In a recent article, Jack Wallen built a case for the ascendancy of Linux over Windows. Now Kris Littlejohn steps up to argue the converse: From cost to software availability to ease of use, he suggests that Vista notwithstanding, Windows is the superior OS.
In the last few years, there has been a surge of renewed interest in using Linux, in both the server and desktop spaces. Several factors are contributing to this surge, all happening at once. First, there is the trend from powerful desktops to smaller, but less powerful, notebooks — and now, netbooks. In addition, more user (and media) friendly Linux distributions, such as Ubuntu, have hit the scene. Finally, here comes the OS everybody loves to hate: Windows Vista — significantly more resource hungry than XP and perhaps released a little too soon. The confluence of these factors has led many users to give Linux a shot.
Largely, however, Linux has still been found wanting. Whether because of some inherent weakness of Linux, a preconceived advantage that doesn't pan out, or the fact that users simply miss their familiar Windows functions, there are a number of reasons why Linux isn't triumphing over Windows. I'm going to look at 10 of these reasons, some that apply primarily to servers, some to desktops, and some to both.
The flip side
Note: This article is also available as a PDF download.
1: Cost comparisons are often misleading
Let's get what may be the most controversial point out of the way early. First, in the server space especially, we should try to compare apples to apples. This means comparing Windows Server to paid Linux. By far the most dominant flavor is Red Hat Enterprise Linux (RHEL), with about a two-thirds share of the paid enterprise Linux market, so this seems the most logical comparison. While there are plenty of free options out there, such as CentOS, for a business running mission-critical workloads, an unsupported operating system is a hard pill to swallow.
There are a couple of ways we can look at cost, neither of which is nearly as flattering to Linux as one might expect. First, we can look at the costs directly related to the acquisition of the platform. RHEL is a subscription-based license, meaning that rather than pay for the software itself, you pay for support. This doesn't mean just phone tech support or troubleshooting (although that is included too, whether you want it or not) but also includes standard patches and bug fixes. Standard support for RHEL 5 Advanced Platform is $1,499 per year per server, or $4,047 for three years. Compare this with $3,999 for Windows Server 2008 Enterprise edition with free patching and bug fixes, and you can basically call it a wash unless you use a ton of phone support. And there are also features that aren't included and must be purchased separately, such as Red Hat Directory Server — thousands more per year.
The other way of looking at cost is total cost of ownership (TCO) of the platform, and this leads into our next issue.
2: Expert talent is more readily available
When looking at TCO, we're not just looking at the software costs but also at staffing and administration costs, costs due to downtime, hardware costs, etc. Of these, staffing is the largest, accounting for more than half of the TCO. Here, Windows wins out because IT pros experienced with Windows are much more plentiful and generally cheaper to hire than Linux experts and because they can often be more productive.
With Linux, efficient management over many machines usually means going to the command line and pounding out a script to automate a process — which is cool. However, with Windows Server 2008, PowerShell is now built in, which means the Windows guys can do that too, arguably better. Add that to the System Center family of tools, where virtually all management tasks are available at the click of a button (and which really have no peer on the Linux side), and Windows is simply easier to manage.
3: Linux isn't actually trying to compete head-to-head
The last reason Linux isn't triumphing over Windows in the server space is that it's not really the primary focus. Right now, both Linux and Windows are gaining in server market share. How is that possible? Old granddaddy UNIX is being thrown under the bus to make it happen. Today, companies are dumping their old mainframe or proprietary UNIX servers for cheaper x86-based commodity hardware. It's easy for a Linux sales guy to come in and make the value proposition: "It's essentially the UNIX you know and love, but it runs on hardware a fraction of the cost."
Unfortunately, the market for UNIX conversions and mainframe modernization is drying up. When those deals are gone, Linux will have to compete head-to-head with Windows to continue its growth, and this is a much harder proposition to make. Why should an organization already using Windows change platforms and have to build whole new skill sets around Linux?
4: Windows offers familiarity and ease of use
Let's face it: Whatever else you might say about Windows, it is easier to use. We love our Start menu and our Task Manager and our system tray. Some of us are even starting to love our Vista Sidebar and gadgets. Young adults today never had to use MS-DOS, even if they started using computers at an early age, so they aren't going to be comfortable at a Linux command line.
Don't get me wrong — Linux has come a long way. But remember how far back it has had to come from — where just managing to install the operating system for a non-expert (and sometimes experts too) was considered a major triumph. There are still too many things in the Linux world that are expected to be done manually, like program installation. A majority of users will say, "I might have to compile something myself? No thanks."
5: More software is available
Here's another one that's a pretty clear edge for Windows. It isn't about being able to play the newest games, even if that is one of the most often raised issues against using Linux. Simply because Windows is the dominant operating system, there is much more (and usually higher quality) software available for it than for Linux. Much of it comes from evil Microsoft itself.
A good example goes straight to one of open source's greatest recent successes: OpenOffice. OpenOffice is great software... considering it's free. I use it when I'm in a pinch on somebody else's computer. It's almost certainly adequate for a light user or a student typing up a couple of essays. As a writer, however, I can't imagine being stuck without Microsoft Office for long. When it comes to features like SmartArt, quick table generation, editing and review functions, and inserting basically any kind of object into a document, there is no comparison. When you go beyond the word processor to the presentation software or spreadsheets, the gap grows even wider.
