June 29, 1999
Web posted at: 11:32 a.m. EDT (1532 GMT)

by Mary Brandel

(IDG) -- It's the late 1950s, and you're a computer operator at MIT running a long job on a computer donated by IBM. The phone rings. It's a request from the president of IBM -- who races big yachts on Long Island Sound -- to run the program that assigns handicap points to the boats.

That request means you have to abort the job that's running, mount a new tape and then restart the current job from scratch.

Or maybe you're a computer programmer. It's late afternoon, and you're picking up the results of the computer program you dropped off this morning. The printout contains an error message. After waiting all day for your program's results, you discover you have left out a comma.

It was in that type of environment that John McCarthy, a professor at MIT, submitted a memo to Philip Morse, then the director of the MIT Computation Center, outlining a new concept called "time-sharing." Unlike batch processing, where programmers submitted programs on punch cards to a computer center, this new mode of computing promised to make computers more accessible to, and interactive with, users.

Encouraged by Morse to pursue the idea, an associate professor named Fernando Corbato and his team developed the Compatible Time-Sharing System (CTSS) in 1961, which was the first demonstration of how time-sharing could be done.

Simply put, time-sharing enables a computer to serve many users simultaneously, so that each person feels like he's using his own private computer.

Functionally speaking, a time-sharing computer stops a long job, copies its memory to a place on a disk, runs something else and then starts up the long job again -- without interrupting the user. "Each user's program has access to the full resources of the machine, and several programs can share time on the machine," explains Tom Van Vleck, who was a junior member of the CTSS team and a developer of Multics, a successor to CTSS.

In addition, computing is interactive. Programmers create programs on the keyboard, and the computer responds almost immediately, with results or error messages.

Although that sounds quite rational to the 1990s user, not even the hardware was prepared for such a radical concept in 1961. For instance, the IBM 7090 used by Corbato's team lacked a keyboard. "Fortunately, we were able to get Teletype machines from AT&T Corp. and Selectric mechanisms [IBM Selectric electronic typewriters] from IBM to solve that problem," Corbato says. "But even then, we had to fight for both upper and lowercase letters."

Working with IBM, the team was able to modify a second machine, an IBM 7094, to solve that and many other problems. "Many of these solutions are still in use today," Corbato notes. They included a hardware timer to interrupt user programs and a way to prevent programs from operating outside of designated memory boundaries.

The 1961 demo was "crude and incomplete," according to Corbato. However, it showed time-sharing to be a feasible concept, and many time-sharing systems would follow the CTSS demo. The finished version supplied time-sharing services to MIT, New England colleges and the Woods Hole Oceanographic Institution.

But most important, the CTSS demo led to Project MAC. Funded with $3 million in 1962 by the Advanced Research Projects Agency (ARPA), Project MAC's first goal was to develop a full-fledged time-sharing system, named the Multiplexed Information and Computer System, or Multics.

Multics wasn't ready until 1969, and it never became a commercial success. However, It is important in many ways. For one thing, Ken Thompson and Dennis Ritchie, the inventors of Unix, used many ideas from Multics.

But perhaps most important, time-sharing brought users into the universe of computer design.

"Time-sharing introduced the engineering constraint that the interactive needs of users [were] just as important as the efficiency of the equipment," Corbato says.

Brandel is a frequent contributor to Computerworld. Contact her at brandel@cwix.com.