody has picked up that the first Internet was also built at PARC. We connected PARC’s Ethernet and the ARPAnet in 1975. That’s what an internet is -- two or more separate networks connected. Vint Cerf and several others claim to be the father of the Internet. Cert pulled together a committee at Stanford around 1975 to create what eventually became TCP/IP, the fundamental protocols that enable computers to communicate over the Internet. He invited some people from PARC. Bob Metcalfe (who was largely responsible for the creation of Ethernet, and later founded 3COM), John Schoch (now a Silicon Valley venture capitalist), David Boggs, and others to help the Stanford committee with the design. PARC lawyers told our researchers that they couldn't tell the people at Stanford what we had been doing with a working Internet.”
According to the account Taylor gave me, the PARC researchers who had already solved the problem of internetworking met with the Stanford group who were trying to come up with the fundamental software architecture for connecting different packet-switched networks together. “The Xerox guys said things like “if you do that, what happens if this other thing occurs?” The Stanford guys would say, “I guess you're right. We need to rethink that design.” They went through this scenario several times, with the PARC guys questioning ideas that the Stanford team originated -- telling them what was wrong with their proposed technical solutions -- but because of their legal restraint=s, not laying out what they knew about exactly how to build an Internet. One of the Stanford guys finally said: " You guys have already done this." And indeed, Ed Taft had already completed PUP (Parc Universal Packet) at PARC before TCP/IP was created. Taft is now at Adobe.”
When I told him about Doug Engelbart’s and Alan Kay’s critical remarks about the design of the World Wide Web, Taylor said, “I agree that the Web is a bad implementation, but it's a miraculous piece of work. What Tim Berners Lee did was to make a whole bunch of software in other systems somehow or the other work together. The Web protocols didn’t create a great meal, but it did make an edible meal out of garbage. Yes there is a lot wrong with it, but it's remarkable nevertheless. In order for there to be an Internet, you need a PC and a local and a global network, helps if you understand desktop publishing and graphical interfaces. We put all those ingredients together by the mid to late seventies. But the Internet didn't take off until mid 90s. Why so long? A lot of people screwed up along the way, especially Xerox. Xerox had the opportunity to put reasonably inexpensive netoworked PCs into people homes when Apple had the opportunity. Apple took advantage of Xerox’s slip, but Apple blew networking. Xerox made some mistakes, Apple made some mistakes. We had all the pieces but nobody could put them together until 20-30 years later. So my own predictions were way off. I thought that by the end of the 70s that the world would have the computing and networking infrastructure we didn’t have about 1993. I missed by 20 years or more.”
What did the original designers do right? “A lot of people today are worried about the Internet becoming less democratic, perhaps centralized or under a few people's control. I don't think that's going to happen because of Wes Clarke, who came up with the idea of the IMP, the Intermediate Message Processors that enabled the packets to find their way around the network without central control. Larry Roberts and other people were talking about having ARPAnet controlled by a central computer in Nebraska. IMP makes it very difficult to centralized control the network. If there had been one computer controlling ARPAnet, we would have already had cause to worry. That single architectural foundation to packet-switched networks makes it difficult for anyone to control the Internet. Nobody ever calls Wes Clarke the father of the Internet, but he has as much right to that title as anyone who claims the title today.
“There haven't been any beneficial changes that surprised me. The porno stuff never occurred to me. That's a negative. I am still concerned about the haves and the have nots. Who is going to have access and who isn't and why. It is certainly based on economics. I hope we will get the cost of being connected low enough so that everyone can afford it. Almost anyone who wants today below the poverty line can buy a television.
What’s the next technoloical revolution that’s going to change the world in the foreseeable future? Taylor agreed with me that it was far easier to foresee technological futures in 1985 than it is today, but added: “I can see a continuing convergence of Internet, television, telephone communications, including mobile telephones, and even appliances. When every little gadget imaginable has a computer in it, and all those embedded computers are talking to all the others through local and wide area networks -- that’s going to make another huge change possible.
Lickider and Taylor were worried, thirty years ago, about whether the cost of access to the technologies they were creating would widen the gap between rich and poor. “I hope everyone is enabled to participate in this continual evolution of technology. One of the most serious world problems is the widening gap between rich and poor. Personal computer and network communications technology, properly managed, has a lot of potential to narrow that gap. The underprivileged and underfinanced do have access to information like never before in history. Twenty years ago, avery few people who had access a lot of money also had a lot of access to information. Everyone else had radio, newspaper, and other sources controlled by other people, and that was about it. There is still enormous promise that information could help have-nots become haves.”
The failure of the Xerox Corporation to capitalize on the astonishing ten year lead that PARC gave them in PCs and internetworking has been documented in books such as “Fumbling the Future” and the history of PARC’s largely unrecognized role in creating the basis for today’s wealthy high-tech industries was recently covered in a book, “Dealers in Lightning.” Taylor points out that “Xerox got a billion dollar business out of laser printers, which paid for all the research costs at PARC, and much more.”
Taylor doesn’t see a lot of growth in the kind of truly fundamental research that PARC pioneered. “Microsoft is doing PARC-like stuff, and IBM still has its research entity, but there isn't a great deal of PARC-like research. Chuck Thacker, Butler Lampson, Mark Brown are at Microsoft research. They really do have some very large investments in research, not just advanced development, which is what many =companies call research these days.” In other words, the company with the largest investment in “creating the future,” PARC-style, appears to be Microsoft.
There are forces at work to try to bring the lesson home to the government that ARPA succeeded wildly. A few proposals, if adopted, would enlarge govt research support in information processing especially.