SCO has finally filed for bankruptcy, a little six-plus years after the original SCO company sold their name and their Unixware business to Caldera Systems -- until then a minor Linux vendor whose only business success was in suing Microsoft. For a couple of years up to the sale I had worked for SCO, and had been vocal within the company about how critical it was that adopt Linux and recast the company based on enterprise-class Linux services. I was one of the first to go when the deal came down. That was always an uphill proposition at SCO, a company that was built on the principle of selling operating systems at $1000 a pop. That price point worked as long as one could save more money on commodity Intel hardware compared to the Unix workstations sold by Sun and others, but by 2000 that model was being squeezed on all fronts -- most seriously from Microsoft and Linux, both of which negated SCO's Intel hardware edge. At the time, I argued that SCO could sell to customers too smart for Microsoft, or to customers too dumb for Linux, but not both: the smart people would just move on to Linux, and the dumb people would go to Microsoft. Since SCO couldn't become Microsoft, its only chance was to become Linux. But in a classic case of companies wedded to their margin models, they did neither. They opted to try to protect their declining market by suing everyone who turned against them, including former Caldera sponsor Novell, former SCO partner IBM, and former customers like AutoZone. The only company they didn't sue was Microsoft, who fed them money to try to poison the Linux market.

When SCO sold out to Caldera, they were still doing around $150 million/year in business, down from over $200 million at their peak. Now they're down to around $6 million/quarter, and still dropping. I'd guess that all of those revenues are on legacy systems, and it's only a matter of time before those dry up completely. Those legacy systems were effectively money in the bank, requiring no new development and few employees. Indeed, SCO shed virtually everyone I knew there -- last time I heard there were still 2-3 familiar engineers on staff, but that was quite a while ago. So what led to the bankruptcy, at least in the Chapter 11 sense, wasn't revenue losses; it was legal setbacks. A federal court ruled that Novell, not SCO, owns the Unix trademark and source code -- something I heard back in 2000 when I proposed open sourcing much of that same code. It looks like the IBM case has turned against SCO as well, and maybe there are others.

I ran across a comment on Slashdot that covers the history fairly well, except for this paragraph:

Caldera didn't want the UNIX business either. They were a Linux business and thought they could convert SCO's UNIX distribution network to selling Linux instead. That didn't work out either; apparently the UNIX resellers didn't want to switch to Linux and Caldera was making more selling UNIX than distributing Linux. So they ditched Linux (and their CEO) and switched to concentrating on UNIX and changed their name to SCO for the name recognition.

Actually, Caldera always wanted to be SCO. They initially hoped they could ride Linux into that position, but they always wanted SCO-like VAR channels, and they always wanted to lock them in with proprietary bits of software. They failed repeatedly, then jumped at the chance to buy SCO. The first thing they did at SCO was to kill SCO's Linux project, which could have added a lot to Caldera's Linux product. They didn't want to convert SCO's customers to Linux because what they really wanted was SCO's proprietary margins. The real mystery is why they thought they could do a better job with a product line that was already failing.

Also, the real ogre here was Doug Williams. He had already dumped a fair chunk of his stock a year before, when SCO was peaking during the Y2K binge, just before the bottom fell out. He sold the company to Caldera not because they would help SCO survive but because they had more cash to burn than any other suitor -- cash that they hadn't earned, that fell into their lap from Microsoft to clear up Digital Research's antitrust suit -- and because they were dumb enough to believe they could make SCO's business model work. It's also quite likely that Caldera had only the vaguest idea what they were buying, although it's quite an irony, given their incestuous history with Novell, that they missed the fine print there.

I always feel sad about this because I had a great time working at SCO. I worked with a lot of great people there, including many who had followed UNIX out of Bell Labs through USL and Novell. It's too bad it's come to this. One thing I believe strongly is that employees should have a substantial stake in their companies, and consequently that companies have an obligation to their employees, as a group if not necessarily individually. SCO used to be pretty good to its employees, but in the end it was Doug Williams who called the shots, and it was Doug Williams who cashed out and sent the rest of the company off to oblivion. I knew him well enough to know that he knew what he was doing.