Linux: Why “The People’s Operating System” can never compete as a desktop OS
Choice is good…..chaos is something else
Paradoxically, one of the greatest strengths of Linux – particularly as it endeavors to gain traction as a user desktop – is also one of it’s greatest weakenesses. One of the primary cases made for Linux is that it offers a dizzing array of choices (along with the fact that it’s open source and free, of course). “Choice is a good thing” is the mantra often chanted by those wishing to describe, and, spread the good news of Linux. The fact of the matter, however, is that Linux offers a lot of choices primarily because it comes in so many variants. Some of these variants, or distributions (called distros, for short) look and operate similarly; some, however, many bear little resemblance to one another in either form, or, function. The look, from the desktop, and the basic operation of these distributions often share little in common. And while it’s nice to have choices, having such significant variations can lead to a host of development and marketing problems.
If one wishes to make an argument for the benefits of Linux simple platitudes, like “choice is a good thing”, ignore and obscure certain realities about the challenges faced by Linux, particularly as it tries to gain market share as a desktop operating system. In the lexicon of the X-Generation, too many Linux fans fail to “keep it real” when discussing the benefits, and likely future prospects, of their favorite operating system. If Linux had many of the benefits of Windows or the Mac OS – but, additionally, offered infinitely more choices – everyone on the planet would be using Linux. There would be plenty of high-end software available for Linux (for doing things like editing audio, and, video). At the moment, there is a dearth of such high-end software for Linux, and, the prospects for the development of such software seem bleak.
Other operating systems may, or may not, offer choices. But when they do, it’s a qualitatively different matter.
For instance, Microsoft Windows offer choices. A Windows operating system, for example, may be released in “Basic”, “Home”, and “Premium” editions. Actually, there are usually six or seven iterations, the differences in which (to a very large exent), are lost upon average computer users. And the differences are lost on most people because the versions look, and, function almost identically. Programs are installed by double-clicking a file called Setup.exe, updates are installed in the same way from one version to the next, and programs that will run on one version will run just as well on the next. The desktops may look a bit different from one version to the next, but it’s usually just a matter of an optional doo-dad that’s either missing, or, included. In some cases there are program features that are available in one edition, and, not available in another. For example, Windows XP Pro (going way back) ships with Microsoft’s IIS Web Server as an optional package (it’s on the CD), whereas it’s not included in XP Home. So you’ve got some choices, but they’re largely inconsequential and moving between these “variants” generally involves a zero learning curve for the user.
The Mac OS takes a completely different approach. They periodically refine their operating system and release it for distribution – and every copy is identical. Driven by Steve Job’s Buddhist like vision of a “perfect whole”, The Mac view is that every user should should share a unified, desktop “gestalt”. If you’ve used one Mac, you can use another – the operating system is always the same.
Linux choices – too often, it’s like “Sophie’s Choice”
So, on one hand, it’s fair to say that “Linux offers more choices” – but that’s a declaration which, by itself, is misleading. Much like a political sound byte, taken out of context, it’s not the whole story. Attempts to summarize (and glorify) Linux with this blanket statement constitute an incomplete description of Linux and it’s benefits – or lack, thereof.
The fact of the matter is that this cornucopia of choices exists, largely, as an unintended consequence of the wild variations that exist from one Linux distribution to the next. Indeed, while Linux offer choices, it also demands choices that are, often, uncomfortable for the user. By choosing a Linux distro, too often a person feels that they are being forced to settle – to settle on a set of features they like in one distribution, while missing features they like equally well in another distro. It’s choice – but it’s a lot like “Sophie’s Choice”. These are the sorts of choices that induce angst. Being forced to make certain choices, where Linux is concerned, can actually be quite unpleasant. That’s the reality of the situation.
The first choice: which distro to install?
For anyone inclined to experiment with Linux, for the first time, simply choosing a distribution could be something of a a headache. Doing a bit of research, I found that the number of known Linux distributions is around 600. In reality, however, most lists of this kind include distros which are no longer active. Realistically, there are probably around 40 or 50 popular distros, and, about a dozen which are very popular as desktop systems. . Far and away, the most popular desktop Linux (as of 2012) is Ubuntu Linux. Ubuntu Linux is based upon a more traditional Linux distribution (not really aimed at desktop users) called Debian Linux. Another extremely popular distro – again, derived from Debian (and Ubuntu) is Linux Mint. There are lots more, too – reviewing, or even listing all of the available choices, is beyond the scope of this article.
And while some users have settled on one of the above distros as their “permanent” and only desktop OS (or another populuar distro) I’ve found that the majority of Linux users tend to bounce frenetically from one distro to the next, always trying to find a distribution that combines all of the features they need. Linux users, it seems, spend much of their time hunting for that perfect desktop distribution – addicted, it seems, to choosing…..again.
Picked a distro? The choices – and headaches – don’t stop here
Let’s say you’ve moved off of square one, and, you’ve chosen to install a particular Linux distro. If your coming from Windows or a Mac your going to find, immediately, that there’s far less software available (particularly of the high end variety – eg. Photoshop) for Linux. Making matters worse is the fact that, of the available software packages, many will work fine on one distro but won’t work (at all) on others.
