Usability of Free Software Desktops: a rant from 2002

In 2001/2002 I wrote for LinuxFormat a two parts tutorial for end users on how to configure Gnu/Linux desktops on limited hardware, that in part summarized what we were then doing in the RULE project. Before the actual tutorials, I had also written down my own, very personal motivations for playing with that kind of tricks. The reason was to clarify, first to myself and then to the editor, what itches I would try to scratch, so that text was never published, and I basically forgot about it. I rediscovered that file, dated September 19th, 2002, only this morning. Reading it again, in these days of Unity, tablets and touchscreens, made me think it may be fun to share it. Remember when it was written, judge by yourself and let me know.

TITLE: Are Gnome and KDE really useable? Are KDE and GNOME suitable for children/students?

A look to the desktop side that KDE, GNOME, and many Linux distribution are missing (not that commercial software never cared…)


Today, everybody knows how and why Linux (and Open Source software in general) is technically superior to closed solutions. It finally has the deserved recognition in the server market recognition it deserves, performing optimally even on obsolete hardware. Many commercial tools have been ported to it, showing that the two worlds can (and, in my opinion, should) coexist.

So, what’s still missing to world domination? The desktop, of course. Apparently, Linux has made huge steps forward in that direction too, with the KDE and GNOME projects: today it is much easier to do WYSIWYG word processing, spreadsheets, games, music, video and games using only Free Software (as in Free Speech of course), in fully graphical and well integrated environments.

Certainly Gnome and KDE are the easiest and more realistic migration path to wean non technical users from the Redmond playpen. I highly respect all the developers, and consider outstanding software many of their applications.

On the other hand, I also think that KDE and GNOME risk to limit or delay the worldwide spread of a Free desktop, because they seem to forget what real desktop usaability should be about.


In KDE and Gnome, you can do everything just clicking on many wonderful icons; all windows have the same borders and fonts; you can cut, paste, drag and drop everything to everywhere. Is this desktop paradise? Yes, if:

  • the Redmond way is the right way
  • you do accept their default tool(s) for every task
  • you do accept their default way of doing things
  • you do have a modern computer, with much more CPU power, RAM and HD space than are really needed

Sure, there are thousands of other Linux applications, all free and customizable. I know it, but your six year old son, your grandmother, or your technically challenged coworker don’t.

They can buy a PC, put the Linux CD in the tray, and have everything installed in twenty minutes. But what do they find on the screen? The confirmation that, even with Linux, to do desktop computing you still must:

  • give up at least 10/15 percent of your display to buttons, panels and taskbars,
  • have the latest CPU and truckloads of RAM and disk space.

Oh, yes, and quite some money to buy all that. Just as commercial ones, free modern desktops perpetuate the dogma that more is necessarily better, forgetting that while software can be free as free beer, hardware cannot.

They say that Linux might even transform any 486 junk into an excellent server, but on the desktop it must waste almost as many resources as other platforms, in the same way (without blue screens, of course). We seem to have lost the passion for doing always more with the same limited material resources.


Shell scripts and commands give the maximum power, efficiency and flexibility, but remembering them all is not easy, and opening a terminal just to type six or seven characters doesn’t make much sense. When we closely look at it, this is the main, if not the only reason, why many people become addict to full fledged mouse-based environments, and limit themselves to the ready made ones.

I know that, without a common graphical toolkit it would take much more time to develop and integrate new applications, and RAM needs would be higher. I am a bit lost, however, whenever I read proud assertions that Unix now has a really useable desktop “because all windows look the same”.

To me, from the point of view of a non competent desktop user, useability means:

MINIMIZE ALL COSTS: without touching real functionality

SPEED, but the one coming from running only what you actually need, not by fan-cooled dual processors

DOCUMENTATION: one common, searchable frontend for everything, be it HTML, man or info

COMMON SHORTCUTS: an easy way to have or remap the same key shortcuts across all applications

FREEDOM TO PICK UP ONLY WHAT YOU ACTUALLY NEED. Recently, I asked in several mailing lists what I should install on a four years old PC to surf the web graphically. I wanted SSL, Cookies, JavaScript but no email/news/ftp and so on. The result was:

  • Konqueror: very good, but you have to install and run a bunch of KDE pieces anyway, so what’s the point?
  • Konq-e: probably the solution, if you know how to compile by hand lots of stuff
  • Galeon: very good, small and ligth on resources… if Gnome and Mozilla are installed
  • Skipstone: almost like Galeon

The truth is that many modern applications are lightweight only inside their complete default environment. If not, you must have quite some skills and spare time to recompile them.

NO PANELS OR ICONS: screen space is too valuable to do in this way what you could still do with the mouse through the root menu, or a file manager window. “Drag any file over the printer icon… TA-DAH!!” Uh? What is faster (i.e. more useable) for a mouse-only newbie? Left click on something
inside any file manager, and select “Print” on the resulting
menu, or drag it all the way to the other corner of the

NO MULTIPLE WORKSPACES: yes, if your boss enters the cubicle while you are chatting with your mistress on a Playboy immersed in company memos and budget calculations. Apart from that, let’s admit that most human beings do things serially to avoid mistakes. Those shock-full screenshots demonstrate how much you can do in Linux, but who really writes, clicks and read into 20 windows simultaneously?