Now of course much of our favorite Windows software can be run using an emulator such as Wine, or on a virtual machine running Windows — but if we find ourselves doing that all the time, why use Linux in the first place?
6: Windows Vista is just a bump in the road
This had to come somewhere. Vista has become the proverbial red-headed stepchild of the operating system world — everybody loves to beat on it, and not completely without reason. Vista hogs resources like no OS before, and initially, at least, there were plenty of issues with both software and hardware compatibility. Without going too far in singing the virtues of Windows 7, however, I think it is reasonable to say that there is no reason to expect Vista's shortcomings to be repeated in the next generation.
Windows Vista was like the growing pains experienced by a teenager when he starts to, well, grow. Too many things were happening at once, and there was bound to be some pain involved. We combined security changes such as UAC and how applications were handled fundamentally by the operating system with lots of nifty but resource-intensive gadgets at a time when so many users were switching away from their powerful desktops, and we got a sluggish OS where things don't always work quite right.
However, Windows 7 is faster. Software vendors and Microsoft have had time to update their code, so now applications are compatible. Before much longer, Vista will be behind us.
7: Hardware continues to advance
Now, while Windows 7 is significantly faster than Vista, I won't try to claim that it will be as friendly to the lowest end hardware as Linux. Fortunately, time marches on — and hardware improves. We can now get a quad-core processor and 8 gigs of RAM in our laptops. Intel has a dual-core Atom processor out, and even if it is made for nettops rather than netbooks, it's a safe bet that a dual-core Atom with netbook-friendly power consumption levels is right around the corner. In any case, as hardware continues to advance, that aspect of the Linux argument will become more and more irrelevant.
Also, while we're on the topic of netbooks, let's not forget that while these may seemingly be the perfect candidate for conversion from Windows to Linux, according to a Laptop Magazine interview of MSI's director of U.S. sales, Andy Tung, the return rate of netbooks running Linux is much greater than the rate of those running Windows.
"They start playing around with Linux and start realizing that it's not what they are used to. They don't want to spend time to learn it so they bring it back to the store. The return rate is at least four times higher for Linux netbooks than Windows XP netbooks"
Both servers and desktops
8: Claims about open source don't stand up to scrutiny
Much of the hype about Linux is really more about open source development in general. The buzzwords all sound good: Open source is about sharing. Collaboration. Proliferation of knowledge. For certain, there is nothing wrong with the open source model, and its use surely helps advance new ideas in software development. As a business model and a model for end-user products, though, it's less reasonable. Here, it causes a lack of standardization. Egos among the different developers collide, and the final product suffers. Let's not forget the old adage "Too many cooks spoil the broth."
Another claim is that Linux and open source software are more secure than Windows and Microsoft software. This is largely based on problems with legacy versions of Windows. Back in the NT and Windows 2000 days, there were valid points to be made, but this is far less true today. The last several years have seen a massive emphasis on security across the industry. And now, with Windows Server 2008, Windows Vista, and the whole Forefront line of products, Microsoft is running a pretty tight ship — enough so that major competitors such as Red Hat are not really bringing up the security argument against Windows anymore.
9: Linux is still too bleeding edge
Since a Linux advocate will be quick to bring up the problems associated with Vista's perhaps premature release, it seems only fair to bring up similar issues with Linux. Though not as high profile as the problems with Vista, the inclusion of KDE 4 with Fedora 9 was perhaps just as much of a flop.
This is no big deal as an isolated incident. But the problem is that it was merely a symptom of a larger issue in the Linux world: Much more emphasis is placed on technological innovation than usability, and often a technology will be dropped or juggled with something new at a moment's notice. Look at Red Hat's hypervisor history. At the end of 2004, Red Hat announced support for Xen — first with Fedora and to be included later in RHEL. Then, fast-forward a year and a half to summer 2006 and it announces that Xen isn't ready and that Novell was irresponsible in including it with SUSE. Less than a year later, in spring of 2007, RHEL 5 was released and Xen was a major feature. And last summer, in 2008, at the Red Hat Summit, it announced the shift from Xen to KVM.
Whether Xen or KVM is better isn't the issue. The issue is that an emphasis on the newest but perhaps untested technology leads to rapid and inconsistent changes in direction that leave users scrambling to keep up.
10: The Linux culture isn't always responsive to the common user
This all brings us to the last reason Linux isn't triumphing over Windows: the Linux community and the users and advocates themselves. Linux is a geek's OS. I don't mean geek in a derogatory way and certainly include myself in the term: a very technically skilled user who derives pleasure from the technology itself.
The problem is that geeks are not the majority. We don't drive the market, the Average Joes of the world do, and Linux geeks in particular don't seem to be receptive to that. In an already unfamiliar and more difficult environment, when people using Linux for the first time encounter a problem and turn to the community for help, too often they're met with ridicule. They don't want to hear, "All you had to do was recompile the kernel."
Of course, this is not the case for all users, and there are plenty of dedicated individuals providing free support in the forums around the world. But until Linux users as an entire community can accept that all users are not programmers or even power users, there is no chance that their operating system of choice will come out on top.
It's in your court...
Okay, you've heard from Jack, you've heard from Kris — so what do you think? Do you see valid points on both sides of the debate? Have any issues been overlooked? Share your opinions and OS experiences in the discussion below.
Kris Littlejohn grew up in a household of tech writers and has been playing with, building/disassembling, and writing about computers and other gadgets from an early age, including a number of articles for TechRepublic.