On Freedom’s Path: The hope of Linux gaining market share

People who promote the freedom of computer users have long dreamed of the day when a sizeable proportion of the world uses a free operating system on their desktop. The dream that ordinary people should control the technology in their hands, rather than letting it control them, has been oft-denied with many bumps and false starts along the way.

Nonetheless, HP’s recent announcement that it will install its Linux-based operating system, WebOS, as a secondary operating system on all its PCs offers a tiny measure of hope to those who struggle for computing freedom. In order to understand the significance of WebOS, first consider the recent history of Free, Libre, Open Source Software (FLOSS) on the desktop.

The dreary history of desktop Linux
While FLOSS advocates many enjoy running GNU/Linux or BSD on their PCs, they have had little success convincing the majority of computer users to run a free operating system on their desktops. While the major computer vendors have long offered servers with GNU/Linux preinstalled, very few offer a free operating system on their PCs for normal consumers. Advocates for freedom may like to note all the successful companies like Google and Yahoo! which run their infrastructure on free software, but running servers in the back room it does little to advance the cause of freedom for ordinary people–in fact many would argue that all those servers running GNU/Linux and BSD in the cloud are privatizing people’s data and stripping their freedoms online.

The real challenge is to get free software on people’s desktops. A number of computer vendors have flirted with GNU/Linux on their PCs from time to time. Back in July 2003, HP announced that it would be selling a $350 PC with Mandrake Linux 9.1 preinstalled, but few seemed to be able to obtain one. In November 2007, Walmart offered a $200 linux-based PC from Everex, which flew off the shelves in a week to never make another reappearance. After two years of promising to offer their premier Thinkpad line with SUSE Enterprise Linux Desktop, Lenovo finally began selling online two models of laptops with GNU/Linux in 2008 (although the built-in card reader wasn’t supported). The experiment only lasted 9 months before the GNU/Linux option mysteriously disappeared from their website. GNU/Linux fans muttered darkly about Microsoft’s monopolistic arm twisting to pressure vendors into dropping their favorite operating system.

The only major vendor with enough spunk to buck Microsoft’s strong-arming has been Dell which first offered Red Hat on a few consumer PC models between May 1999 and July 2001. After receiving hundreds of requests for Linux, Dell began offered Ubuntu on three of its PC lines in May 2007. It has always offered Ubuntu in some fashion since that time, but checking of its US web site reveals that it only retails one model of consumer desktop with Linux today. Nonetheless, few buy the PC with Ubuntu preinstalled, since there is no option to select a Linux operating system when customizing the purchase in the main Dell storefront online. Instead, shoppers have to know to go to http://www.dell.com/ubuntu and click on the “Shop Ubuntu” link to find a computer with Linux preinstalled. Obviously Microsoft has cowed Dell just like all the other major vendors.

There was a brief hope back in 2007, when ASUS first introduced their eee PC netbooks with Xandros, that maybe GNU/Linux would finally burst into the mainstream. In mid-2008, Gartner estimated that 4% of PCs were shipping with a Linux operating system. IDC market analyst, Richard Shim, opined that netbooks would offer Linux a chance to grow. “It’s an opportunity for Linux to kind of incubate,” Shim said. “They have an optimized chance vs. Windows on price.” Intel worked assiduously to promote the use of Linux on their Atom mobile processors. In July 2007, Intel released preliminary code for its version of Linux for netbooks, named Moblin (short for “Mobile Linux”). It was a FLOSS stack which could be easily adopted by any distribution, as Ubuntu, Mandriva and Linpus did. Acer planned to run its Aspire One netbooks on Moblin and Dell installed Ubuntu’s Moblin Remix on their Minis.

