Operating Systems and UNIX - History of UNIX
(Page 2 of 4 )
It all began in the year 1965. Early on it was very difficult to get computers to communicate with each other. Often an interpreter was required, even for an application to communicate with another app on the same computer. Operating systems were constructed around a single machine and a new, more powerful computer required a new operating system.
A small group of people assembled in the year 1965 to put an end to this. The goal was to create an OS able to support multiple users. The result of their work is known today as the Multix (Multiplexed Information and Computing Service) mainframe timesharing system.
Bell Labs financed the project; however, the system failed to materialize as predicted, and frustration was replaced with failure and a cut of support from Bell Labs in 1969. A small group of the team working on Multix did not give up, and pursued their dream of creating this OS.
In the spring of 1969 Ken Thompson, Dennis Ritchie and Rudd Canaday meet for an informal talk about what an OS should do from the point of view of the researchers. Once they finished, Canaday picked up a phone and dictated its notes to Bell Labs' dictation service. What they received back the next day in a printed form was a bunch of "butchered" acronyms like inode, eyen and so on. In the end this sketch became the basics of the system that today we know as UNIX.
In the year 1969, a computer game named "Space Travel" was quite popular. This was a hard game to play, as it accepted only text commands. It was based on the Gecos computer. Ritchie got his hands on a PDP-7. This was a medium-sized computer for that time, as you can see in the picture below:
He managed to port the game to this computer, and by doing this he also handled issues related to the operating system. Later that summer he extended all that with the paper file system, worked out important traits like the notion of process, added a couple of user-level utilities like copy, print, delete and editing of the files. He wrote a command interpreter -- what you may know today as a shell.
The name UNIX came into the picture only later on, in 1970, when Brian Kernighan suggested it as a traitorous mean reference to Multix. Encouraged by the improvements, Bell Labs bought a new PDP-11 in December 1970. Now the operating system had to be ported from the PDP-7 to the PDP-11. First UNIX was written in Assembly.
Nevertheless, Ritchie wanted to do it in a higher-level language. In 1971 he initially tried FORTRAN; however, he gave up after just a day. He decided to write a simple language and called it B. The new language had multiple issues. First, it was interpreted. Because of this, it would always suffer performance penalties. The second and more pressing problem was that it was word-oriented and was not good for the new byte-oriented computers like the PDP-11.
The new language received an addition in the form of types. Therefore, the name changed to "New B" (NB). Writing a compiler to it turned out to be a harder task, so this delayed the project all the way into 1973. However, Ritchie managed to pull this off, and the writing of UNIX in the new language C could be started.
Ritchie and Thompson at a PDP-11
Next: Properties of UNIX >>
More Miscellaneous Articles
More By Gabor Bernat
/ 6 
