I'm not a fan of the GNOME 2 release series. It was my main desktop for years until I replaced it with KDE 4, which was far more innovative. A few months ago, I wrote pointing out some shortcomingsof GNOME 2, questioning the demand for what seemed to me like a desktop that had long outlived its usefulness.
GNOME 2 is not likely to become my main desktop again, but I realized that there were nine reasons (if not more) that users might prefer a GNOME 2 derivative over more recent desktops. Many of these reasons are not a specific software feature, but higher level design approaches, or even things completely external to the desktop and its performance.
GNOME 2.0 recently passed its tenth anniversary. If you compare it with GNOME 2.32, the last release in the series, the improvement in widgets, utilities, and everything else is obvious. However, what didn't change in the entire series was the classic desktop layout with a configurable panel, a menu, and launchers on the desktop.
This layout makes GNOME 2 clones like Mate instantly familiar to both anyone who has used a Windows release in the past fifteen years or who used a free desktop like Xfce or LXDE. Users might take a while to learn the details, but they immediately understand the general navigation.
That doesn't mean that GNOME 2 is intuitive, or even well designed if looked at impartially. But it does mean that users can quickly stop focusing on the desktop, and turn their attention to their tasks, where it belongs.
2. A Customizable Panel
Once, a panel that you could easily position or add icons and widgets to seemed a given. Then the KDE 4.0 series premiered with a panel with next to no customization options (a limitation long since removed), and the GNOME 3.0 releases did the same in the name of saving users from clutter.
Admittedly, some users leave the panel exactly as it comes. But, for others, a panel is a convenient place for useful utilities. Many depend on a panel so much that they have two. Mate, for instance, largely reserves the bottom panel for the task bar, which means that you can have over half a dozen windows open and still clearly read their names.
Personally, I wouldn't exaggerate if I said that a customizable panel is basic to the way that I work -- and I suspect I'm not the only one.
3. A Classic Menu
A classic menu, whose levels spill across the desktop is an ungainly mess. All you can really say about the classic menu is that it is better than the alternatives.
Unlike a menu confined to a single dialog window, a classic menu assures that you never lose your place in the menu hierarchy.
Unlike a menu on an entirely separate screen or overlay, like those offered by GNOME 3 or Unity, it's quick and doesn't distract your attention from what you are doing, and rarely buries administration items more than three clicks from the desktop.
4. Windows Open At Medium Size
Presumably inspired by the interfaces for mobile devices, both GNOME 3 and Unity open most apps maximized. Exceptions are only made when an app's window is so small that opening maximized would be ridiculous.
In comparison, GNOME 2 and its clones open apps in a medium-sized window. This is usually the most sensible option, since you don't always open an app to use it exclusively or even immediately. Often, you open it to use together with another window. Opening it maximized distracts you as you scramble to rearrange the open windows so you can use two or more together.
This rearrangement -- not incidentally -- is all the harder when the tools for adjusting windows’ sizes are either not close at hand or invisible until the cursor rests on the title bar. Title bar buttons may seem like clutter to the developers of GNOME 2 and Unity, but they are an accessible positioning of basic tools for anyone who works with more than one window open at a time.
5. Control of Virtual Workspaces
GNOME 3 arranges virtual workspaces for users. On the whole, it manages well, and should be credited for introducing users to the concept.
However, like any tool that adds a level of complexity to basic desktop functionality, virtual workspaces need to be under a user's control. For one thing, some users will always want to avoid the added complexity. For another, users want to organize their workspaces by naming them and, in some cases, like terminals, have them persist between desktop sessions.
Virtual workspaces in Mate and other GNOME 2 clones could do far more, such as allowing different icons sets on each virtual workspace. But any degree of control is preferable to what GNOME 3 offers.
6. Launchers on the Desktop
Many GNOME developers seem to regard launchers on the desktop as clutter. GNOME 3 does not permit them at all, while Unity easily allows templates and documents, but requires users to go through the file manager if they want to add application launchers.
Yet the desktop is the quickest place to access commonly used applications, placing them only one click or double-click away. By allowing app launchers on the desktop, GNOME 2 automatically appeals to a large section of users that other GNOME desktops do not.
7. No Catastrophic Failures
GNOME 2 and its successors sometimes crashed on me, requiring a restart of X Window to restore the desktop. But I can't remember a crash that permanently brought the desktop down and required laborious recovery, such as creating a new account and transferring personal files to it. I can't say the same of KDE, GNOME 3, or Unity.
Exactly why the GNOME 2 desktop is more robust is uncertain. Maybe with a simpler desktop, there's less that can go wrong. Or perhaps its code is more mature.
8. Lower Memory Requirements
In the past, GNOME 2 was sometimes dismissed as bloated, with its last releases realistically requiring half a gigabyte of RAM. Compare to a window manager or a desktop like Lubuntu, that figure seems high. But, compared today to KDE, GNOME 3 and Unity, all of which require a gigabyte of RAM for decent performance, it seems economical. Not only will it run better on older machines, but on modern machines it appears far more responsive.
9. Fewer Assumptions About Users
GNOME 2 was designed to accommodate users with different work habits. Users could make launchers on the panel or desktop part of their workflow, or open apps entirely from the menu. They could use virtual workspaces, or ignore them. Essentially, it could be customized for the user, rather than forcing users to work the way its developers expected.
For all its experimental features, KDE still has this flexibility. However, both GNOME 3 and Unity are built on the assumptions that there is one way to use them efficiently, and that users should adjust to the built-in expectations. If these assumptions fit the way you work, you are unlikely to have any problem with them. But the trouble is, many people find the assumptions awkward and distracting. Such people are likely to feel less restrained in GNOME 2.
A Possible Tenth Reason
Reading comments on various forums, I am almost tempted to add a tenth reason for using GNOME 2 clones. The design of GNOME 3 and the GNOME project's general ignoring of user complaints has apparently caused considerable anger -- far more than even KDE 4.0's release did. Unity has received similar complaints. To many GNOME users, these modern desktops continue to feel like a betrayal by developers.
To those who continue to feel this way, what better way to register your anger than to revert to an earlier release series, or to activate a group of extensions like Linux Mint's Cinnamon that converts GNOME 3 back to GNOME 2? But this reasoning, if it exists, is probably unconscious.
A GNOME 2 clone is never going to be the most innovative of desktops. To many eyes, it will look antiquated. Yet it gets the job -- any job -- done without any fuss, and respects users' choice, and that combination is rare enough in modern GNOME that solutions like Mate and Cinnamon are likely to have a wide appeal for years to come.