Though some popular Windows and Mac programs have been modified – or “ported”, in the language of programmers – to work on Linux the number is extremely small. And since various Linux distros utilize different core packages for creating desktop graphics (and more) a program may have been ported to work on Ubuntu, but, not on Fedora. In other words, the mere fact that a program has been ported for Linux does not mean that it will neccessarily run on the version which you happen to choose.
Let’s talk about another choice. Linux will allow you to select from a variety of desktop environments – they all look different, feel different, offer different functions, etc. More choices – wonderful, right? Well, yes and no: you’ll find that a significant number of software packages which work with one desktop environment won’t work on another (for instance, what works on the Gnome desktop may not work on the KDE desktop). With a comparatively smaller selection of software out of the shoot, you’ll actually narrow that selection even further based upon your choice of desktops.
Is it just me – or does it seem like our real choices are dwindling, here?
What about installing software? In defense of Linux, modern versions generally allow you to simply double-click to perform a software install. Some things, however, still have to be installed from a command line (this, too, varys considerably from distro to distro). Uninstalling packages can be easy, or again, fairly involved – and the process differs from distro to distro, and from software package to software package. It’s never as simple as using the Windows Control Panel to simply choose “uninstall this program”.
Simply getting the software can be something of a pain in the neck, as well, and – you guessed it: the steps involved vary from distro to distro. In almost every Linux distro there are distro-specific steps which must be taken in order to download and install the most popular software (adding something called “repos” in Redhat-style systems, editing a file a called sources.list in Debian-style systems. etc.). It’s not as if these things are incredibly difficult to learn, but, a lot of regular computer users simply don’t want to fiddle with this sort of thing. And, if they’re experimenting, non-geeks will be even more repelled when they find that what they’re learning does not apply to other distros.
Finally, even in 2012 (almost 2013) there are still some fundamental issues with the Linux core architecture that really can drive a desktop Linux user to distraction. Foremost among those problems has been a persistent issue with audio. Sufficing a technical treatise on how Linux handles audio, let’s just say this: the audio setup, under the hood, is fairly complex and the results are often unpredictable. Frequently, you can’t listen to an mp3, and, simultaneously watch a YouTube video without something crashing. What if you’d like to capture some audio from your computer, recording the sound you here through your speakers? You can download the open-source program Audacity and try it – but I’ve never had it work with any of my sound cards. And, on an editorial note, though it’s frequently touted as one of the premiere open source package available I really don’t like Audacity. The interface is clunky, it’s difficult to configure and operate, and – for my money – in form and function it’s light years behind my ancient version of the old Windows program “Cool Edit Pro” (circa 1999 – something from a previous century).
Linux: on the positive side
To be clear, Linux makes an incredible server platform – and it’s completely free. Further, there are actually some great open source programs that will install on almost any version, most notable among them are Open Office and The Gimp. Too, most Linux distros are easily installed (via a graphical installer) and the OS does a remarkably good job of installing the correct drivers for most common devices. The Linux filesystem does not become fragmented (in any meaningful sense), and, though there are viruses that effect Linux you’ll probably hit the Power Ball Jackpot before you get one (I’ve used Linux for eleven years and never had a virus).
I’m actually crazy about Linux. I’ve used it a server for over a decade, and, I tinker with it as a desktop (off an on). But, contrary to ravings of some Linux fanatics, it’s far from becoming a serious contender in the desktop computer market. And that’s not the idle musings of some old guy at his computer – I’ve got data to back it up……
The Linux DESKTOP EXPLOSION : a 1% market share – and growing!
If you do a Google search for “market share linux”, I promise, you’ll find that the first five pages of results (or more) are pages that suggest that Linux is making terrific gains in the desktop computer market. Yet, if you look at the actual statistics it would seem that Google needs to tweak that darned algorithm, one more time. Or maybe those search results are due to the fervor, and, frequency with which fans tend to inflate the successes of Linux via WordPress. At any rate, research (from a variety of sources, as of 2012) indicates that Linux currently enjoys a mere 5% of the operating system market, totally, and only a 1% share of the desktop market (the other 4% represents Linux servers).
A rare failure of the open source model: why desktop Linux will remain marginilized
Like a lot of people, I’ve always been crazy about Linux. As a completely free operating system, built on an open source model (anyone is free to modify the code) it has a Bohemian – even Utopian – kind of allure. Linux has always felt like a wonderful, benevolent, incredibly successful socialist experiment. Had it been born in China, and not Finland, it might have been called “The People’s Operating System”. With it’s exceptional stability and low-overhead it reigns as, arguably, the best server platform on the planet. Ironically, however, the same things which have made Linux such a huge success as server software are things which work equally hard against Linux as it trys to become a mainstream desktop operating system. The open source model – as opposed to a standard business model – is absolutely doomed to failure in the desktop marketplace.