Nonetheless, the dream of widespread usage of Linux on the desktop quickly evaporated as Microsoft aggressively discounted the price of Windows for netbooks. Despite the fact that Microsoft was eager to move to Vista, Microsoft made a special exception for netbooks which didn’t have enough RAM to handle Microsoft’s latest memory-hogging OS. Microsoft slashed the price of Windows XP on the netbook to a rumored price of $25 and the price difference between Linux and Windows netbooks evaporated (probably helped by some backroom pressuring along the way). Windows came to dominate the netbook sector just like other types of PCs. Today, all of the major computer vendors including ASUS have dropped Linux on their netbooks in most parts of the world. Dell claims to still offer Linux on its Mini 10, but it is impossible to buy from their North American or European web stores at last check. Effectively, the dream of Linux breaking into the mainstream on low-end netbooks has died, although some hold out hope that ARM-based netbooks in the offing may revive that dream, since Microsoft Windows doesn’t yet support that architecture.

Microsoft’s monopoly on the desktop may be weakening slightly around the edges. Gartner projects that the world market share of Windows will drop from 95.2% in 2009 to 94.4% in 2014 in non-consumer usage and to 92.8% among the consumer installed base. Those few percentage points of loss in market share, however, are projected to picked up by Mac OS X. Advocates of freedom regard this as little gain, since Apple is arguably worse that Microsoft when respecting its users’ freedom, despite the fact that Apple operating systems are constructed on top of Darwin, which contains a free Mach 3 kernel and BSD command line tools.

Advances in mobile Linux-based operating systems
This dreary picture on the desktop, however, is contrasted by the recent advances of free operating systems on smartphones. Mobile devices were once the demesne of a plethora of proprietary systems, but Linux has proven so adaptable and cost-effective, that it has become increasingly difficult for proprietary systems to compete with the rapidly-evolving Linux. In recent years, Intel and the ARM manufacturers have worked closely with the Linux developers to incorporate drivers for their mobile processors into the Linux kernel.

Nokia, the largest producer of cell phones in the world, foresaw that its venerable Symbian operating system, which had 63.5% of the smartphone market share in 2007, was simply not flexible enough to compete at the high end of the market. Nokia developed Maemo, a Linux operating system based upon Debian which used the GTK+ toolkit for its applications. First released in November 2005 for its flagship 770 line of smartphones, Maemo was highly applauded in the technical press. The 770 and its successors were explicitly designed to be customized by the user and to place very few restrictions on its use. Nokia encouraged the installation of third party software found in the other GNU/Linux distributions and even placed many of them in its “extras” repository to be downloaded. The technorati delighted in Maemo’s ability to run many of the standard applications ported from Debian.

Although Linux enthusiasts snapped up the prized smartphones running Maemo, Nokia’s operating system did little to advance the cause of freedom in the mainstream. Maemo found its way into few hands, since Nokia only installed it on its expensive 770, N800, N810 and N900 models, while the vast majority of Nokia’s smartphones continued to use proprietary Symbian. Inside the feuding Nokia, Maemo was treated as a kind of interesting side-experiment, rather than the focus of its core business. Although the Maemo interface got better with each new release, Nokia never marketed them with the passion of Apple, whose iPhone took the market by store upon being released in January 2007.

Likewise, Nokia made the mistake of not trying to develop Maemo into a common platform like Symbian, whose development costs could be shared among many companies. Instead, Nokia allowed Google to steal its thunder with the development of Android.

Google, which wanted to guarantee its search revenues on the emerging mobile sector, decided to develop the Android and ChromeOS operating systems and the Chrome web browser to ensure that customers using smartphones, tablets and netbooks would be guided to Google’s services. Google poured millions of dollars into developing a superior Linux-based operating systems for mobile devices and gave it away for free. Google partnered with almost every major cell phone manufacturer which wasn’t already developing their own platform like Apple, RIM, and Nokia, so that Android has became quickly become ubiquitous across the smartphone industry. Between the fourth quarter of 2009 and 2010, Android grew from 4.7% to 33.4% of worldwide smartphone market share, according to Canalys. Today, Android is either the number 1 or number 2 operating system being shipped on smartphones, depending on which market analysis firm is to be trusted.

