Tuesday, February 03, 2009

Why Linux won't challenge Windows on the Desktop (yet)

Every year begins with a renewed optimism (on the part of the Linux community) that this will be the year Linux takes a bite out of Windows on the desktop.  While distros like Ubuntu and Mint, and the latest KDE and gnome UI's have significantly improved, they all suffer from the same flaw - they are all (directly or indirectly) designed to look like windows.  What's worse, they are all stuck in Win9x/NT4.0 days.  Sure, in the last few years the community has added visual effects that in some cases outshine (and definitely outperform) Vista Aero.  However, at the end of the user experience is Windows-like.  In fact, in some instances it is worse than Windows.  The number of places you have to set proxy settings, the agony that is wireless networking, and the sheer number of operations that can only be accomplished by dropping into a terminal make Linux less user-friendly than Windows.

The key issue is that Linux is an operating system written by very technical people, and for the most part, for very technical people.  I can appreciate that (and actually run several distro's on spare systems and VMs).  However, I really don't want to have to modify an interfaces file in order to connect to a wireless network using WEP with a default key other than 1.  I would love to get through a day without using sudo.  And I don't want to have to use a cryptic command line to find out if my USB key is mounted as sda or sdb.

Given the paradigm similarities, it's going to be hard for a fragmented Linux community to compete.  Especially when Windows is continually refined.  I've posted before about how Microsoft is stuck in incrementalism.  This was partially a criticism, but partly it is just the nature of Microsoft being the victims of their own success.  When over 90% of the world uses your software you can't just pull the rug out from under them and change the paradigm.  But even Microsoft has recognized the need for change. 

Widespread use of cell phones and smart phones has driven adoption of task-based operating systems.  Other small form factor devices such as MIDs and netbooks have provided a logical beachhead for task-based user experiences on PCs.  Microsoft has dabbled in these waters with Origami, and do-it-yourselfers have done it by putting Android on the HP 2133 Mini.  A task-based UX on PC's represent an opportunity for Linux like never before.  However, the community needs to let the engineers step away and give the designers some time to play.  HP's Ubuntu-based Mi experience on the Mini 1000 is one example of this direction.  Jolicloud - a much less publicized distro aimed at netbooks - is another possibility.  The key to a task-based UX is that the fundamentals - wireless networking, peripheral devices, etc. - have to JUST WORK.  No terminals, no complicated command lines, no administrator or root privileges required.

Time will tell if task-based will take hold with netbooks.  However, at the very least it should be a hint that the Linux community should stop trying to out-Windows Windows.  If anything, Linux should take a hint from Apple and position away from Windows instead of positioning better than Windows.  More and more people are valuing computers based on how they use them, not speeds and feeds and the granularity of control they have to configure them.  Phones have further encourage task-based interaction with devices.  Finally, the pervasiveness of the Internet and the rise in cloud computing have blurred the lines of defining a computing device.  Linux needs to embrace these trends  and change the game.


No comments:

Post a Comment