While I’ve touched upon some of these points, above, here’s an overview of why desktop Linux represents a rare failure of the open source model – and why desktop Linux will never attract more than a tiny fraction of desktop users:
- Because the Linux code is open source and can be modified by anyone (wihout violating any copyright laws) programmers from around the world are free – and in fact encouraged – to find and fix problems with it’s core architecture. Someone in Zaire may find an issue with the kernel and develop a patch. They share it with community and everyone benefits. Someone in Belgium may make an improvement in the way in which Linux handles certain types of secure connections. Again, they share their work with the Linux community and everyone benefits. This open source tradition of finding, and, sharing knowledge is what paved the way for the phenomenal growth of Linux.And, though originally a command-line system modeled on Unix, people are also free to extend Linux by developing slick looking desktop versions – complete with great graphics, audio components, and everything else offered by something like Mac Or Windows. Linux desktop development, however, differs fundamentally from Linux core development.
Whereas core development focuses on issues of operating system stability, speed, security, etc. desktop development has far more to do with fashion, taste, and aesthetics. It’s a creative process in which the developer may, or may not, model a distro strongly along the lines of an existing distribution. Some desktop developers simply tweak existing distros, whereas some write a considerable amount of code to create a strikingly original version. Scattered around the globe these developers may work independently, or, in small collaboratives. But, what works so well for Linux core development completely thwarts any chances that Linux might ever gain widespread use as a desktop operating system.
- The problem is obvious: with so many people working independently (or in small groups) desktop Linux distributions have evolved absent the sort of standards that people have come to expect in a desktop operating system- the sort of standards provided in a business model. As you move from one desktop Linux distribution to the next, you often find that they contain disparate and conflicting software packages for performing identical functions. It may be the audio structure, or the the core libraries used to produce the onscreen graphics (or a combination of these things, and more). As mentioned earlier, this means that what runs on one distro will often not run on another – and this is a state of affairs that neither a Windows nor Mac user would ever understand.
In fact when considering Linux as a desktop, in this sort of conversation, it’s difficult to really even define desktop Linux. With so many different distributions, varying wildly in their forms and functions, how would desktop Linux ever be marketed? Perhaps it’s why you often here people say, “I’m an Ubuntu user” as opposed to “I’m a Linux user”. Desktop Linux, really, still lacks a clear identity.Perversely, the open source model for which I have so much admiration amounts to something of a birth defect in the case of desktop Linux. In short, the lack of standardization, and the extraordinary fragmentation of desktop Linux amount to a stake in the heart for any hope that it could ever be a serious contender in the desktop marketplace.
- Linux is free, and free is hard to beat. We all like free. Free, however, doesn’t work very well when it comes to developing a robust, reliable, standardized desktop operating system – the sort of thing which can compete with Windows, and, Mac OS. And “free” works equally poorly for developing the high-end software that users normally find available for a desktop operating system. The Polyanna view that, somehow, developers who are working for free (or who are relying on donations) will produce an operating system on a par with Mac OS – or an open source rival to Adobe After Effects – is a delusion shared my many in the Linux community. It just ain’t so.
A case in point: As I’ve mentioned, the Gimp (modeled on Photoshop) and Open Office really consititute the only high-end software packages which have been standardized to run on almost any version of Linux (and Unix). One similarly ambitious project has been undertaken by programmer Paul Davis. Working largely on his own, Davis has been developing Ardour. From Wikipedia” “Ardour’s intention is to provide digital audio workstation software suitable for professional use.” Essentially Davis has set out to produce a Linux alternative to the popular, and fairly pricey, Pro Tools package available for Windows and Mac. And Ardoir has received rave reviews from all corners. But there’s a problem. Davis’ original plan was to offer the program to the Linux community for free, counting on donations to support the project. But while the download links got clicked thousand and thousands of times, it was as if the “donations” button was invisible. For all of his work, Davis reported that in 2009 donations topped out at around $25,000. As a top-flight programmer developing a wildy popular, complex software package for Linux he was making about the same money he would have made working full-time at McDonald’s. The Ardoir episode amounts to a reality check for anyone believing that an open source model will ever produce an operating system, and associated software, that can compete with retail counterparts. And, it amounted to a serious reality check for Paul Davis: while the project remains active, as of 2012, Ardoir can no longer be freely downloaded.
Davis learned what Adobe (and Redhat) have known for a long time: Linux desktop users are a lot like crime – they simply don’t pay. It’s been one more epic failure of the open source model when applied to Linux as a desktop operating system.
A final case in point: A few years ago, Redhat – a dominant force in Linux from very early on – announced that they were getting out of the desktop Linux market, for good, and adopted a business model focused around their Redhat Server line. While they remain committed to helping develop the Linux desktop through Fedora the move was a tacit acknowledgement that Linux, as a desktop, was a losing proposition. It appears to have been a smart move; through the sale of licensing fees, and, support services Redhat quickly became a billion dollar corporation.
People got out of bed today, all over the world, and used desktop Linux to surf the web, do their business, and chat with their friends – and, I’m sure that they enjoyed their Linux experience. And I’m certain, that 50 years from now, computer users all over the world will still get out of bed and use desktop Linux to do the same……about 1% of computer users, world wide.
Guy Merritt is a contributing
editor at DIYTechTools.com