What does the Windows 8 Developer Preview have to do with the Linux desktop? Not much, you might think at first, especially since the final version is likely to be vastly different than the preview.
But as I explored the preview, I couldn't help asking myself: is this the first hint of how major desktops will look in the future on all operating systems? It's a distinct and -- for a Linux desktop user -- an alarming possibility.
If you haven't already downloaded the Windows 8 Developer Preview, then Jason Perlow's informal video will give you the general idea (as well as the key to installing it virtually: use the newly released VMWare Workstation 8).
The preview is really two desktops in one. One is a standard Windows desktop that differs little from Windows 7. The other is the default Metro API, which resembles the Windows Phone.
Scrolling horizontally, the Metro desktop consists of clusters of widgets, with square ones for applications, and rectangular ones for utilities and online services. These widgets can be moved around, but a garish desktop is nearly inevitable, partly because of the two different sizes, but also because, so far, the widgets can't be themed. Metro apps open full-screen, and intentionally do not support plug-ins -- for those, you need to use the legacy interface.
Verdicts about the Metro desktop would be premature, because in the preview it barely exists. Yet the trends could be summarized as more of the same. Metro is a desktop where users can do even less than ever before without drilling down into the remote depths of the interface, and can do little easily except launch applications. In Metro, Microsoft appears to have labored mightily, only to produce a glossy modern version of Windows 3.1.
It Can't Happen Here - Or Could It?
One of my first thoughts while looking at Metro was that I've been too hard on GNOME and Unity in the past few months. For all my misgivings about the directions of these two desktops, their current versions are far preferable to the Metro preview.
If GNOME and Unity sometimes make me feel that I'm constrained and distanced from the tools I use, then Metro makes me feel that I am locked into my chair with steel bands and forced to type with down-filled mittens.
But, thinking more carefully, I realized that the difference is largely a matter of degree. Right now, it's an extremely large degree, but what's worrying is that it isn't a difference in kind. If Metro is successful, then the difference in degree might easily diminish, or even vanish altogether.
Consider: the free desktop has always looked to proprietary rivals for design inspiration. KDE's Kicker menu looks mightily like recent Windows menus. Similarly, Unity's imitations of OS X have become the source of endless jokes -- jokes that have a sting in them because they contain a grain or two of truth.
Sometimes, such borrowings can be justified because they make sense. The arrangement of system settings into categories, for example, is more efficient for users than a long list of options, and never mind that Windows had it first.
At times, however, the justification is less obvious. KDE's default placement of the menu in the lower left corner, for instance, is usually justified on the grounds that Windows users expect to find it there, rather than as an efficient arrangement.
Too often, I suspect, developers are simply unable to shake the now-obsolete conviction that the Linux desktop lags behind Windows and OS X, and needs to emulate them. Examples of free desktop developers trying to innovate in their own right, as GNOME 3 or the KDE 4 series does, are still relatively rare. And even these examples are not completely free of a sense that proprietary desktops need to be imitated.
The possibility that Metro could be imitated on the free desktop seems all the more likely because many developers already subscribe to the same design principles as Metro. Development of GNOME 3 and Unity in particular already seem driven by a drive towards simplicity that users don't seem to want, apparently on the grounds that what might be suitable for a novice must also be suitable for intermediate and advanced users.
Even more importantly, most free desktop developers seem to have decided -- apparently independently of each other -- that mobile devices should be the model for all computing desktops.