The Open Road

In what may come to be seen as a deeply symbolic moment in the history of operating systems, Red Hat is on the verge of surpassing Sun Microsystems' market capitalization for the first time.

Sun, perhaps unfairly, represents a fading Unix market. Red Hat, for its part, represents the rising Linux market.

As I write this, Red Hat's market capitalization sits at $2.62 billion, while Sun is just ahead, at $2.7 billion. The stock prices are way out of whack with revenues: Red Hat pulled in $627 million in 2008. Sun? More than $13 billion.

Such is Wall Street's confidence in Red Hat's Linux focus, however, that the market capitalizations between the two companies are almost at parity.

Both companies, of course, have product portfolios beyond Linux or Unix. Sun, in particular, has been significantly expanding its portfolio to include full systems that comprise software (OS, database, storage, portals, etc.), services, and hardware.

Given enough time for its open-source strategy to play out, Sun's market capitalization will likely recover and outpace Red Hat's. But for now, a symbolic moment is about to occur. The inauguration of the Linux-based economy?

Though SCO still has the option to appeal, a federal district court judge Dale Kimball has now effectively written its death sentence in the form of a somewhat blistering final judgment (PDF), as Groklaw reports.

SCO, once the bane of the open-source world, is effectively dead. The company, which long ago stopped trying to make useful products and instead morphed into a boutique law firm, has seen its revenue slide into oblivion while Novell, which stood up to SCO and has now won in court, has seen its Linux revenue jump.

Lesson? You can only milk a weak intellectual property claim for so long. Ars Technica gives the details of the final judgment against SCO:

Judge Kimball determined that SCO was subject to a contract with Novell, which it violated by lifting SVRX confidentiality provisions in a licensing agreement with Sun. This move exceeded the authority granted to SCO under the terms of a 1994 asset purchase agreement that enabled SCO to sell limited SVRX licenses to third parties on behalf of Novell. Judge Kimball also determined that SCO breached its fiduciary duty by neglecting to remit the requisite portion of the licensing revenue to Novell. In addition to the $2,547,817 that SCO was originally ordered to pay to Novell in a previous judgment, SCO will also have to pay $918,122 in prejudgment interest and $489 per day from August 29 until November 20.

It's possible that someone will invest more money in SCO to take a gamble on the lawsuit turning in SCO's favor on appeal, but this is doubtful. SCO's claims have been demonstrated in court to be false, leaving any would-be litigant with no real possibility of winning.

The largely unsung hero in all this? Novell. Novell has stuck with the litigation for five years. Thank you, Novell. I may disagree with the company on other issues, but on SCO and other patent trolls we can agree.

Last week, IBM announced that over the past 10 years, it has gone from also-ran to first place in the Unix server market, claiming 35 percent of the $61 billion market in 2008 to Sun's 29 percent share, as noted in the Post-Bulletin.

Good for IBM, but the real news in this Unix market share carve-up isn't Unix at all: it's Linux.

Linux server growth has been outpacing Unix server growth for some time, with Linux gaining more than a full percentage point of market share to land at 13.4 percent market share earlier this year, according to IDC.

As more and more enterprises switch from Unix to Linux to gain performance and shed costs, Red Hat and Novell stand to gain.

For those wondering how big Red Hat and Novell can become on operating-system revenue alone, keep that $61 billion number in mind. Most of that $61 billion is hardware-related, but it meant approximately $650 million in Linux server sales for Red Hat and Novell over the past year. As Linux eats into Unix, Red Hat and Novell can expect to grow linearly with it.

Or perhaps they can do better. As the operating system comes to include big-ticket items like virtualization, perhaps the roughly 1 percent of the overall server market that Red Hat and Novell claim in software revenue will become 2 percent...or more.

This is the big opportunity. It's a good time to be selling Linux server-related software subscriptions.

This is one of the funniest, best posts I've read in a long time. Randy Jensen wrote it back in 2007, but it just popped up on Digg today. I had never read it. It's hilarious.

