Compatible Time-Sharing System

~ the obit ~

The compatible part of the name was a concession. Fernando Corbató's group at the MIT Computation Center wanted several people typing at the same machine at the same time, but the machine was an IBM 709 that a lot of researchers depended on for ordinary batch work, and nobody was going to let them break it. So CTSS ran the normal batch monitor underneath as just another job. Your FORTRAN deck kept working exactly as before while, in the gaps, a handful of typewriters got to talk to the computer directly. That is where the compatibility was: with the old way of doing things, which is usually how a radical idea gets permission to exist.

It was first demonstrated in november 1961. The 709 was swapped for a modified 7090 in 1962 and then a 7094, and by 1963 MIT was running two of them. Users called them by the color of their cabinets. The blue machine served the Computation Center and went back to IBM in late 1968. The red machine belonged to Project MAC, arrived in october 1963, and kept running for another decade after its twin was gone.

What CTSS actually changed is hard to feel now because the alternative is so far away. Before it, you wrote your program, punched it onto cards, handed the deck to an operator, and came back the next day to find out you had a typo. Debugging happened at the speed of a queue. After it, you sat at a terminal and the computer answered you. Every habit of working with a machine that anyone now has, including the expectation that a mistake costs seconds rather than a day, starts on hardware that was already obsolete when Nixon resigned.

The red machine was switched off eleven years before I was born. My whole relationship to this one is inherited: I learned the shell before I learned where the word came from, and the file I edit to configure it ends in the letters rc because of a command written for a computer I will never see. There is no personal memory available here, only the equipment.

~ the rap sheet ~

~ what it did first ~

CTSS was the first computer system to ask for a password. It had to be, because it was the first one where two strangers were logged in simultaneously and the question of whose files were whose suddenly had teeth. Around 1965 it also got a command called MAIL, which let one user drop a message into another user's directory to be read whenever they next logged in. Electronic mail was not designed as a communications medium. It was a way to leave a note for somebody sharing your computer.

Louis Pouzin, visiting from France, wrote a command called RUNCOM that would execute a list of other commands stored in a file. He also coined the term "shell" for the program that reads what you type and does something about it. Both survived. TYPSET and RUNOFF, the editor and the text formatter, became roff, then nroff and troff, which is what Unix manual pages are still written in. A remarkable share of the furniture of modern computing was built by about a dozen people around one room in Cambridge, Massachusetts.

~ the funny part ~

In 1966 an administrator managed to swap two files: the message of the day, which every user saw when logging in, and the password file, which was stored in plain text because nobody had invented the alternative yet. Everyone who logged in that evening was greeted by the complete list of everybody's passwords. It was a friday at about five in the afternoon.

A Project TIP staffer named Bill Mathews stopped it by deliberately crashing the machine with a single instruction that sent the processor into a loop it could not escape. Taking the whole system down was the fastest available fix. Staff then spent an unplanned evening changing passwords one at a time. The first password system in history and the first password leak in history are separated by five years and one filename mixup.

~ the long tail ~

CTSS was never meant to last. It was a demonstration that the concept worked, and the moment it worked, the same people started designing the thing that would replace it. Multics was already underway by 1965, sharing staff and half its ideas, and it in turn drove the Bell Labs programmers who left to write Unix. MIT's other lab looked at CTSS, disliked its restrictions, and wrote the Incompatible Timesharing System as a joke at its expense. Whole families of operating systems exist as arguments with this one, either extending it or rejecting it.

The red machine ran on at Project MAC until 20 july 1973, kept alive by people who still had work on it after everyone else had moved to newer systems. Roger Roach pulled the switch. No press covered it. A 1961 experiment in letting more than one person use a computer had by then quietly rewritten what a computer was for, and the machine that proved it went out the way most of them do, with one staffer in a room and nobody watching.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »