April 17, 2014

.comment: Not Forking But Branching - page 3

The Light at the End of the Tunnel Might Not Be a Train

  • February 21, 2001
  • By Dennis E. Powell

Whether the bout between Linux Lewis and Microsoft Tyson leaves Linux standing in the marketplace, ears intact, is in some respects only of interest to commercial Linux distributors. The very skilled could start from scratch, download everything necessary, and put together their own Linux system, and there are more than a few people who have done just that (while resisting the temptation to burn the results on a CD and pronounce themselves distributors). These systems represent the ultimate in specialization, of course -- you're going to tailor your own system to your own needs.

Just one level down are those who began with a distribution but who have so changed and upgraded it that it now has very little of the originally distributed code. A lot of people do this to one degree or another, from upgrading the kernel (and the things in the Changes file that a kernel upgrade dictates) to installing a new XFree86, to compiling all sorts of new versions of everything. Given sufficient skill -- not difficult to acquire -- these users have little reason to install a new distribution at all, ever. This is one of the delightful advantages that Linux has over the popular commercial operating system. It is for this reason that I, a Caldera user, am untroubled by the company's recent not-quite announcement. Orphaning Linux is pretty difficult, and that will continue to be the case as long as we have good compilers and access to source code.

These things are comforting, but it would be hard to make the case that the community doesn't benefit greatly from the commercial distributions. A lot of developers of the software that we use every day are able to do their work only because they are subsidized by a company. This group ranges right up to Linus himself. And setting aside the generally mindless distribution cheerleading and the overly broad default installations, distributions provide the best and most convenient starting point for users to build the systems best suited to them. As noted above, I think this could be improved upon to the benefit of all, but even as it stands, distributions are what moved Linux from the hobbyist category to the serious contender on many fronts that it is today.

So please don't misunderstand. I'm not discussing the survival of Linux but the survival of distributions. Linux will always be around. With a rich array of distributions to meet differing needs, it would also fulfill its promise as a way for Linux companies to make money and benefit the community overall. It would also make development for Linux a more rewarding proposition.

Still, this assumes more sense than some of the distributors have employed so far. The head of SuSE has been quoted as saying that it is not possible to make money with Linux (he later issued a non-denial denial that was more on the lines of "boy, I wish I hadn't said that"). The thing is, he's absolutely right if the model being followed is a huge number of distributions that do essentially the same thing, all chasing the same amorphous market. There are many of us who believe that Linux is a great desktop operating system and can become even greater. But the unhappy fact is that the vast majority of computer users probably don't care what operating system they have (and I'd be willing to bet that an impressive number don't know what operating system they have). Those of us for whom the computer is as much an end unto itself as the means for doing things are and shall probably forever remain a diminishing minority. That's why IBM's involvement in Linux, despite a few bucks tossed to various desktop projects, has nothing to do with the desktop. IBM has learned from bitter experience that it doesn't want anything to do with bottom-level end users. Wanna talk Linux on the mainframe? IBM is the place to go. Wanna set up a Linux box? IBM will sell you a machine, even one with a distribution installed, but after that you're pretty much on your own. IBM is specializing. A non-reactive approach to the Linux marketplace -- how refreshing!

Distributions, many of which are wounded, are retrenching, and some of them will disappear. Nothing will change this as long as the business model extends no farther than the single word, Linux, as it apparently has been by SuSE and others. That business plan worked for a few minutes early last year, just long enough for some fortunes to disappear. (One can imagine people standing at major intersections in Orem, Utah, holding signs that say, "Will trade Caldera stock options for food.") The conclusion being reached, by SuSE and others, is that Linux is not good business. The conclusion that should be reached by Linux distributors is "we screwed up." They have all been clones of each other in every important way. One-size-fits-all fits no one very well. Neither does a vague pursuit of Microsoft -- you've got to have something to sell beyond stability.

The solution? Pretty easy, really. First, sit down and hammer out the Linux Standard Base, then stick to it. At the same time, agree to cooperate. This doesn't mean eliminate competition but instead declare specializations. Let some distributions -- SuSE and Mandrake, maybe -- say that they're going after the desktop. Let some others, as Caldera either has or hasn't, depending on which day you ask, decide to go after the enterprise. Stop trying to do everything, because it doesn't work. The industry has outgrown that model and the distros are being left behind. Put development money in real innovation.

Will this happen? I have no idea. If it doesn't, though, we'd all better learn Slackware and Debian.

Sitemap | Contact Us