Interview of Albert Vezza
Total Page:16
File Type:pdf, Size:1020Kb
[3B2-9] man2012020008.3d 26/4/012 17:8 Page 2 Interviews Albert Vezza Editor: David Walden Albert Vezza is notable for his roles Siemens, not AT&T, modem. The dial-up network with J.C.R. Licklider, the Massa- didn’t support 1,200-baud full duplex, so I designed a chusetts Institute of Technology simple protocol and a little circuit that turned full du- (MIT) Laboratory for Computer plex into a half duplex, but we were not allowed to con- Science (originally Project MAC), nect ‘‘foreign equipment’’ to the dial-up network. An the ARPANET development, Info- AT&T line was leased along with time on ITT’s undersea com creation and management, cable to Europe’s dial-up network. However, ITT’s establishment of the World Wide onsite engineers in New York City refused to connect Web Consortium (W3C), and support of the Internet our leased line to the undersea cable. I phoned people Engineering Task Force (IETF).1 up the ITT chain of command, demanding connection, until I reached an executive vice president. The VP called MIT’s president, who suggested that ITT comply David Walden: When I first knew you, you were with my request. With our leased line connected, we working with J.C.R. Licklider in his Dynamic next had to convince operators at ITT and dial-up oper- Modeling Group at MIT. First off, how did you ators in London, Paris, and Berlin that a quiescent line come to MIT? was still in use so they wouldn’t disconnect the circuit. Albert Vezza: Upon graduation with a master’s degree Luckily, it all was working about three days before con- from Northeastern University’s Graduate School of ference started, and people around the institute knew Engineering, one of my job interviews was for a posi- of my efforts. tion at the MIT’s Electronics Systems Laboratory. The About that time, Rob and Tom were leaving to start a interviewer, a pilot in the Air Force Reserve, and display terminal company, the group was disbanding, I argued quite vigorously about the frequencies of Tac- and I had my CV ready to move on. But Licklider, tical Air Navigation systems that I’d worked on at who was taking over as director of Project MAC from Stromberg-Carlson. I left the interview discouraged Robert Fano, convinced me to stay at MIT and join a with my prospects for working at MIT. A couple of new group he was starting. days later I received a job offer from MIT. When the project for which I was initially hired Walden: The Dynamic Modeling Group? ended, I joined an Electronic Systems Laboratory team Vezza: I don’t know how it happened, but I soon found assigned to Project MAC. About that time, MIT Project myself managing the group—talking to students and MAC and the Technical University of Berlin were plan- staff members about projects that needed to be done, ning a conference in Berlin. Tom Cheek and Rob Stotz worrying about budgets, buying more equipment, and (our group leader) had developed a computer display so forth. Licklider didn’t have the time or inclination using a Tektronics storage tube that communicated at to deal with the daily activities of the group, but he 1,200 baud (full duplex) with the CTSS (Compatible was definitely the intellectual leader—a magnet for stu- Time-Sharing System). MAC’s Assistant Director Dick dents and staff members. Mills wanted to demonstrate that display at the confer- Lick had several objectives for starting the Dynamic ence. Other members of the group were busy; as the Modeling Group. One objective was the creation of dig- newest member of the group, I was brash enough to ital modules that could be connected tinker-toy fash- say that I’d do it. ion, the way analog computer modules are connected. An apparently simple task quickly turned com- He also wanted students and staff to become immersed plicated. For European compatibility, we needed a in real-time graphics experimentation and become part Background of Albert Vezza Born: 30 July 1934 Infocom, 1984–1986; Massachusetts Institute of Education: Rochester Institute of Technology, BS, Technology, 1986–1996; Corporation for National 1962; Northeastern University, MS, 1964. Research Initiatives, 1996–2005; Foretec, 1998– Professional Experience: US Navy, 1953–1957; 2005. Massachusetts Institute of Technology, 1964–1984; 2 IEEE Annals of the History of Computing Published by the IEEE Computer Society 1058-6180/12/$31.00 c 2012 IEEE [3B2-9] man2012020008.3d 26/4/012 17:8 Page 3 of the ARPANET experiment he knew was and said, ‘‘You better make sure it works.’’ coming. Along the way the experimental That’s all he said. I said, ‘‘Lick, I may need a MDL (MIT Design Language) was developed, lot of resources for that.’’ He said, ‘‘As mainly by Chris Reeve but with help from much as it takes.’’ other students and staff. We did experiments I spent the next year or so, working with with an Evens and Sutherland display and Bob Kahn, worrying about the ICCC demon- with communication between our PDP-10 stration, down to the smallest detail about mainframe and Imlac displays. how to put a false floor in the hotel in Wash- Lick loved the latest toys, and he didn’t ington, DC, so that we wouldn’t step on like packaged products. Thus, we always wires and disconnect things and personally had lots of hardware and software to be installing a 20-wire modification missing developed and integrated. It was a full-time from one Imlac terminal. MIT sent a delega- jobforseveralofustofigureouthowto tion of 20 or more staff to the demonstration make that all work. For instance, our PDP-6 and had a like number back at the lab to en- running the ITS (Incompatible Timesharing sure the operation of our systems. System) didn’t have a disc operating system, so we bought discs and a smart freshman, Walden: After that you went back Scott Cutler, adapted the Artificial Intelli- to concentrating on the Dynamic gence Labs’ PDP-6 disc operating system to Modeling Group. our discs. Another effort was connecting a Vezza: The group was working on many proj- large memory bank to our PDP-10, done by ects, and Lick was being asked to do a second recent MIT graduate Pits Jarvis. tour at ARPA in 1973.3 I took over the group At one point, Licklider invited me to a lec- and immediately scouted around for a new ture by Robert Kahn in the director’s confer- mission. I heard about a Morse code chal- ence room about the ARPANET. Afterward lenge that the Department of Defense had Lick said, ‘‘You’d better get the staff busy.’’ given to George Heilmeier.4 A couple of my So I found out more about the network and staff members were hams, and I knew Morse that we needed an interface to the ARPANET code from my Navy days. With the group’s interface message processor (IMP, or packet knowledge of the subject, we wrote a proposal switch). MIT graduate Bob Metcalfe designed that was accepted.5 and built it. (He was in graduate school at Harvard but learned of the opportunity Walden: Michael Dertouzos had taken from one of his MIT fraternity brothers.) over Project MAC by this time. I ascertained he was capable, and he said he Vezza: Yes, and I still had a job because my could do it. He had a lot of chutzpah. group was one of the better funded ones in During that early ARPANET period, Steve the lab. Crocker was holding ARPANET Network Working Group meetings trying to develop Walden: How did you come to help the user protocols for remote connection, Dertouzos with Project MAC? Or was it file transfer, and so on. To get everyone on the Laboratory for Computer Science6 the same page, a meeting of the NWG partic- by that time? ipants was hosted at MIT. The purpose was to Vezza: It became LCS shortly after Mike took make the various host-level protocols on over. Mike always had associate directors to every operating system on the ARPANET help with various jobs. Because I was in interoperate with their counterparts on charge of one of the larger groups, he asked every other system. We had a matrix on a me to be one of his associate directors. white board of systems and services. When a service interoperated with its counterpart Walden: Infocom was started while you on another system, it was considered alpha were still at Dynamic Modeling. tested. By the end of the week everyone had Vezza: Infocom was started around 1979. At a an idea of what work was still ahead. meeting in my dining room, 10 of us decided After that meeting we started planning the to initially invest $10,000 each. first ARPANET public demonstration at the We started the company to create a data- October 1972 International Conference on base system for a personnel computer. How- Computer Communication in Washington, ever, Mark Blank, who was the prime mover DC.2 I remember telling Lick, ‘‘You know behind Zork on the PDP-10 while at Dynamic there’s going to be a demonstration in Wash- Modeling, said, ‘‘We’ve got Zork already; why ington at ICCC,’’ and he just looked at me don’t we try to market it for an Atari or the April–June 2012 3 [3B2-9] man2012020008.3d 26/4/012 17:8 Page 4 Interviews Apple-II?’’ Zork was a large program, so for at one such meeting—this was 1994, maybe easy portability among machines, Joel Berez even late 1993—and someone said, ‘‘We and Mark built a Zork virtual machine on should start a Web consortium.’’ The seven which Zork ran.