Several weeks ago, I ended a comparison of the KDE 4 and 3 desktops by saying "Unless a project takes over KDE 3 development, sooner or later it may become unusable with the latest generation of computers."
What I had missed -- free software being a large place where events move at near-light speeds -- was that a project had already taken over KDE 3 development. It's called Trinity KDE, and is organized by Timothy Pearson, who has been releasing Kubuntu releases that use KDE 3.5 for some time. According to Facebook rumor, he has been planning to revive KDE 3 for some time.
I had hoped to interview Pearson about his rationale and motives, but an expensive server crash has kept him too busy to talk, although I still hope to talk to him in the next few days.
Meanwhile, you can easily fill in the blanks about why the project is being enthusiastically received. Although the KDE 4 series has become more usable with every point release, there are still those who condemn it as too bloated, complicated, or resource-hungry. Others suggest that KDE 4 has fewer features, although that is less true with every release, and is sometimes an impression due to a renaming or rearranging of features. Still others, I suspect, are simply nostalgic for a desktop that they used for years.
The truth is, the KDE 4 series has never entirely recovered from the disastrous KDE 4.0 release -- no matter whose fault that was. Core KDE developers like Aaron Seigo can now make wry jokes about that events they found personally disheartening, but others have never recovered from the gap between expectations and reality with KDE 4.0.
Trinity KDE 3
What we don't know, of course, is how many would support a revived KDE 3, or whether those who disdain KDE 4 are a sizable minority or even a majority, or simply very loud. But, with Trinity KDE, we are about to find out -- and witness, as well, just how complicated bringing a mostly dead desktop back to life can actually be.
Taking a test drive
Another thing you can do while waiting to learn more about Trinity KDE is take it for a test run in its Kubuntu incarnation. Labelled KDE 3.5.11, this release is available as a Live CD, which you can install in Ubuntu's usual seven simple steps.
At this late date, a detailed review of KDE 3 seems ludicrous. After all, most of those likely to try Trinity KDE at this stage are likely to be doing so because they already like KDE 3.0. Under these circumstances, it is enough to say that most of the long-familiar aspects of the late KDE 3 series are faithfully included in Trinity KDE, from the initial setup up wizard to the classic menu, the five different types of panels, and Konqueror positioned as both a file manager and web browser, instead of abdicating file management to Dolphin.
So what is different in Trinity KDE compared to the original KDE 3 series?
First of all, basic desktop operations like opening and closing windows and menus are noticeably faster in the Trinity KDE version of Kubuntu than the KDE 4 version. However, the importance of this speed no doubt depends on your hardware -- on a new machine with several gigabytes of RAM, the difference still exists, but is greatly reduced. Perhaps Trinity KDE might become a choice for older machines, although Xfce or LXDE would be a faster choice.
Second, as you explore Trinity KDE via Kubuntu, expect a few changes. For example, instead of the KDE Control Center, you'll find a System Settings window much like that in KDE 4, with high-level categories and a General and an Advanced tab. Similarly, you'll also find support for ICC color profiles in the desktop, and an entry for the task manager in the panel's right-click menu. However, most of the changes are described on the Kubuntu page as being behind the scene improvements or bug-fixes.
You can see the same mixture of old and new in the selection of applications. Trinity KDE is supposed to be able to coexist with KDE 4.x and run its applications, but Trinity KDE's own choice of applications are those designed to run with the KDE 3.x series -- and this limitation is probably the project's least appealing feature.
The trouble is, almost all of Trinity KDE's native apps are represented by versions two to three years old. Some of Trinity KDE native apps, such as digiKam, KMail, or Amarok, are not that far from approximating the latest KDE 4.4 apps, but they lack some of the features and stability of their KDE4 namesakes. When you do see a late model app, like KDE 3.2, it is a third-party app.
These native apps are the end results of long development cycles, so most people should find them adequate for general use. Still, they raise an interesting question: Is using older apps a fair trade off for using the desktop of your choice?
While some might agree, others might argue that this choice is a reversal of priorities, and the apps are what matters.
For that matter, is KDE 3.5.11's speed and stability worth doing without the innovations of the KDE 4 desktop, such as folder view, the semantic desktop, the social desktop, remotely run widgets, or arranging apps in tabs in the same window?
The question naturally arises, because the impression is that Trinity KDE is a long way from having such features. But, then, its supporters might not want it any other way.
Expected growing pains
Besides the question of how many people will want the KDE 3 series extended, Trinity KDE also faces a number of other challenges, as Sebastian Kügler points out.
To start with, the KDE 3 series depends on the Qt3 toolkit, which is no longer supported by TrollTech, the company that develops it, although documentation is still available online. That means that Trinity KDE is faced with either working with an outdated toolkit, or of taking on the large task of porting the KDE 3 series to Qt4.
After all, the rewriting of KDE with the 4.0 release was not just a whim of bored programmers. A large part of the reason for the rewrite was that KDE developers were frustrated by the limitations of Qt3.
Moreover, the reason people use major desktops like KDE is partly because of the extensive ecosystem of applications that spring up around them. To be a major desktop, Trinity KDE either needs to attract a large number of developers to support this ecosystem, or else convince projects that have embraced the KDE 4 series to also maintain KDE 3 versions. While neither task is impossible, both will require strong organizational skills to carry out. A project's community does not simply develop overnight.
None of these challenges is insurmountable, but collectively they do mean that Trinity KDE already has programming and organizational obstacles ahead. Personally, I view the KDE 3 series as good in its time but superseded, but Trinity KDE's mere existence is so quixotic that I find myself hoping that it will become a robust, living alternative, and not just a shambling shadow of KDE 3 hobbling along on limited contributions.
Besides, the free software world can always use a new desktop -- even if this one isn't altogether new.