Fifty years ago, a major project that ultimately seeded much of today’s computer technology was created at MIT: Project MAC, and the Multics operating system initiative within the project. Out of this came hundreds of computer scientists and programmers, who went on to found and work at leading computer companies and create milestone products whose legacies live on today.
In honor of Project MAC’s and Multics’ fiftieth anniversary, MIT is hosting a two-day conference, “MAC50: Computing The Future — Celebrating 50 Years of Computer Science and Artificial Intelligence, ” on May 28 and 29, offering a mix of retrospectives, what’s currently being done, and where computer science visionaries see we’re headed.
Speakers at the event include Ethernet inventor Bob Metcalfe and leaders from companies including Akamai, Boston Dynamics, and iRobot. While no more seats are available for the live event, MIT will be recording and archiving the sessions (check the MIT site later for availability).
Here’s a short look at Project MAC, and Multics — not just way back when, but also how it has influenced and led to the mobile, desktop and Web tools we now use daily — including some recollections and other thoughts from some of the project’s alumni, on what they did back then and there, and since.
PROJECT MAC: FUELING COMPUTER SCIENCE
At its peak, Project MAC (which, by different accounts, stands for “Mathematics And Computation,” “Multiple Access Computer,” or “Machine-Aided Cognition”) had around 400 researchers.
Over time, Project MAC turned into MIT’s Laboratory for Computer Science (LCS) and famed AI (Artificial Intelligence) Lab, and a decade ago, LCS and the AI Lab came back together as MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL).
Project MAC’s impact went beyond its specific research activities, according to Paul Green, a Multics alum who is currently a senior technical consultant at Stratus Technologies Inc. and a co-organizer of the Multics reunion. “Project MAC led to the start of an official computer science curriculum at MIT. When I joined Project MAC as an MIT freshman in the fall of 1969, computer science was part of the Electrical Engineering department. Now, MIT’s ‘Course VI’ is the Department of Electrical Engineering & Computer Science.”
Ten years ago, MIT combined the Laboratory for Computer Science and the AI Lab into MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). The many researchers who worked at Project MAC, LCS, the AI Labs and CSAIL over the years includes a sizeable piece of computing’s roll call of honor.
Project MAC’s AI Group included director Marvin Minsky and LISP inventor JohnMcCarthy. Multics alums include Ken Thompson and Dennis Ritchie, who originated Unix, and Brian Kernighan, a leading early contributor to Unix; Dan Bricklin and Bob Frankston, co-creaters of VisiCalc, the first electronic spreadsheet; and Bill Poduska, founder of Prime Computer and Apollo Computer. (See the Multicians web site for more.)
Companies founded or strongly driven by Project MAC alums include Digital Equipment Corporation, Prime Computer, Stratus Technologies, Bolt Beranek & Newman, and RSA Data security. According to MIT, CSAIL’s researchers have created over 100 companies , from 3Com, Akamai, and iRobot to ITA Software, Lotus, RSA Data Security, and Thinking Machines Inc.
MULTICS, HELPING MAKE COMPUTER RESOURCE SHARING EASIER
“Multics promoted the notion of time-sharing access to an expensive resource — here, computers that cost millions of dollars in the 1960s and 1970s — similar to the way in which today’s researchers share access to tools like the Hubble Space Telescope up in Earth orbit, and CERN’s Large Hadron Collider in Switzerland,” said Paul Green.
“The basic idea behind CTSS — one of the first time-sharing operating systems, developed and in use at MIT — and Multics came from MIT Professor Emeritus Fernando J. Corbató,” said Tom Van Vleck, who worked on the Multics operating system for about sixteen years in the ’60s and ’70s.
“In 1961, when Fernando J. Corbató was a Research Associate in the MIT Computation Center, he wanted to find a way to share the unused time on the mainframe computer,” Green said. “He later told us that the idea for time-sharing came to him because the only time he could get access to the mainframe was late at night, often after midnight, and he wanted a way to get his computer jobs run during the day, using those slices of idle machine time basically going to waste.”
Exploiting CTSS timesharing and remote access was a major focus of MAC’s, says Van Vleck. “Corbató provided the vision, leadership and continued engagement — he led Project MAC’s computer systems group, saw that CTSS was limited, and that a better system was needed. Under Corbató’s leadership, that system became Multics.”
