Disappointment with GNOME 3 has users pining for GNOME 2. But it's not always wise to resurrect the dead.
In the past year, the GNOME 2 release series has become the zombie of the free desktop.
Although replaced by GNOME 3, it refuses to die. Its mourners flock to Xfce, in which they see a resemblance of the deceased. It lives an undead existence in a crippled version in the current GNOME's fallback mode.
More recently, Linux Mint has taken on Dr. Frankenstein's role, trying to revive GNOME 2 in MATE, and reincarnate it as Cinnamon.
Since most of these efforts are not supported by the GNOME project, they have an air of heroism. Users are insisting on having what they want, no matter what developers decree, and I have to admire their persistence.
However, I find myself asking: What was so great about the GNOME 2 series? It had reached its limits a couple of years before GNOME 3 was released and its last few releases were distinctly minor.
Maybe as a primarily KDE user, I should keep quiet. Yet I can't help remembering that GNOME 2 had at least 7 annoying features that I, for one, would be happy to see dead and buried.
So before people rush to support its revival, maybe they should take a moment to remember all that is wrong about GNOME 2, instead of idolizing it uncritically, as if their brains had been eaten by something shuffling from its grave.
Tools No One Wants to Use
Like a zombie trailing bits of itself, through much of its history GNOME 2 continued to ship with tools that nobody wanted.
For instance, Epiphany is a Mozilla-based browser with fewer features than Firefox, and little, if any, corresponding improvement in speed. Similarly, although Evolution was popular seven or eight years ago as a personal information manager, in recent years it's been neglected the way that Microsoft once neglected Internet Explorer.
Some distributions also used to install with the AbiWord word processor and Gnumeric spreadsheet, the only parts of the GNOME Office Suite that were contemplated before the release of OpenOffice.org's code cut the legs from under all the other alternatives.
Not that anything is wrong with AbiWord or Gnumeric. AbiWord is one of the few word processors designed according to how most users actually work, while Gnumeric's calculation speeds are second to none, especially with complex formulas. But, as with Epiphany and Evolution, the inclusion of AbiWord and Gnumeric has more to do with the history of GNOME than with the needs of users.
The Applications / Places / System Menu
GNOME 2's three-part menu might be called its main identifying feature. No installation is without it. Yet, for all its omnipresence, it's irritating to use.
For one thing, the Places menu is uneditable. Unlike the other two menus, you can neither add nor subtract items to the Place menu, nor remove the top-level menu. Even deleting one of the directories listed doesn't update the menu. Moreover, the less you use the default sub-directories in your home folder, the more useless the Places menu was.
Similarly, the division between Applications and System menus might make sense, except that one top-level item in Applications is System Tools. You never know when to look in System Tools or the System menu.
For that matter, the distinction between Preferences and Administration in the System menu can blur, too, depending on the distribution. Such problems should ensure that this triple-headed dog of a menu system isn't missed.
The Classic Menu
GNOME 2's Application menu is what is now described as classic. As you open each menu level, the menu drops open like pieces of fleshing dangling from a decaying arm. With too many items in one level, the menu descends to the bottom of the screen.
With too many sub-levels, it takes over the horizontal space on the desktop, throwing all open windows into the background.
An extreme example of this problem is the infamous Debian menu, which may list everything installed on the system, but also goes at least five levels deep. Assuming your menu is in the top left of the screen, two-thirds of the horizontal space on the screen can be taken over by the menu.
The one solution for these problems that was ever tried was to limit the number of menu items. This solution has the advantage of keeping new users from being intimidated by the sheer number of applications, but has the disadvantage of preventing the discovery of any application not listed.
The classical menu might have been adequate when the average hard drive was 200 megabytes (although I recall problems even then). But, today, when users routinely fill two or three terabyte hard drives, it's an inadequate anachronism.
The Stagnation of Applets
Applets are the small applications that can be added to the GNOME 2 panel. In their heyday, they made the GNOME panel a place that users could customize at will.
