The Ubuntu story is fascinating. It was created by Mark Shuttleworth, a 30-something South African entrepreneur who made his fortune in the 1990s, flew to space in 2002, and then decided to knock Microsoft off its perch as the world’s leading desktop OS. (AKA fixing bug #1) In just two years, Ubuntu has become widely recognized as the desktop version of Linux to beat.
Of course, a big part of its success is the $10 million a year he’s reportedly sinking into Ubuntu’s parent company, canonical. Still, there’s no way you could build a full-featured OS from scratch in two years with $20 million–to say nothing of a desktop OS with hundreds of applications and support for multiple architectures. Ubuntu is a very thin layer of commercially-developed (but free-as-in-speech) enhancements to off-the-shelf free software. Most importantly, Ubuntu is built atop Debian, a Linux distribution that focuses on stability and using exclusively free software.
Below the cut I’ll give some of my initial impressions of the OS, which necessarily will be a little bit more technical than the usual TLF fare. I’ll consider some of the economic implications of Ubuntu in a future post.
I last played around with Linux in 2002, when I installed Debian on an old Power Mac 7300. The improvements since then are striking. In 2002, you still had to boot into a text-based installer that made you do things like partitioning your hard drive manually. A lot of little things didn’t work right, and in many instances, you were required to troll newsgroups and then pop open a command line to edit text files in order to get stuff to work right.
In contrast, the Ubuntu installation process was dead simple: download a CD image, burn the image to a CD-R, boot the computer from the CD, double click the “install” icon, and then answer about 5 questions on non-technical subjects like your preferred language and time zone. (I did have to edit the X config file at one point to get the GUI to come up properly, but I think that’s due to my having a 6-year-old machine running an obscure CPU) The installer automatically partitions the hard drive and copies the necessary files to it. There were lots of bells and whistles that weren’t there on my previous experience: sound worked flawlessly, there were attractive splash screens and reasonably helpful status messages, and when I was in doubt I could consistently choose the default and it would work just fine.
Things get even better once you boot into your freshly installed system. A nice graphical software-update tool pops up offering to upgrade your software to the latest version. You click “OK” and it does its job in the background without having to do anything on the command line. Useful utilities like OpenOffice and GAIM are installed by default and work flawlessly. There are a dozen games pre-installed, and a variety of helpful tools like calculators, dictionaries, and an address book.
The desktop environment was a program called Gnome, which I continue to find clunky. It seems determined to take over as much of my screen as possible, with a menu bar at the top of the screen, a panel at the bottom, and a menu bar in each window. The windows themselves also seem to take up a lot of unnecessary space, with a tall title bar and wide borders around each window. So I converted to KDE–a process that did require me to open up a command line, but only required typing one command–and am much happier with its look and feel. Most importantly, it allows me to put application menu bars at the top of the screen where they belong. The KDE preference panel also strikes me as vastly superior to Gnome’s alternative, although maybe that’s because of its similarity to the Mac OS X preference panel.
Neither Gnome or KDE are up to the level of fit and finish of Mac OS X, though. My biggest pet peeve with X (the Windowing system that serves as the foundation for Gnome and KDE) is its idiotic use of text selection for “copy” and pressing the middle button for “paste.” Selecting text and cutting-and-pasting are distinct operations. I often want to select text for deletion purposes without overwriting the contents of the clipboard in the process. Although most applications now seem to support proper copy and past commands, there are enough old style apps around to make it irritating to use.
If I were to switch to Ubuntu as my full-time OS, I would sorely miss other Mac OS innovations: notably Expose, Apple’s insanely great window-switching utility. And I’ve heard (although I haven’t tested this) that Ubuntu still falls well short on the laptop side, with Mac OS being superior on rapid sleep and wake (something Windows doesn’t do right either) and on rapidly and transparently finding and logging on to wireless networks.
But those are details that can doubtless be ironed out in future posts. It seems to me (based on a few hours of tinkering) that the biggest obstacle now facing Ubuntu is that it has to swim against the tide of the Windows and Mac OS network effects. In many cases, the features that Ubuntu lacks are features that are provided for free to other companies. For example, there doesn’t appear to be a user-friendly Flash player for PowerPC-based Linux machines. And Google Maps doesn’t seem to work with Konquerer, the default browser of KDE (it works fine with Firefox, however). These aren’t a case of Linux users being too cheap to pay for needed software: Adobe and Google give these products away for free to the users of other platforms. It simply reflects the fact that there are not yet enough Linux PowerPC and Konquerer users, respectively, to provide support.
By the same token, there doesn’t appear to be a convenient way to access the iTunes Music store or load my MP3s on my iPod. This is a deal-breaker for me. But again, Apple give iTunes away for free to Windows users. If there were more Ubuntu users, doubtless Apple would create a version for them too.
Of course, there are two ways that these problems could be solved. One would be for Apple, Google, Adobe, and company to produce proprietary versions of their software for Linux. In some cases, however, the solution is even simpler: just publish the specs and give the Linux community permission to create their own open-source clones of the software. In many cases, there are open source developers who would be happy to do the work themselves if they could just get ahold of the necessary information.
Both of those problems would begin to ease if Ubuntu began to take market share away from Windows. Linux support would no longer be a luxury to satisfy a few vocal geeks: it would be essential to serve a growing part of your customer base. The Linux community, then, faces a kind of chicken and egg problem: users won’t adopt the platform until third parties start supporting it, and third parties won’t spend the resources to support it until it has a significant number of users.
If they ever do solve the chicken-and-egg problem, though, its growth could be explosive. The promise of hundreds of free applications, no viruses, and free software updates for life should be a big draw for users. More importantly, as hardware costs continue to decline, the Microsoft Tax will be a larger and larger fraction of the cost of a PC, giving OEMs a substantial incentive to migrate their customers to a free alternative.