Multics, Unix, and How Microsoft Revolutionized the Mainframe World of 1980 with Xenix

The series setting the record straight on computer history continues. Only this time I won't be presenting a graphical interface or a particular computer, but the history of Unix and how Microsoft managed to revolutionize that market.

In 1969, Ken Thompson wrote the first Unix in assembler on a DEC PDP-7. A successor to Multics, a (relatively) failed project of MIT, General Electric and AT&T. Thompson's goal was to do a bit of word processing and play around. Three years later, work began on rewriting Unix from assembler to C. The first versions were given away to universities for free – from these the BSD line later evolved, on which Mac OS X is built today. AT&T's Bell Laboratories, the former developers of Multics, took notice of Unix and took on Thompson and his work. In the late seventies, Unix was the operating system for mainframes, full stop. The then still young company Microsoft, which made most of its money with the DOS operating system traffic light programming and word processing on the Macintosh, began work on its own version of Unix, based on Bell Laboratories' Version 7. On August 25, 1980, Microsoft released Xenix. According to Microsoft's plans it was to become the standard operating system for PCs. The plan didn't work out, though, because at the time a hard disk cost around 10,000 marks, and a hard disk was a prerequisite for Xenix. On top of that, Xenix required 256 KB of RAM, while IBM PCs shipped with only 32 or 64 KB. So Xenix remained reserved for mainframes and DOS took over the personal computer. As early as 1983, Microsoft spun off its Xenix division and from then on called it SCO. SCO Unix became the most successful Unix. Siemens licensed a version from SCO for its own operating system, Sinix. Xenix was deployed thousands of times over in banks and industry as a server and timesharing system. In the recent past, SCO, which still owns the rights to Xenix and some other parts of Unix, caused a stir when it accused the Linux community of having taken parts of the code in the Linux kernel from an SCO Unix version.