Although I regularly use KDE, Xfce, and other desktops for GNU/Linux, I keep returning to GNOME. Sometimes I use the default Metacity window manager, and other times the quicker Sawfish, but, with either choice, GNOME has an uncluttered look that allows me to focus on my work rather than my software. It also contains enough customization that I can easily set my increasingly long list of preferences with a minimum of effort.
However, my loyalty is far from unqualified. Despite being in development for a decade, GNOME still lacks one or two utilities that I consider essential. Other default tools are lacking in functionality, or could stand a modern redesign. I also question some of GNOME’s policy decisions.
If I can be excused from building dream castles and planning to move on the first of next month (as Harlan Ellison would say), here are the improvements that I would most like to see in GNOME:
A font manager: Font administration is a basic necessity for any graphic designer. Designers need to load and unload fonts, so that they don’t overload their system with thousands of fonts. They also need to enable or disable fonts by groups, so they can load all the fonts for one project with a minimum of mouse clicks. KDE 4 can manage fonts this way, so why can’t GNOME? If nothing else, it could make Fontmatrix part of the default installation.
A multiple-item clipboard: Desktops on different operating systems have had single item clipboards since the early 1990s. However, as anyone who has used MS Word’s clipboard or KDE’s Klipper can testify, the ability to store multiple threads of text or images for reuse is a valuable editing tool.
Moreover, a multiple clipboard for GNOME already exists. It’s called Glipper, and installs as a panel app. Glipper does have the drawback of supporting only text, but making it support some basic graphics formats shouldn’t require much development.
Better graphics for games: The standard package of GNOME games is varied enough, but they look like they were written in the 1980s. Compare AisleRiot solitaire to Pysol, or Five or More to the equivalent Kolor Lines in KDE, and the need for redesign becomes obvious.
Granted, I’m only talking about games, not essential productivity apps, but sooner or later, almost every user opens one, if only to take a break or to wait for a critical download. They’re some of GNOME’s frequently used programs, and they create a poor impression of the desktop as whole.
A better file manager: For all its ability to burn DVDs, Nautilus remains a mediocre file manager at best. Its default view shows only a limited number of file attributes, and its navigation is limited to a small directory tree that is hidden on the bottom left and shows only the current path, requiring you to click endlessly to get anywhere. A much better directory tree is available if you use the command
nautilus --no-desktop --browser. But some distributions, including Ubuntu, don’t offer that alternative, while others hide it the usually crowded System Tools menu.
Other problems? To start with, the only way to open another directory view is to start another instance, which takes up unnecessary space on the screen. The ability to open multiple directory views in Nautilus would save space and make drag and drop operations more efficient.
I’d also appreciate a way to turn off the account-centric view that emphasizes the current user’s home directory and desktop. That option may be fine for new users, but more experienced users are as likely to want to view system files as personal ones, and to them the extra layer of abstraction is just an annoyance.
Inspiration for the redesign of Nautilus could come from KDE’s Dolphin or Krusader, both of which are much more useful as file managers. Alternatively, a leaner browser without extras like Emblems that nobody uses could come from Xfce’s Thunar.
An overhaul of the System menu: The decision to offer a centralized set of configuration tools, like the Control Center in KDE 3, or a set of independent menu items, as in GNOME, is largely a matter of design philosophy. In the abstract, I could tolerate either.
Yet, that said, GNOME’s System menu could use a major reorganization. What falls under Preferences or Administration is next to impossible to remember, or even reason out. Even worse, the menus contain separate entries for such items as Keyboard Configuration and Keyboard Shortcuts, or Network and Network Tools.
Most irritating of all, some functions are split over a number of different menu items. Probably the most dispersed ones are the preference settings for preferred applications. Default choices for some commonly used programs are set in Preferred Applications. However, to set the programs to use with cameras, scanners, or tablets, you go to Removable Drives and Devices. Choices for accessibility and for general look and feel are similarly scattered. Perhaps the solution is for the System menu to abandon the alphabetical order of other menus and group related items together.
More choices in Preferred Applications: Right now, Preferred Applications set only the web browser, mail reader, terminal, and music player. These are some of the most commonly used categories of programs, but why stop there? Developers might welcome a setting for text editors, and office users one for word processor or office suite. Nautilus’s settings for various MIME types could also go in the same dialog window.
Better access to Gconf: Gconf is the master GNOME configuration file. It is sometimes referred to as the GNOME Registry in an echo of Windows usage, and also because, like the Windows’ registry, Gconf is lengthy and complicated to read.
Some programs, like Ubuntu Tweak, give you control of some of Gconf’s arcana through a graphical interface. However, intermediate and advanced users might appreciate an editing tool that allows them to access and navigate Gconf itself.
The removal of Epiphany: In theory, Epiphany is supposed to be the default GNOME browser. But, while some users might prefer Epiphany to its fellow Mozilla-based browser FireFox, they’re a minority.
Some distributions remove Epiphany as a matter of course, but a surprising number don’t, including Debian. Instead, it lingers, showing a surprising tendency to spring into action, no matter what the Preferred Application settings for a web browser happen to be.
Eliminate Mono: I don’t start automatically at the mention of Mono simply because it is an implementation of Microsoft’s .NET. Nor am I made more than cautious by unproven suggestions that Mono might be vulnerable to claims of patent infringement. However, I do question GNOME’s insistence on a feature that so many people object to, especially when so many other programming languages are available.
I also wonder about the long-term effects of writing add-ons in another language instead of using C like the core of GNOME. Could speed and efficiency be compromised? If so, then giving developers a choice of languages could come at the cost of a major cleanup down the road.
However you look at Mono, it just doesn’t seem necessary. The default Debian installation of GNOME does quite nicely without it, and so could other distributions.
Searching for the perfect desktop
None of these lacks, irritations, or disagreements is enough to make me abandon GNOME permanently — although KDE 4 is looking very appealing, I must admit. Many can be remedied with customization or installation of a few additional pieces of software, and I’ve done enough installs that I can quickly set up to do things my way.
Still, every once in a while, one of the deficiencies I’ve mentioned is enough to send me roaming after alternatives, especially when I think of how long I’ve waited in some cases for improvements. Yet, in almost every case, a list of problems that is equally long sends me back to GNOME in a preference for the problems I know rather than new ones.
Probably, the only way to be completely satisfied would be to write my own desktop, something that I lack both the skill and time to do. Failing that, though, I’ll settle on GNOME.