Home > Articles > Operating Systems, Server

  • Print
  • + Share This
From the author of

Why KDE Still Doesn't Matter

For all of KDE's strengths, still a few things keep it from really making a difference outside of its loyal following—and particularly outside of the Linux world.


Many common tasks, such as setting up a Web server or configuring the network, force you to run non-KDE applications and might even require you to pull up a command line. You won't find pointers to many of these tasks in the control center or even on the K menu. This means that the novice user is forced to figure out what KDE can do and what KDE can't do—and, worse, how to do what KDE can't.

On a Macintosh, you can configure everything using the Apple menu. In Windows, the Control Panel serves the same function. But on Linux—even with KDE—there's no one place to go to configure everything.

This is an unfortunate side effect of the fragmentation of the Linux desktop between GNOME and KDE. No distribution is willing to fully commit to one; nor is it able to commit to both. As long as this is true, Linux will always be a difficult environment, no matter how pretty the user interface is.


The help files in KDE are better than those in many open source tools, but the quality is variable. Similar to the integration problem, there's no one place to go to ask, "How do I set the background wallpaper?"


KDE is licensed under the GNU General Public License (GPL) and the GNU Lessor General Public License (LPGL), whose restrictive languages scare off commercial developers and even many open source developers. A component developer who wants to create a new KPart rightfully feels uncertain whether she will be required to release her source code. Even an application developer must worry because she will have to link against the Qt toolkit. In both cases, the KDE team claims that the answer is "no" (provided that the vendor pays for a commercial Qt license), but, given the history of the GPL, vendors have reason to be concerned.

After RMS and Debian's claim that KDE couldn't be linked against Qt because of license incompatibilities (an issue that was resolved only by Qt switching to the GPL), who can blame a non-GPL developer for being wary about developing for KDE? This doesn't apply just to closed source developers, either. Qt's original free license, the QPL, was certified as an open source license, but that wasn't good enough to satisfy the GPL guardians. Again, everyone can argue over who was to blame and whether the QPL and GPL were actually incompatible, but that doesn't matter. It's the uncertainty that helps keep vendors away.

The only way to resolve this impasse is for the GPL and the LGPL to be tested in the courts. Only then will legal precedents define what can and cannot be done with (L)GPL code. Few companies are interested in being the test case, of course, because they almost certainly will not win. Even if they win in the courtroom, they will likely lose in public opinion, particularly if the courts decide that the GPL is unenforceable or otherwise void. This fact alone could forever keep Linux off the desktop, but that's a subject for another article.


KDE suffers from the same problem that plagues most open source projects: playing catch-up. Many of the new features in KDE start off as, "I saw this on Windows, and…." There are also some copied features from the Mac and even some from BeOS, but most of the bigger ideas seem to come from Windows.

Don't misunderstand me. KDE does include a number of user interface features that I haven't found other places. For example, being able to easily vertically or horizontally maximize your windows is sorely missed on Windows and the Mac (this could be borrowed from other Unix window managers, but I haven't noticed it anywhere else). KWord promises to blend word processing and desktop publishing in ways in which competitor products haven't succeeded (hopefully bringing the best of Word and FrameMaker together). I'm not saying that there isn't any original work going on, but KDE needs to push further into user interface design if it's going to be sold as anything more than a cheap version of Windows with less software.

  • + Share This
  • 🔖 Save To Your Account