While the development of Chrome has been plagued by delays, Google is already lining up industry partners for its next operating system. Chrome looks geared to have a similar impact on the emerging tablet and nettop market as Android did in the mobile phone space, although it is still too early to know for sure.

Google’s Android has so dominated the high-end smartphone market, that Nokia has given up trying to develop its own Linux competitor to Android. Nokia recently announced that it would no longer develop its own Linux OS for smartphones, and instead it would ship its premier smartphones with Microsoft’s proprietary Windows Phone 7 operating system. This announcement felt like a death knell for advocates of FLOSS (free, libre and open source software).

Nokia is the largest producer of mobile devices in the world, and thus arguably the largest producer of computers in world as cellular phones acquire much of the power and functionality of a traditional computer. For many people in the world, the cell phone is effectively their first computer, so it is vitally important that they being using a device which runs with free software.

Why is WebOS important?
So if WebOS does little to encourage freedom, much less the adoption of a flexible and open operating system, why is it important? HP is the largest PC vendor in the world. According to Gartner, in 2010 HP shipped 62.8 million PCs, accounting for 17.9% of the global PC market. Although its market share has slipped in recent years, roughly 1 in 5 PCs on the planet come from HP or its subsidiary brand Compaq and there is no computer firm with greater global reach. HP may have the highest failure rate in the industry, according to Square Trade, but HP’s products are marketed to all regional and income sectors worldwide, especially in developing countries. In my travels through 20 Latin American countries, HP is the brand of PC I have most encountered, even in out-of-reach places like Bolivia, where higher quality PCs from Apple, Lenovo and Asus are virtually unattainable. What this means is that people of all walks of life are going to boot up their PCs and see Linux for the first time in their lives. It may be a crippled form of GNU/Linux, encumbered with proprietary software and little scope for expansion. It may only be useful as a glorified portal to the internet, but the fact that doesn’t have that Windows logo and you don’t have to click on the Internet Explorer to get the internet means that it may train a whole new generation of people to not automatically assume that Windows = Computer.

All of this is a good thing, even though I have my doubts that it will lead to a stampede of new Linux adopters, despite the optimistic predictions of pundits like Steve Vaughn. HP historically has shown little willingness to buck Microsoft’s dominion over the desktop and is very unlikely to promote WebOS as a full-fledged PC operating system. Like SplashTop, which inspired it, WebOS will be installed on a tiny partition and will offer few options for expansion. It is unlikely to have OpenOffice.org in its repositories or any of the other necessary application to be a Windows replacement. After a good experience surfing the internet in WebOS, a tiny percentage of the audacious may get curious and eventually stumble upon Ubuntu or some other distribution which truly affords the user some meaningful freedoms, but WebOS certainly won’t advertise itself as “Linux” or “Open Source” and thus does little to introduce users to the wider horizons of FLOSS. It would be highly surprising if WebOS leads even 1 out of every 20 HP owners to eventually install a full-blown GNU/Linux or BSD distro as a Windows replacement. 1 out of 100 is more likely.

Nonetheless, WebOS will advance the cause of freedom in a different manner. No major hardware manufacturer will want to shut themselves off from a fifth of the PC market. With HP demanding that all their PCs be compatible with WebOS, it will force every major manufacturer to finally release Linux drivers for their hardware. In a couple years, every wireless card, every webcam, every cardreader will finally *just work*, because every manufacturer can no longer ignore Linux as a market niche. It will be just like Europe demanding PCs with lead-free solder in 2006. Despite the fact that Europe only consumes 20% of the world’s PCs, within a year or two, every PC on the market was lead-free, because nobody wanted to shut themselves out of a major market.

