It was back in 2002, according to Sam Trenholme, the creator of the secure DNS server software MaraDNS. That was the year that forces conspired to make sure Linux on the desktop would never become a reality. Linux as a server was another matter entirely, but to him the "Linux desktop" is as dead as the Amiga.

How did this happen? Sam cites three things.

• Mac OS X was out and getting application support; UNIX was finally on the desktop (and it wasn't Linux).
• Windows XP came out: Microsoft finally had an OS for end-user desktops with server-level stability. [I question whether XP's stability is "server-level"; that to me would imply being able to hot-add memory or do other things that are way out of XP's league. "Suitably stable" is more like it.]
• Loki games, a company that made games for Linux, went out of business. This was the final nail in the coffin for commercial desktop applications for Linux.

In his view, this created a death trap. The end user had at least two major, robust and well-supported desktop environments to choose from, each with their own thriving software marketplaces. (You could argue that multiple versions of Windows also constitute "choice", for those of you who abandoned Vista for XP.) And without games -- one of the first and most lucrative software markets for any computing platform these days -- Linux was doomed to not attract the kind of developers it needed to keep alive casual interest in supporting the platform.

I know there are people reading this and muttering, "You're kidding! I'm using Linux as my desktop right now and I love it!" Fair enough; I myself have a machine dedicated to being a Linux desktop, which I use regularly. Much of the software I use day-to-day is open source. The server I host my personal site on runs, I believe, CentOS. I could go on.

What I'm saying is that you can expect to not have a great deal of company except from the other few and proud Linux users out there. If other folks who are not computer-centric start to use Linux, they will be well-protected from that fact. You can expect Linux to roll merrily along, but don't expect it to become as broadly recognized or as automatic an option for the desktop as Win/Mac.

I made a comment not long ago about this dilemma as it applied to a friend who just bought a new Vista-equipped notebook. She's already "arrived", as it were. Rhetoric about "freedom" means nothing to a person who's already quite happy. A desktop version of Linux would have to not just be a Windows-killer, but a Windows-and-Mac-combined-ecosystem-killer. And there are takers, Canonical/Ubuntu being the biggest and most obvious. I give it 18 months before we have a good idea if they're bringing something worthy to life or just continuing the death march.

InformationWeek Analytics has published an independent analysis of the current state of open source adoption. Download the report here (registration required).

Follow me and the rest of InformationWeek on Twitter.