When opposites attract

While Linux has been an unquestioned success, most of that success has come on the server. Success for the Linux desktop, despite significant advances, is still a long, long way from challenging Microsoft Windows on the desktop.

Part of the problem, no doubt, is Linux's desktop diversity. The same broad range of choices that Linux users find so valuable among their server products and suppliers looms as a weakness as Linux seeks approval by the world's non-techies. Most of the computerphobes I know want as little decision-making as possible -- they just want tools that work and they don't mind entrusting Microsoft or Apple or anyone else to make the decision for them. Which choice to make is less important than the fact that a viable (and supported) choice exists.

And don't even try to talk to these casual users about why the approach they're using may be technically inferior to another that is (according to some techie) worth trying. Inertia, as a force against abandonment of old techniques and adoption of new ones, can't be ignored on the desktop.

The desktop duet
In this context we have the ongoing rivalry between GNOME and KDE, the two Linux desktop environments that have pretty well knocked any other approaches to the sidelines. While both are fairly large (read: RAM and CPU hogs) and there's still demand for light and nimble approaches such as XFce, the two biggies are definitely the main challengers for King of the Linux Desktop. Problem is, opinions on which one is tops are, shall we say, quite varied.

Most distributions have chosen clear favorites, with Red Hat and Debian defaulting to GNOME, and many of the other major players picking KDE. I personally like the approaches of Red Hat and Mandrake, which have their preferences but allow you to install the other, or both at the same time, making it easy for users to experience both environments. I occasionally switch between the two.

Consumer choices or consumer confusion?
But most people, especially the non-computer-savvy, don't see value in such choice. They want a winner and a loser, a VHS and a Betamax. This view was best expressed by Dave Whitinger in a Linux Magazine article some months back, in which he advocated a single standardized Linux desktop. Whitinger received a bunch of negative feedback from folks who didn't want to sacrifice Linux diversity and innovation in order to simplify a choice for non-technical users. Still, Whitinger had a point, even if (to me) his original solution was overkill.

But users like Whitinger weren't the only ones interested in reducing confusion. Software developers, who have so far been forced to take sides, are also interested in a resolution. As a result of the dueling desktops we have fantastic apps coming out of both camps -- such as KDE's Konqueror browser, and the Nautilus file manager being developed for GNOME by Eazel.

Compounding the problem, we also have projects such as KWord (for KDE) and AbiWord (for GNOME). These two word processors are coming along nicely but neither is really anywhere near fully cooked. How much further along might the community be if all these people were to get together and produce an open source Microsoft Word killer rather than undertake this parallel development?

Not long ago, my colleague Joe Barr at LinuxWorld suggested that he might switch to GNOME just to be able to use its excellent Balsa mail reader.

If that happened it would be a shame. We shouldn't force a situation in which applications written for one environment won't work well on the other, and in which users have to balance the available apps when choosing their environment. Someone wanting both Koffice and Gimp would be forced to have one of these packages operate at less than maximum functionality. To be sure, the basic functions are all there -- all of these applications are using the X Window system -- but the added modularity possible in either of these two desktops is unavailable for applications written under the other system.

So far.

A meeting of minds

In what may be (we can hope) one of the most significant events to happen in the history of Linux desktops, folks in the GNOME and KDE camps are talking to each other about integrating some of the low-level modular functions. These subsystems -- called KParts on the KDE side and Bonobo in the GNOME world -- have different underlying technologies, but they serve generally the same purpose. They allow developers to make components that are highly re-usable between applications, similar to Microsoft's Component Object Model (COM) architecture. The good news is that Kparts and Bonobo developers are being encouraged to combine their efforts -- or at least provide some framework that lets apps written in one system maintain all their component functionality on the other.

It all started when a troller entered GNOME and KDE developer areas and goaded them all for not doing more to interoperate. This resulted in GNOME developers joining a KDE chat area where the discussions started. From there, enthusiast Casey Allen Shobe (who prefers to be known as "Rivyn") started a publicity campaign to encourage the talks to continue. He started collecting names for an online petition of sorts, to show parties on both sides how much support exists.

I think this is very helpful, and I encourage you to add your name (either at the bottom of the current list, or directly by mail to Rivyn. The Open Source world evolved from an attitude of "If you don't like the program, rewrite it yourself," but many of the people who would benefit the most from Kparts/Bonobo cooperation are likely to not be programmers.

Rivyn, though aware of the technical and political hurdles required to bridge this particular gulf, is optimistic that it can happen -- especially if his effort can generate enough support. "It is my hope, that with enough names on the list, we can convince the developers to get along," he said.

Everyone benefits
So far the results are encouraging, but it's far from time to celebrate. The discussion raised by the proposal, within both the GNOME and KDE developer communities, has been far more positive than negative. Once the inertia of the politics and such is overcome, the technical end should be doable.

I share Rivyn's hope. The implications of having a truly interoperable system, where components will work no matter what desktop you choose, are far-reaching. Indeed, it will make the choice of desktop purely one of personal preference, a far easier counter to the forces of inertia than a lengthy explanation of which apps work best with which desktops.

Furthermore, a common desktop/programming architecture allows application developers, both commercial and open source, to concentrate on the best possible code and tools rather than on which side to pick. It's even a choice that Whitinger can live with, given that he's not going to get his wish -- which he maintains despite the flames -- of a single dominant desktop. Well, he's not going to get that any time soon, but this is the next best thing.

I personally have no problem at all with multiple Linux desktops. Let KDE and GNOME slug it out, let them take ideas and maybe even code from each other -- that's what free software is all about. It's no surprise that newer versions of KDE and GNOME look far more like each other than when they started. Who knows? Maybe rather than one killing off the other they'll just sort of quietly meet in the middle one day. But I'm in no rush, and it's fine with me if that never happens and the two keep pushing each other ahead. What's important is that end-users and developers won't be penalized by this soon-to-be-friendlier (we hope) rivalry.

Is the GNOME/KDE rivalry hurting desktop acceptance of Linux? Let us know in the TalkBack below or in the ZDNet Linux Forum. Or write to Evan directly at evan@starnix.com.