The Dawn of Desktop Linux is Upon Us

I'm dating myself a bit by doing so, but I will confess to having a been a mostly loyal Macintosh user since 1984. Indeed, I had the pleasure of regularly using the historically significant original 128K Macintosh with a single 400K drive. "Distinction" may be a better word than "pleasure," though, as the early criticisms of the original Macintosh were spot on and I can recall the delight when my then employer purchased a second external 400K disk drive. My limited imagination couldn't imagine anyone wanting anything more.

Regardless, I was a true believer in the graphical user interface and the elegance of Apple's way of doing things. By the mid-1990s, however, Apple had dramatically lost its way. The quality of both its machines and its software had deteriorated badly. The hardware had become slow, clumsy, prone to breakage, and confusingly incompatible with other Apple hardware at the edges (think modems and network cards). Decision-making with regard to software updates amounted to balancing whether the new version would be more or less dependable than the current version. Neither was particularly good and both had different problems, so your decision came down to what you most needed to with your machine. The times were dark and I found myself, albeit with much reluctance, contemplating moving my then employer to Windows.

Of course, Windows 95 was not the end all be all of operating systems itself. It fell fall short of the promises that had been made for it. Nonetheless, it appeared to be "good enough", a phrase which probably accurately describes Microsoft's successful product strategy, and it had positive momentum, which Mac OS lacked. I was frustrated and haunted by the lack of choice. Fortunately, my ultimate decision was to wait just a bit longer.

My patience paid off as Steve Jobs and NeXT managed to pull off a takeover of Apple that was widely and erroneously reported in the press as Apple's acquisition of NeXT. Apple's hardware finally started to utilize standard parts. It was also being built to higher standards of quality. In the software department not only did we get a credible roadmap to a modern operating system that ultimately became Mac OS X, but the software updates of the Jobs era could be trusted to work as advertised and almost universally offered improved features and/or performance. In the "old days" one could be proud of being an Apple customer. In the dark days one had no choice but to be embarrassed. As we came out of the dark days, we could stop being embarrassed, but best remained inconspicuous. And, today we can once again be proud. I have personally used every significant desktop environment of the past fifteen years and like Mac OS X on my PowerBook better than anything else I have seen by a comfortable margin. Mac OS X is simply an excellent platform.

Which, in a roundabout sort of way, brings us to desktop Linux. I started playing with Linux in early 1997 with the RedHat Linux 4.0 product. At the time it was a "playing" sort of thing. I had a sense that moving from WebStar/MacHTTP on a Macintosh server to Apache running on Linux might be a good thing and I wanted to find out. Sure enough, it was a good thing. I also wanted to find out, however, if there was any point in running a Linux Desktop. I had not used XWindows a lot and wanted to see what it was all about. And, of course, 1997 was in the middle of Apple's dark ages, so I was looking for possible alternatives.

I did manage to get X running, which I recall being something of an accomplishment in those days. Actually, that can still be a challenge today, though the situation is much, much better. Despite getting things running on an old Compaq 486 that was lying around, it was immediately and painfully clear that Linux was not going to be an alternative on the desktop. It quickly found a home that it has since had in my server room, but I held out no hope that we could replace even a single Macintosh with a Linux workstation.

The times have changed. While Mac OS X is my desktop workhorse, I do keep a copy of the "testing" version of Debian running as a desktop system and it is indeed getting to be very good, though not yet enough so that it wouldn't scare away less technology-inclined users. There is no doubt in my mind that if Mac OS X were to drop off the face of the planet tomorrow, I could use Linux as a desktop system, though I would miss the elegance and seamlessness of the Apple system. Desktop Linux is now "good enough" for a significant portion of the user population, especially in a corporate environment that has support personnel available to get things started. My experience with excellent CD-based Knoppix distribution suggests that we are not that far from being "good enough" for everyone.

Desktop Linux does indeed seem to be poised to go further than "good enough" and it will be the interest of profit-seeking companies that will push it to be a viable alternative for most any computer user. The traditional open source developer community has done a great job of moving desktop Linux to the point where it is usable by a significant minority of users. I suspect, though, that this community will have difficulty moving beyond this point, however, as open source progress depends on the "scratching of itches." Developers who are smart enough to contribute to the development of desktop Linux now have a system that is more than good enough for themselves, so the motivation to go "all the way" is not terribly strong for individual developers in the aggregate. Companies, Sun and Novell in particular, are starting to pick up the scent of potential profits.

Sun's schizophrenia with regard to Linux is worrisome, especially in light of the serious possibility that their investments in SCO were made to stop its momentum. I am convinced enough of Sun's complicity in that mess to have sticken them from my choice of vendors. Nonetheless, the mysteriously named Java Desktop System, which is really a whole lot of SUSE dressed up in more coherent and attractive attire, looks really impressive, though I must admit to not having actually given it a try yet. By the looks of things, however, it seems that they have not only smoothed out the rough edges and simplified things, but also built a package of software that makes sense for many business users.

Similarly, the entrance of Novell into the game is interesting and, while their plans have not fully been disclosed, their purchase of Ximian and SUSE lead one to believe that they, too, will be releasing a strong desktop product. Ximian has a long history of involvement in the GNOME project which has been the basis for their Ximian Desktop product, which has always had a fit and finish that was above and beyond the mainstream Linux distributions.

Sun and Novell are clearly interested in the commercial market and both, despite their foundering of late, have entensive connections into that market. With their motivation to profit from desktop Linux, one can be rest assured that they will be putting resources and effort into making the user experience as good as possible, something that independent developers arguably would have a hard time doing.

Of course, the press have raised the issue of what this new corporate involvement means in terms of the developer community, especially with regard to the relationship between the community and the newly involved commercial entities. The impact is, in all liklihood, nothing but positive, which is the true gift of the GPL, or General Public License. The Free Software Foundation, which created the GPL, wrote the license to be a means of advancing their very firm ideological beliefs. One does not have to subscribe to the license's underlying ideology, however, to benefit from the practicality of the license, at least as it relates to "platform" types of technologies. The real beauty of the GPL may be that it establishes a framework in which groups that otherwise compete with each other can work together to advance computing for the benefit of everyone. Indeed, had the GPL been adopted for UNIX itself, it is within the realm of possibilities that the rise of Windows as a server platform would have been mitigated and all of the UNIX vendors would have done better in that business.

Thus, the contributions of the commercial players to the core platform technologies will remain "free," as in beer and speech. At the same time, the commercial entities will pioneer the "packaging" to the masses, the lessons of which can be learned by those seeking to create non-commercial dsektop distributions. Aguably, in the minds of many, including myself, Apple has already set the target. Mac OS X could exist just as nicely on a GNU/Linux foundation as it does on a Mach/FreeBSD foundation. There is no reason an end user on Mac OS X ever has to know that they are running on a sophisticated *nix foundation, though it is there for those who wish to utilize it. This column is being written using vim on Mac OS X, for example. Mac OS X is proof that *nix-based operating systems can be easy enough for anyone to use.

And the lessons will, I think, be learned and implemented by those who wish to create competitive non-commercial distributions. Knoppix is an excellent example. Its ability to autodetect and configure its hardware environment, and the usability of its graphical interface is exceptional. Thus, the "free" software community can create fit and finish. It's just a matter of getting the right leadership in place to make it happen.

What is finally being learned is that Microsoft, through its massive relatively low priced shipments of Windows, commoditized the market for core platform technologies, that being the kernel, the basic libraries, and the windowing system itself. Competitors can certainly improve upon the Microsoft offerings, but the general market is not willing to pay any premium for better core technologies. What they will pay a premium for is better fit and finish (read usability), better security, and better stability with the qualification that the applications they need be available on any system they choose. The GPL allows commercial competitors to work together in conjunction with independent developers to create viable computing alternatives for regular desktop users. The core technologies are now good enough and the "fit and finish" is starting to appear. 2004 will see desktop Linux become a reality and, if the parties involved work productively together, we could see desktop Linux as a viable mainstream alternative in 2005. Choice is good for everyone. Sonate delectum!