from
Jargon File (4.4.4, 14 Aug 2003)
timesharing
[now primarily historical] Timesharing is the technique of scheduling
a computer's time so that they are shared across multiple tasks and
multiple users, with each user having the illusion that his or her
computation is going on continuously. John McCarthy, the inventor of
{LISP}, first imagined this technique in the late 1950s. The first
timesharing operating systems, BBN's "Little Hospital" and {CTSS},
were deplayed in 1962-63. The early hacker culture of the 1960s and
1970s grew up around the first generation of relatively cheap
timesharing computers, notably the {DEC} 10, 11, and {VAX} lines. But
these were only cheap in a relative sense; though quite a bit less
powerful than today's personal computers, they had to be shared by
dozens or even hundreds of people each. The early hacker comunities
nucleated around places where it was relatively easy to get access to
a timesharing account.
Nowadays, communications bandwidth is usually the most important
constraint on what you can do with your computer. Not so back then;
timesharing machines were often loaded to capacity, and it was not
uncommon for everyone's work to grind to a halt while the machine
scheduler thrashed, trying to figure out what to do next. Early hacker
slang was replete with terms like cycle crunch and cycle drought for
describing the consequences of too few instructions-per-second spread
among too many users. As GLS has noted, this sort of problem
influenced the tendency of many hackers to work odd schedules.
One reason this is worth noting here is to make the point that the
earliest hacker communities were physical, not distributed via
networks; they consisted of hackers who shared a machine and therefore
had to deal with many of the same problems with respect to it. A
system crash could idle dozens of eager programmers, all sitting in
the same terminal room and with little to do but talk with each other
until normal operation resumed.
Timesharing moved from being the luxury of a few large universities
runing semi-experimental operating systems to being more generally
available about 1975-76. Hackers in search of more cycles and more
control over their programming environment began to migrate off
timesharing machines and onto what are now called workstations around
1983. It took another ten years, the development of powerful 32-bit
personal micros, the {Great Internet Explosion} before the migration
was complete. It is no coincidence that the last stages of this
migration coincided with the development of the first open-source
operating systems.