Jensen compares different operating systems to types of houses, and comes up with some gems like this:

Windows is the government housing. The houses are built quickly, cheaply and go up anywhere and everywhere. Unfortunately since they were all built so cheaply, you end up spending twice as much later to keep the place standing.

And this:

Linux is the carpenter/designer's dream house. The foundation is rock solid and is inspected every year by thousands of the best carpenters all over the world. The house is everything you want, if you know what you're doing. It's all there for you. The doors, windows, walls and roof. You 'get' to assemble it yourself any way you like....

Everyone on the block is envious and wants to know how you bought such a beautiful house so cheaply. After you explain to them that you built it yourself, they leave in disgust and never talk to you again because you're one of "them".

As for the Mac? "The floors are so clean there is a mirror reflection of every item in the house and the lights are positioned perfectly to make everything look just perfect...just the way the owner wanted it, so don't touch!" Classic.

Paul Murphy at ZDNet has a great post on why many MCSEs (Microsoft Certified Systems Engineer) struggle to gain proficiency with Linux. In part it may stem from apathy, but Paul (following a long post from a would-be Linux adoptee) identifies something more fundamental:

A divide between theory and practice.

As Paul writes:

...[T]here's a great divide between the Windows and Unix camps: a divide one side doesn't recognize and the other doesn't want to cross. It's the divide between training and education: the difference between rote learning and the application of theory to practice.

... Read more

I found this article on O'Reilly's (Microsoft-sponsored) Port 25 page fascinating. For all Microsoft's attempts to own the budding minds of students, it may well be that Microsoft has become too corporate, too sterile to be of interest to the creative mind:

Even back in my day, you could go to a "Windows lab" and work with Visual Studio or go to a "UNIX lab" and use vi and gcc. And you know what? All the fun was in the UNIX lab. And not just for me. There was just a difference in the attitudes and ethic across the two lab environments. People in the Windows lab were trying to get their project in before it was 11:59 PM, while people in the UNIX lab were goofing off, playing with code, and... trying to get their project in before it was 11:59 PM.

What is it about UNIX, vi, emacs, gcc, perl, and INSERT-HERE that makes it fun to play with, while Visual Studio just makes you want to... well, work?

In the enterprise, this alleged Microsoft attribute might be considered a Very Good Thing. But is it? Do enterprises really want automatons that punch in and code to spec? Or do they want innovation that changes the game?

As Computerworld reports, the search for simplicity is driving Linux and Windows into the data center in ever growing numbers. Gartner expects Windows to hit $19.6 billion in 2007, growing to $22.2 billion by 2012, 13% growth. Linux will pull in $8.6 billion in 2007 and boom to $12.2 billion by 2012, 40% growth. Linux, then, is growing faster, but largely because it's starting from a smaller installed base.

The big loser in all this is Unix, which will not lose revenues (holding flat at roughly $16 billion), but won't grow them. Consolidation of operating systems in the data center primarily involves Linux and Windows going forward:

Users will also continue to seek to simplify their environment, often by cutting back on the number of operating systems where possible. In an audience poll, the majority of attendees - 80 percent - were either reducing the number of supported operating systems or maintaining that number, with just one-fifth opting to add to their operating system mix.

The end of Unix? Not today or tomorrow. But Sun, IBM, and HP need to be worried about trends that point to growth outside their Unix platforms.

In some ways, it's just another nail in SCO's coffin, but Steven J. Vaughan-Nichols reports that a German court has found SCO guilty of lying (gasp!) with regard to its claims that Unix code found its way into Linux. But with SCO's directors trying to reap final monetary gains from their mendacious ways, it's probably good to drive as many nails in as possible to ensure SCO truly dies. Quickly.

In the United States, SCO's Linux/Unix litigation has been stalled out while the company's bankruptcy trial is being dealt with. In Germany, however, several court cases have found SCO Group GmbH, SCO's Germany branch, guilty of lying about Linux containing stolen Unix code.

Truth was never a strong point with SCO's management team. At least now we can thank the Germans for officially declaring this fact. Remember, SCO friends: Lügner gehen zur Hölle.