While there are no longer any systems running Multics — the last one, at the Canadian National Defence site in Halifax, Nova Scotia, was shut down in 2000 — Multics’ legacy is present in today’s computers and the Internet.
“We see the use of shared resources today even on our desktop computers, tablets and smartphones,” said Paul Green. “Here, not sharing by people but by several applications running on a device at the same time. Multics’ time-sharing concept is how huge numbers of people can be accessing the same web site and even web page, sending electronic mail, or watching a video on YouTube. Multics concepts and programming tools fed into what became Unix and Windows.”
MIT Professor Emeritus Fernando J. Corbató remarks, “On reflecting on the CTSS and Multics efforts, I think that one of better outcomes is that there were hundreds of young programmers exposed to an environment of building complicated and advanced computer systems. It is gratifying that so many of the participants have gone on to have successful careers and make further contributions.”
MANY PEOPLE, MANY PIECES AND PROJECTS
The work done at Project MAC and with the Multics timesharing system — like that done being done by computer scientists at other universities and companies — comprised lots of computer tools, software utilities, and other innovations.
“My BSEE thesis work at MIT in 1969 was to build part of a compiler for Multics,” recalled Jack Haverty, who went on to work at Bolt, Beranek & Newman Inc. (BBN) and at Oracle. “My MS/BS thesis adviser in 1970/71 was Professor Licklider, and after I got my degree, I stayed on as Staff in his group for about seven years.”
“I worked on Multics from 1969 to 1986 (Project MAC 1969-1972 and Honeywell 1972-1986),” said Melanie Weaver. “I worked mostly on various aspect of Multics runtime, including the default condition (a.k.a. exception) handler, and maintaining/extending the linker, binder and BASIC compiler after their original developers left. I also supported the efforts of some non-PL/1 compiler developers to integrate into the unique Multics runtime and developed some lighter-weight runtime environments. I am currently working with utility data using SQL and Microsoft .NET.”
Ken Pogran worked at Project MAC and then at MIT LCS, from 1970 to 1972 as a graduate student, and as a research staffer from 1972 to 1980. “I was involved in developing software to connect Multics to the ARPANET,” said Pogran. “I wrote the first FTP server and a lot of email-related stuff for Multics; was one of the co-authors of RFC-733, which was the first semi-official written standard for email, and also worked on the then-new area on LAN (Local-Area Network) technology.” Pogran went on to work at BBN, among other places; currently, he is Director of Business Development at EGH , a software company in Lexington, Mass.
Bob Metcalfe was an undergraduate at MIT from 1964 to 1969, and a Harvard Ph.D. student 1969 to 1973.
While at Project MAC, Metcalfe also, among other things, built Arpanet IMP interface hardware to put Project MAC’s DMS on the Arpanet, and, recalls Metcalfe, “I worked with Danny Cohen at Harvard to take an aircraft carrier landing simulator at Harvard, send its images over the Arpanet to MIT for processing, and return them to Harvard for display,” Metcalfe said.
Metcalfe is best known to many as the “Inventor of Ethernet,” the network technology today used in all office and home wired and wireless networks.
“My talk at MAC50 will be about how Ethernet WAS invented at Project MAC, but only in the larger sense,” Metcalfe said.
“My Harvard Ph.D. thesis was on ‘Packet communication, and included thin-wire best-efforts interprocess communication,” he said. “From Project MAC, I went to the Xerox Palo Alto Research Center (XEROX PARC), where I put Xerox on the Arpanet and, in 1973, invented Ethernet for the Parc Alto personal computer. And while consulting at Project MAC in 1979, I founded 3Com, another networking company — so Project MAC helped me ‘invent’ Ethernet twice, in some sense. But of course nothing is invented in one day, and I got my Internet start at Project MAC.”
“The people at Project MAC were university researchers trying to solve interesting problems — and educators seeking to train the next generation of engineers and computer scientists,” Green said. “They invented the field. And Multics’ efforts to create a utility-grade infrastructure for computing is a lot like what we today are calling ‘the cloud.’”
So don’t dismiss those thunking refrigerator-sized objects bedecked with blinking lights and spinning tape drives, now seen only in museums and movie and TV scenes, and the people who created and worked on them. MIT’s MAC50 rightly celebrates the role these computers and computer scientists played in the remarkable tech toys we have and use today.