8 thoughts on “Usability of Free Software Desktops: a rant from 2002”

  1. I thought the rant was invalid, it was also most certainly invalid for 2002.

    I have been involved in IT since the 80s, and I watched personal computing grow. While technically minded people like you and me will happily work and play on Command line interfaces, ordinary non technical people won’t and for physiological reasons can’t (they are simply unable to relate to a blank screen, in many cases they have trouble interacting with a graphical interface). Until Graphical Interfaces were available personal computer take up was was relatively slow.

    Linux without modern, even flashy, graphical interfaces stands no chance of being taken up by ordinary non technical users.

  2. I disagree with the last paragraph – about multiple workspaces.

    I don’t think I’m a particularly heavy duty user and my desktop would be hopelessly cluttered without multiple workspaces. If I’m just sitting down to check my mail, I’ll probably only have that one window open. But if I’m sitting down to get something done I’ll have:

    * email, a file browser and a web browser open on one workspace.

    * music player, volume control and another file browser in a second workspace

    * A remote desktop session into each of two or three servers, almost full screen, each in its own workspace

    * A telnet or ssh session into each of several applications, one for each current customer service issue, all on one desktop and a web browser for reading reports and documentation.

    * at least one spare workspace for scratchpad type work

    I rarely have less than three workspace actively in use, often as many five or six so my window manager (jwm) is set up with eight. They don’t cost anything if I’m not using them, as I have a decent size monitor at 1920×1080 (though only one).

    If I were using a system with less screen real estate, I’d opt for flwm as my window manager. It offers multiple desktops but no “tray” (task bar) nor icons – everything works off of the right-click menu – probably fits your “No panels or icons” system perfectly.

    1. HI Liam,

      personally, I confess that I don’t really get workspaces much more than I did ten years ago, when I had a much smaller monitor than today. I have roughly the same number of windows open as you, on average, and get by without any need to have them in different workspaces. Probably because I can’t get anything done by looking at more than 2/3 windows at the same time. I wonder how much it depends of one’s type of work, and how much from one’s type of brain, since every brain works in a slightly different way.

  3. I have to agree with Tracyanne to a certain degree.

    Most computer users don’t care whether their computer’s performance is caused by fan cooled dual processors. They don’t care that the graphical desktop takes up screen real estate because the alternative of having to use a terminal is much more difficult.

    As long as the user can run the applications that they want to run, which for the majority will be browsers, audio tools, email applications and possibly office applications, they don’t really care that their desktop is cluttered with taskbars and icons.

    Now techies are different. We like our screen to be uncluttered. We like to maximise our computer’s resources by having applications that perform well. We aren’t scared to run applications from the command line or give up some of the pretty graphics for performance and functionality.

    In the Linux world both parties are catered for. If you want an all singing all dancing desktop you can have Unity, Cinnamon or Gnome 3 but equally if you like substance over style then you can run lightweight desktops or no desktops at all. You can run everything from the command line if you so wish. Windows has never given you that option. (Well since Windows 95 anyway)

    1. HI Gary, Traceanne

      my point was that sure, novices don’t care if the desktop is cluttered etc… but also that they don’t really need such a desktop (especially not if it slows the system down). And this doesn’t necessarily mean “command line only”, think to WMs like blackbox & CO. Sure they _look_ intimidating, but with a good 5 minutes tutorial, I do believe a novice would be just as productive / efficient / self sufficient as with Gnome, KDE etc.. And I don’t think that is the ONLY way, only that having only overcharged DEs would be very, very bad for _nongeeks_.

  4. I think were all missing the whole reason for GUI’s and extras. As these devices are being designed for everyone, and everyone learns/remembers/works differently based on personal psyche and influence, along with tasking.
    Look at the swiss army knife for instance… crappy knife but with enough tools to get a job done (A tool for a multitude of people) which is why its popular. Its not as efficient as a bayonet or a machete, but not everyone needs a bayonet or a machete.
    The very same instances compare to the PC industry. Programmers (Abstract thinkers for the most part) only need a terminal. But secretary’s may need additional visual stimulation to remember where the options to parse, save and spell check a document are. You can also add the many proprietary formats and now you have a mess of options involved. All this for someone who may need a schedule to remind the boss of things to be done. The boss may be a plumber, and I’m sure he would rather concern himself with plumbing than VI commands.
    Point is that there are a variety of people in numbers that far outnumber the amount of people needing computers that are adept to the concepts of text in a terminal. This is why we all do things differently and need a “swiss army knife” for most General Purpose Computing. But there are still points where using an optimized machine are better (Servers, etc…).
    As time moves on, we will be optimizing and using less desktop space, doing more with less. But don’t forget all of the external variables involved in making a usable/sell-able platform for the times we are in or have been in.

    1. Hi Vince,
      my point was not to ignore “all of the external variables involved in making a usable/sell-able platform”. It was to suggest that the way to deal with those variables was not the best one

Comments are closed.