|Main||In Other News||Security Roundup||Tips of the Trade|
"Each time I get a new version of GNOME, there's this feeling of anticipation and exhilaration a feeling that this new version of GNOME can't possibly turn out to be as bad as the last one," fumed Linux enthusiast Nicholas Petreley, "But so far, each new version lives down to the same low standards set by the previous one."
Mr. Petreley was using the occasion of the recent release of GNOME 2.6 to spend some time bashing the desktop, which is, for all intents and purposes, the heir apparent to the Unix world's workstations for reasons we've delved into before.
Any other year, the whole matter might be a question of an enthusiast overfocusing on minutia. With GNOME's near ubiquity, and its lead spot on desktop offerings from Sun and Red Hat, potential problems with the project bear some examination. To consider a Linux desktop deployment is, with most credible enterprise offerings available today, to consider a GNOME deployment.
We'll admit to approaching each new GNOME release with some trepidation, too. Once upon a time, GNOME offered a veritable cockpit of configurable options for users to confront, allowing them to massage the environment into a custom fit all the way down to "When we drag a browser window around, we'd like to see its contents, but when we drag a word processor window around, we'd like it to present an empty frame, unless we're resizing it, in which case we want to see what's in it." Sounds like a horrible way to spend your day, clicking on all those little boxes and options. But we were younger back then, and time seemed less valuable.
Partially out of humane concern for frightened newbies and partially because people at all the best (read: nerd) parties knew it was sexy to talk about "usability" in knowing tones, the GNOME project toned down some of the flexibility and even got some money from Sun to conduct a true usability study, which recommended the people programming GNOME quit acting like they were their own audience. At about the same time, members of the original Macintosh design team started building the Nautilus file manager, a key component of the GNOME desktop, which did a lot to get GNOME developers talking about usability.
Some of the simplifications made in the name of usability have infuriated people, like Mr. Petreley, who have heard the word "usability" abused enough at all the best (again, read: nerd) parties that they've come to associate it with moronizing and dumbing down not what the word means when someone who's made a profession of it uses it. In this context, it is more like "refinement" or even "being considerate of your users." It doesn't help that discussions on these refinements have been conducted in the typically abrasive and dismissive tone developers tend to use when they're separated from end users by a pair of mail clients and the entire Internet.
So is GNOME 2.6 a moronized monstrosity fit only for small children and simpletons? Or, is it a reflection of several years of usability refinement? If you're an IT decision maker (say that a few times ... we get chills every time), is it something you'll want to inflict on your users when it turns up in Sun's Java Desktop System or Red Hat's Desktop?
We've spent a lot of time experimenting with GNOME 2.6 during the past few weeks, and we're inclined to say it's no worse than anything else we've dealt with in recent years. Early OS X releases constituted the hoodwinking of an entire user community that had no idea it was paying for the privilege of running two years of beta software. It took a veritable bucket brigade of third-party software developers to let us stand the sight of a pulsing blue button. We've been similarly troubled by Windows XP and its obvious anxiety over OS X, and we've looked at the latest from the KDE project, which provides a cluttered riot of over-configurability.
There's an obvious opinion present in pieces of GNOME. The feature that infuriated Mr. Petreley was the project's so-called "spatial" interface, which eschews a hierarchical navigation approach by reusing the same file manager window and remembering the size, shape, and location of the window the next time it's opened. The GNOME developers have a nasty habit of declaring certain preferences (like hierarchical navigation) anathema and banishing them from the control panel. They infuriate long-standing users in the process because the so-called refinements are aimed at attracting users they don't yet have, rather than acknowledging the ones they do.
Since we're wearing our "IT decision maker" hats, the question we have to ask is: What will inflicting an interface on our users do to their productivity? In the case of GNOME 2.6, we've got to say "nothing disastrous." If they're moving from Windows or Macintosh machines, they'll have to deal with interface shear no matter what. Nothing GNOME does precludes corporate users from surfing the Web, reading their mail, editing spreadsheets, and working on a word processing files. There are decent touches for mounting network drives, dealing with potentially malicious files with extensions that don't match their actual content, and "just working" when a wide variety of file types are encountered. GNOME mostly, in fact, just works.
That might sound like damning with faint praise, but from our perspective there are much bigger issues than faintly religious-sounding scuffles over the relative usability of desktops and whether they let you configure this, tweak that, or conform to some cherished notion of what's "more usable." How those desktops interact with hardware, the rest of the network, and legacy file formats are the real issues. What we found was that when a system is properly configured underneath (something the IT team should work on before rolling the machines out), GNOME handles that stuff just fine (and so does KDE, for that matter). In fact, in our completely unscientific usability study, it took our subjects less than 10 seconds to locate the Solitaire game. We're not sure what else the corporate desktop needs.