However, at some point, the GNOME project turned away from the idea of applets. Aside from one or two exceptions like the Tomboy note applet, new applets stopped appearing in default installations. What remains is a few utilitarian applets like clocks and system monitors, and a few traditional jokes like the Fish or Eyes. In other words, long before GNOME 3 designers dismissed applets as clutter, applets were being unofficially discouraged in GNOME.
By contrast, look at the supporting ecosystem to be found in KDE's widgets, that environment's equivalent of applets. KDE widgets include tools for maintaining different profiles of key applications, quick links to social applications, a magnifier, and a virtual keyboard -- and more are being developed every week. Discouraging applets, you might say, destroyed an important part of GNOME 2's habitat.
Controlling the desktop from the file manager
Want to adjust how icons are displayed in GNOME 2? How the Trash operates? How your system handles external media?
In all these cases, you don't go to the System menu in the panel or open the right-click menu on the desktop. Instead, you open the Nautilus file manager and select Edit -> Preferences. Not only does this arrangement tie you to a brain-dead file manager, but it could hardly be more unintuitive if the designers were trying to be unhelpful.
Exactly why the arrangement made sense to developers in the early years of the millennium has never been clear to me. However, mercifully, most other desktop environments have long since left this setup in its grave.
Useless Default Icons
Most installations of GNOME 2 come with at least two default icons: Computer and Home. These defaults are not removable from their right-click menus, which, for most users, means they can't be deleted at all.
That wouldn't be so bad, except that neither is needed. The Computer icon duplicates part of the contents of the Places menu, while the Home icon is unnecessary because the file browser opens in your Home directory anyway.
Both icons could be replaced by the file browser in Applications -> System Tool, which is a modified version of Nautilus for administration -- to say nothing of the only version worth using in the first place.
Many people complain about the design assumptions in GNOME 3. However, those assumptions first sprang up in GNOME 2.
You may remember Linus Torvald's famous complaint seven years ago:
This "users are idiots, and are confused by functionality" mentality of Gnome is a disease. If you think your users are idiots, only idiots will use it. I don't use Gnome, because in striving to be simple, it has long since reached the point where it simply doesn't do what I need it to do.
Torvalds was complaining that, in the effort to provide streamlined, polished applications, GNOME and many of its core applications were removing functionality from the graphical interface.
Torvalds himself has flip-flopped on his choice of desktop environments several times since making this remark. However, if you compare almost any GNOME core application with its KDE equivalent, his remark is -- if anything -- even truer than when he made it.
Whether you're comparing GNOME's Banshee music player to KDE's Amarok, Brasero to K3B, or Shotwell to digiKam, the tendency is unmistakable. On the one hand, in all of the GNOME applications, at best beginner to intermediate functionality is supported. On the other hand, in all the KDE applications, functionality exists for every level of user, including the advanced.
This tendency might appeal to users who are first learning an application. However, it can quickly become frustrating as users want to do increasingly more complex tasks or if problems emerge. Like zombies trying to cope with the modern world, in such cases, applications designed for GNOME 2 often lack the necessary capacity.
As Dead as John Cleese's Parrot
Such shortcomings tend to be overlooked when people look back at GNOME 2. For many users, GNOME 2 has the advantage of familiarity, and they have long ago learned to overlook such shortcomings -- especially when viewing it through a lens of nostalgia.
Yet the fact remains: GNOME project developers were not merely justifying their wish to do something new when they began working on GNOME 3. GNOME 2 was approaching the end of its natural life span, and keeping it on life support -- let alone keeping it healthy -- would have required more and more work with every passing year.
Increasingly, many parts of GNOME 2 would have had to be removed and replaced with ideas transplanted from other sources.
The only trouble is, developing GNOME 3 has proved as difficult as keeping GNOME 2 on life support. GNOME 3 exaggerates the design philosophy that began in GNOME 2, and introduces questionable assumptions of its own.
Under the circumstances, reacting to GNOME 3 by retreating to GNOME 2 is understandable. However, whether it is advisable is another matter. As Linux Mint's Cinnamon has already started to show, GNOME 2 requires extensive surgery after a year of neglect, and the effort is not trivial.
In the long run, the wiser strategy may have been to let GNOME 2 rot in peace.