Reading through the “article” that Hubert Figuiere mentioned on Planet GNOME (and the accompanying reply), I couldn’t help but think “What planet does the author of that article live on?” I know RedHat decently enough to do maintenance tasks, and Debian well enough to do most anything I’d ever want to, and prefer Debian, but so what? Does that mean that I am suddenly transformed from a vaguely competent programmer and administrator into a clueless n00b when faced with some other distro? Most of the learning about GNU/Linux is learning about the GNU side, very little of which changes from distro-to-distro. (Apache doesn’t suddenly store it’s config in XML if you’re running on OpenBSD.) /etc/init.d, /etc/passwd, /etc/fstab, and /etc/exports are work-alike no matter what distro you’re on (init.d may include some distro-niceties, like gensplash or RedHat’s functions, but if you stick a script that takes a “start” and “stop” arguments in there and symlink it, it will still run).
But on a deeper level, systems administration means knowing what you know, and knowing what you don’t, and knowing how to take what you don’t know and make it something you do (most of that is just knowing—in the Matrix, “You’re faster than this” sense—you can figure it out eventually). If your administrator is unable to figure out TurboLinux coming from a RedHat background, you’ve got a much, much bigger problem than just having a strange new distro to deal with. And those kinds of problems cannot be solved by dumping a metric ton of cash on Sun Microsystems (which, of course, your RedHat-only admin will know even less about) for their proprietary, “If we don’t support it you’re up shit creek,” software.
It’s the same “Our marketroids are far stodgier than you, which somehow makes the pay-per-view-style, crapshoot, ‘Your call is important to us, but your time isn’t,’ 3-hour tech-support wastage superior to being able to personally e-mail the programmer who actually wrote the code, or fix it yourself,” nonsense underneath all the verbage about support, TCO, and “upgradability.”
Really, though, the whole article feels like a rant from someone who shelled out a lot of money for Sun’s certification classes, and feel as though a goodly share of that will be useless in the rapidly expanding free-software-using/developing world; crass, understandable, and not at all the fault of the free-software-using/developing world.
Update: I should probably note that in spite of the fact that I’ve even though I’ve mentioned Sun negatively in the last couple posts, I’m not singling out Sun for bashing—the second computer I ever touched (at age 12) ran SunOS 4.1.3, and I wasted many an hour playing xbattle against my brother from a networked X terminal. It just happens that the last proprietary software problem I ran into was from Sun, and the article in question for this post covers the broader package of what I had to wrestle with.