Network Effects in Operating System Usage
Operating systems are at the core of our computing experience. Â They allow us to execute programs, manage files, communicate with hardware, and network with other computers. It makes sense, that as a result, users would use the best operating system out there for their daily computing needs. However, due to the concept of network effects in technology, this is not the case.
First, I should explain my experience and preference with operating systems. Until one year ago, I had always used Windows. It had almost always suited my needs, and, being highly proficient with computers, I could fix almost any problem that came up. Then, last year, during my first year here at Cornell, I joined a project team that almost exclusively developed in Linux. As such, I decided to try it out. After a couple months, I learned how Linux was superior to Windows in many respects. I also realized some of its failings, having to do with compatibility. At present, I run both Windows and Linux on all of my machines, sometimes simultaneously through the use of a virtual machine. Both systems have their merits, which I will explain. Note that I have never tried using an Apple operating system and as such, I will avoid offering opinions in that area.
First, consider the current state of operating system use. According to recent studies, approximately 77% of users use Windows, 7% use Mac OS X, and a little over 1% use Linux-based operating systems. The remaining percentage is attributed to mobile operating systems. From this information, one might assume that Windows is the undoubtedly the best operating system. This however, is far from the case. While an in depth analysis would require a course in Operating Systems, consider the basics. Running benchmarks shows that Linux is significantly faster than Windows at most performance related tasks. Additionally, in terms of system security, Linux beats windows again. Finally, considering monetary cost, Linux, being free, is clearly superior to Windows’ steep pricing. So, the question remains: why would people stay with Windows when Linux is superior in terms of core functionality of the operating system?
The answer lies in the fact that operating system usage is heavily reliant on network effects. The value of an operating system to the average user takes into account the percentage of the population that is using it. This makes more sense considering the functions that an operating system provides. Users want access to the best programs. Companies, wanting to reach the most amounts of people, naturally develop for the most popular operating system. Therefore, the quality of the programs is directly proportional to the percentage of users of an operating system. This creates more dependencies, since these programs in turn have network effects. For instance, consider word processing. In order for two users to collaborate, they have to use the same program. In order to use the same program, they must use the same operating system. Note that I am ignoring the case of multi-platform programs and universal file formats for simplicity. Therefore, the principle reason that Microsoft is dominant as an operating system is the fact that the majority of users use it. Linux can never gain traction as a dominant operating system because too few people use it. To the vast majority of users, the lack of programs and compatibility is a serious deterrent. In fact, this is the primary reason that I continue to use Windows. Some extremely powerful programs in Windows simply cannot run on Linux.
As further evidence of the power of Linux, consider other computing networks. Consider servers, essentially computers that are purposed to host specific functions to users on the internet and otherwise. Servers are far less reliant on the specific operating system of their clients, and as such, it is not surprising to discover that approximately 64% of servers run Linux. Additionally, supercomputers are even more Linux biased with over 91% of them running a Linux variant. Supercomputers require almost no network effects in terms of their value since users of supercomputers almost always write programs specifically for the task at hand. The steep percentage using Linux in these cases is no coincidence. Power users recognize that Linux is a superior operating system in terms of performance.
However, despite all its power, Linux lacks the fundamental ability to become a dominant force in consumer computing. In order for people to begin to switch to Linux, the usage needs to hit a tipping point, at which, the value for most users to switch will exceed the cost. Not much can be done to the cost of using Linux. Linux is quite easy to use and is free. Its only downside is the lack of usage, which cannot be changed by fixing software. The only other option is to raise the perceived usage of Linux through advertising. However, this is equally unlikely since Linux is open source and cannot afford starting advertising campaigns. As such, Linux will continue to remain a niche operating system until the market changes in some manner, lowering the cost of Linux even further. With recent trends towards cloud computing, I would not be surprised if something like this does happen in the near future.