Konqueror vs. Nautilus and Epiphany. It's difficult to compare these applications. Konqueror acts as a hybrid between file manager and web browser, just like Explorer for Windows. Gnome took another route and split these two tasks between the file manager Nautilus and the browser Epiphany. I guess it's completely impossible to compare Konqueror, the file manager and Nautilus because both handle this task completely differently (read more about this here). The browsing capabilities, however, show a small advantage for Epiphany, mainly because it uses Mozilla's rendering engine, which has, due to the tremendous success of Firefox right now, more support. I took Konqueror on a couple of small surfing sessions and visited all my bookmarked sites. It had severe problems with two out of 28 websites, all the others displayed flawlessly. I understand that these sites are composed in old-school, bad HTML (and I would help rewrite them today). When I do web design, I never have problems with KHTML. I guess Konqueror just has to improve its routines to handle HTML in compatibility mode (a programming job that's probably quite a pain in the ass, so to say). I hope that Apple's cooperation will bring KHTML on par with Gecko fast.
The main advantage of Gnome over KDE is definitely the better menu structure and the rigid rules on interface design that the Gnome Human Interface Guidelines impose. KDE's menu structure is a big mess compared to Gnome's. Just compare the screenshots over at osdir.com between Ubuntu Warty (screenshots 26-32) and SimplyMepis 2004.04 (screenshots 52-65). The KDE team has to do something about this because it effectively worsens the usability of KDE. Take JuK or Kuickshow as positive example and throw some more bloat out of the other applications. The KDE Control Center, for example, is a nightmare. I definitely think Gnome's way of handling things is better here: the Gnome team tries to reduce all programs to their most important tasks so you have easy access to these most important functions. All the other options can be changed too, for example, through the command line or the GConf-Editor. Many KDE applications on the other hand try to present every possible option to the user, but that doesn't make sense to me: that just diminishes the usability convenience for all users, experts and novices alike, because it takes longer to find the few important, everyday features that are so clearly presented in Gnome applications. So, adding more features to an already "full", if not supercharged user interface offers nothing but diminishing returns. Also, who is the targeted user audience? Experts are able to edit text-files or use the command line anyway, novices are just scared away by too many features.
Let's stop here because these are the applications that I use regularly. There are of course a lot more applications, but as I don't use them I did not feel qualified to write about them. I intentionally left the system configuration tools out: if you've made it this far, you probably know that I don't like the KDE Control Center, and the Gnome System Tools, as presented in Ubuntu Linux 4.10, are nice, but far from being complete. I also left out one program that I use all the time, the GIMP because I think that it does not have a KDE competitor. Additionally, the GIMP is also not a real Gnome application.
So what's my final verdict? Well surprise, surprise, KDE 3.2.3 won't win over a Gnome aficionado like me. But I was impressed by a lot of things in KDE and I think I'd be able to switch to a KDE-based distribution pretty easily if I had to. Now that I've seen how applications like JuK or Quanta Plus have evolved, I see Gnome's problems more distinctly. I'll hope that Gnome 2.10 solves some of these issues. If not, I might take a close look at KDE again.
Finally, a prediction for 2005, since it's the end of the year and everybody around me is making predictions too: desktop environments will become even more important. We will pick our distribution of choice even more because of the flagship desktop they include. Distribution makers will be well advised to pick either KDE or Gnome and throw their entire force behind making their selected environment as polished as possible. This is probably the best way to produce a really well integrated distribution. At the end of the 2005, we will not have the "Linux Desktop", but the "Gnome Desktop" and the "KDE Desktop", both Linux, GNU and X powered, with a highly modified and integrated OpenOffice.org as Office Suite (and maybe Firefox as browser). Many people will choose inferior programs because they come with their desktop environment instead of a mixture of the best applications available (just the way every single Rhythmbox user effectively does right now). And I don't think that this will be necessarily a bad development because this will hopefully boost development speed until KDE and Gnome both are full featured, nicely usable desktop environments, each with its own set of quality applications for every major task, each on par with, let's say Mac OS X. That also means we will see less graphical applications that don't "belong" to either desktop, thus probably even less commercial 3rd party applications. And the ones that do survive (Firefox, OpenOffice.org) will do their best to integrate with Gnome and KDE. In other words, happy times ahead, folks!
Christian Paratschek, 28, likes living in his Gnome cave, but eventually comes out to examine the wooden huts that the KDE Neanderthals have built recently. Then he mumbles something about UI design and goes back to sleep in his fully HIG-ified cave.