We can hope that the recalcitrant hardware manufacturers which have historically refused to work with the free software community will now be willing to open up their specs like AMD/ATI and VIA do, and even produce free drivers as Intel does, but it is far more likely that they will give us nasty binary blobs, as nVidia, Epson and Brother do. Still, even binary blobs for the Linux kernel may have beneficial effects over the long term. First of all, having a driver for everything even if non-free, means that GNU/Linux will finally run on all hardware. It is hard to convince people of the value of a free software operating system, if they can’t connect to their wireless network and control their printer. Some people may become convinced of the value of computing freedom from reading the FSF web page, but most don’t become true believers until they have actually used a free operating system. It is one thing to install a free application like Mozilla Firefox in Windows, but the utility of freedom only really dawns on most people the first time they start using a system like GNU/Linux. Comments abound like, “Can you really do that?” “Hey wait, that is cool!” “Geez, there are THOUSANDS of awesome programs here! Let’s try downloading this one!” The joy of exploring something new is the first steps down the path of freedom.

It is only after using free software for a while, that many people begin to appreciate the importance and implication of computing freedom. But the first step is to just get the dang thing running, and although we would like people to choose hardware which is compatible with free software, we have to reach people with the hardware that they currently own. Most people are not going to throw away their current PC, just so they can try out GNU/Linux for the first time. Nor would we want them to, considering all the environmental damage that would cause. That would just convince them that GNU/Linux is too expensive and demanding for their tastes. We should not look askance at using a few binary blobs to get people to start using a 99% free system.

This does not mean that the FSF is wrong about the evils of proprietary binary blobs, but those binary blobs can be necessary evil to get people started on the road to freedom. Once 99% of the system is free, that last 1% which are binary blobs really starts to get irritating and people begin to realize how nasty they are; but first people have to get accustomed to 99% freedom and all the joys of an almost free system.

Acquiring the values of freedom is a process and most people don’t begin that process aware of the implications nor fully convinced of its importance. Once they have been using free software for a while, however, they begin to notice how horrible it is to use that binary driver from nVidia or a proprietary application like Skype. Still, having a few proprietary bits is often necessary to get people to migrate to GNU/Linux in the first place.

Only later, do they start to discover how inflexible proprietary drivers are. Then, they start to have moments of realization:

Hey, I want to try out the latest Release Candidate of my favorite distro, but wait that proprietary driver isn’t compatible the latest Linux kernel. Why doesn’t NastyCorp make a free driver for WidgetX so any kernel hacker can dig into the code and fix this problem? Why doesn’t NastyCorp provide any information about its product, so we can write decent driver? Look! CoolCorp has the full specs online for their WidgetX and they have free drivers which work with any version of the kernel. Dude, I got screwed! Next time, I am going to buy from CoolCorp rather than NastyCorp. RMS was right. Binary blobs suck!

After using proprietary applications in a free system, people begin to make comparisons between proprietary and free software:

What a pain! I know that I have been using NastyApp for years, but why does it force me to go to its web site to download it? Looking at all these screens of legalese and clicking all these “Accept” buttons sucks! Why can’t I just type “apt-get install NastyApp” like all my other favorite apps? And why isn’t it incompatible with PulseAudio and the rest of my system? Dude, look at that bug in NastyApp! I had better jump onto their website and report it like I do with all my other favorite apps. Hey!, where is the public bug tracker? Where is the developer mailing list? Is anyone paying attention? What the hell! I don’t want to read all this corporate bullshit. Where is the source code? Maybe I can find the problem. Oh wait, this is proprietary software, I can’t even look at the code like the rest of my apps. Wait! I just found FreeApp which does the same thing as NastyApp, but it doesn’t have any proprietary crap in it. Maybe it doesn’t look as slick or have as many features as NastyApp, but at least it doesn’t come with any crazy restrictions. Dude, I am only going to use free software in the future!

What all of this means is that we are inching little by little toward greater freedom for users. Although it may not be written anywhere explicitly in the business plans of companies like Google and HP, they are opening up new spaces where the agenda of freedom can nestle around the corners. It is up to the users themselves, however, to take advantage of those freedoms and not wait passively for the companies to grant it on a silver plate.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s