Comment Articles

Unix and Linux: Lessons learned, lessons passed on

Steven J. Vaughan-Nichols AnchorDesk

Published: 24 Jul 2000 16:00 BST

Just to shock the pants off those who don't know their history, let me quote from a 1980 issue of Microsoft Quarterly, "The XENIX's system's inherent flexibility ... will make the XENIX OS the standard operating system for the computers of the '80s." Yup, that was Microsoft beating the drum for Unix. Soon, thereafter it made the sweetheart deal of all time and got the Quick and Dirty Operating System (QDOS) on the cheap. Four years later, Microsoft abandoned the Intel Unix market by selling Xenix to SCO because QDOS had become MS-DOS and was much cheaper. Every copy of Xenix Microsoft sold, on the other hand, cost them an AT&T licensing fee.

If I were crass, I might say that given a choice between quick, dirty, cheap and profitable versus effective, flexible and a bit more money, Microsoft made exactly the choice you'd expect it to make. On the other hand, Bill Gates is a billionaire, so operating-system ideological purity certainly doesn't count for everything.

The SCO Story

While it hasn't made tens of billions, SCO has made hundreds of millions over the years. It hasn't been easy. In the late '80s and early '90s, SCO had to battle with several other companies that were bringing Unix to the Intel platforms. It was in 1990, that I went from just using SCO Xenix and Unix and writing about them, so I had a front row seat.

SCO, thanks to the boost already given to Xenix by Microsoft, was the early leader and never let go of its lead despite the best efforts of companies like Interactive Unix. It took five years, but by 1995, with critical battles over how to bring System V Release 4 Unix to the PC done, SCO was the clear winner.

But -- and here's two lessons for today -- one reason why SCO took so long to win in the PC Unix wars was that Unix applications were incompatible from one company's Unix to another's. Because of these resulting increased application-development costs, the Unix companies -- despite endless attempts to work together (such as the The Open Group) -- made little practical improvements in making a truly universal Unix.

Those problems persist to this day. The bottom line was that instead of unifying against the growth of Windows, the little Unix companies wasted their time and efforts fighting and competing with each other. SCO was the best at this game and was the survivor. But, being the biggest fish in the Intel Unix pond didn't mean much in the wider ocean as more companies turned to Intel servers powered by NT instead of Unix.

Read on...