Obligatory disclaimer: I worked for SCO from 1991 to 1995. I was in the techpubs team working on SCO OpenServer 5, released in mid-1995. So I think I have some insight into SCO's corporate culture as it then was ...

Kids, the company that filed this lawsuit is not SCO.

Engineers at SCO were bolting together PC based UNIXes back before Linus got started. From the late 1980's, they inherited Xenix -- a descendant of AT&T System 7. In 1988 they bought the rights to AT&T SVR3.2 (for an eye-watering sum -- in 1994, each box SCO sold was encumbered with about $200 in royalty payments to other companies). By 1992, SCO UNIX 3.2.4 was a cash cow, and they needed something new.

I'd rate the rot as having set in by late 1991. Before then, SCO was an exceedingly cool place to work -- one of the early UNIX start-ups, SCO was the outfit with the hot tub in the courtyard of the original company offices and the source of numerous interesting legends. There was a lot of cross-fertilization with SGI and Sun at the engineering level back in the late 80's, and some of that survived into the 1990's. But the ACE Initiative killed it dead -- led to a 15% downsizing in 1991, when SCO was forced to admit that it couldn't market Open Desktop against Windows and hope to win. Then there was a string of bad decisions that effectively doomed the company to ossification and slow decline.

First there was the decision to build OpenServer in the first place. Then when SVR4 appeared to be making ground and SVR4.2 (UnixWare 1) came out, there was IIRC a quiet attempt to clone the SVR4 kernel. (The AT&T copyright declarations were retained in the headers, but by 1995 SCO's main product bore about the same relationship to SVR3.2 that a heavily customized rice burner bears to a showroom model.)

But SCO was, at this point, still a real software company. The UNIX dev team had more than 200 engineers working in it. Then the rot set in for real ...

(Historical aside: I first met Linux in 1993, as the system a bunch of SCO's engineers were running on their home machines. But when I left in early 1995, there was an attitude of complete denial in SCO's management -- Linux was a toy system that could never be relevant.)

Anyway. Why did I leave?

The main warning to me that the company was probably not a good long term career bet happened three months ahead of the functional freeze on OpenServer. One lunchtime managers came around our cubicle farm and pitch-forked us into coaches, drove us for two hours around the M25 motorway, and dragged us into a hotel at Heathrow where we were given glasses of grape juice and ushered into a theatre. The lights dimmed, the sound system came up playing "Things Can Only Get Better" (gack!) and the board of directors ran on stage punching their air. The occasion? It was to announce the retirement of the CEO and his replacement by the CFO (yes, the head bean counter). Said CFO promised to grow SCO's revenue base from $200M/year (in 1995) to $1Bn/year by 2000. I took one look at this stage, considered the Linux box (1.2 kernel) back home, and went back to my cubicle and started updating my resume.

There's a point to this long, discursive ramble. After I left SCO, I kept an eye on it. Sales didn't do well, although the Tarantella middleware product -- Doug Michel's pet, after the board panicked, kicked out the accountant, and invited him back -- did okay. The UNIX dev team languished, became an appendix to HP and IBM with Monterey, and in the end was downsized repeatedly until it no longer existed. Finally, SCO split in two.

The important corporate bit now follows. SCO had two arms; the Tarantella middleware arm, which was doing okay, and the UNIX arm, which was in a death-spiral. As I understand it, SCO Inc sold the UNIX arm to Caldera (then flush with IPO dollars), renamed itself to Tarantella Inc, and is presumably doing okay, albeit as a smaller software company in a different field. Caldera retained some of SCO's UNIX marketing and sales staff, but basically treated SCO's software as a cash cow. Caldera were set up with lots of money by Ray Noorda, but don't seem to have had a clue how to sell software. And the company now known as SCO is actually Caldera.

So what's going on?

Caldera has always been a money hole. Caldera peaked at something like 4% of the market for shrinkwrapped Linux distros, and never quite seemed to get the engineering side together. While Redhat's contribution is well-known, and SuSE have done a lot of solid engineering work (much of which is GPL'd -- the device drivers, for example -- rather than the much-more-visible and proprietary admin GUI), who remembers what Caldera tried to add to the community? (Yes, they tried to build yet another admin front end -- in the end, nobody else bothered using it.) Caldera basically targeted the commercial market before it was ready to buy Linux. Then they bought a sadly run-down product from SCO, and failed to promote it effectively. That's because Caldera still think they're selling software licences, rather than support and services. Now nobody wants to buy their licences (when they can pick up something equivalent for free) they're trying to attach a cost to the free stuff.

It's somewhat sad that SCO's name is being dragged through the mud this way; it feels like someone I knew who's been dead for years just clawed their way out of the graveyard mud and began shambling around town looking for brains to chow down on.