A Tale of Two Operating Systems – Info Gadgets

An operating system shouldn’t be viewed as a standalone piece of software that a company ships once in a while. It is, instead, the sum total of the entire ecosystem of solutions that comes to life around it. Over the long haul, the platform stands or falls due to the health and vitality of that ecosystem.

A successful platform lasts decades. As such, every architectural decision made early in its life cycle becomes critical in the long run, much more so than any dozen decisions made later in its life.

When analyzing the recent history of operating systems, a pattern emerges that is worth a closer look. At any given point in the recent past, in fact, two operating systems have competed for dominance in the general-purpose computing arena. Examples are: Unix vs. Windows, Netware vs. Windows, MacOS vs. Windows, Linux vs. Windows, iOS vs. Android. By dominance, I am referring to some fuzzy formula based on unit volume, total revenue, application availability, total cost of ownership (TCO), performance, reliability, and developer mind share.

The names have changed over the years but the story is often the same. When it comes to building a platform, there are basically two strategies available: rich or reach. (I didn’t invent these terms, they’ve been around for decades.) Both can be successful but they outline markedly different approaches to ecosystem development.

In the case of the former, a rich, coherent, and elegant end user experience is prized above all else. Apple’s products belong into this category, where a “walled garden” approach delivers a superior end-user experience at a premium price, achieved by tight control over the ecosystem, a reduced number of choices at each integration point resulting in a smaller test matrix and hence fewer opportunities for breakage. This approach can be highly lucrative; the iPhone covers less than 15% of the smartphone market in terms of unit share but accounts for 90% of the overall profits.

By intentionally reducing the available hardware choices and maintaining complete control over the out-of-box experience, Steve Jobs forced the Apple ecosystem to “move up the stack” and innovate in applications instead, while also improving overall product quality. Combined with a central app store, this also resulted in better quality control at the application level.

In the case of the latter strategy, the top priority is to reach as many customers as possible. Windows falls into this category, as does Android. Microsoft’s mantra of “a computer on every desk and in every home” was the rallying cry for Windows for many years. By necessity, the reach strategy requires an open ecosystem of partners and many integration points into the platform. It often yields a less cohesive end-to-end user experience as various integral parts of the product are designed and tested by different vendors.

Windows-powered laptops are available from dozens of vendors in a dizzying array of hardware and software configurations. When your platform strategy is to reach the broadest audience, these OEMs (Original Equipment Manufacturers) often end up competing on price. As a result, they choose the cheapest available piece of hardware to save a few pennies on the COGs. The OEMs have to make money somewhere in the value chain, so they then also add what has euphemistically been called “crapware” to the system.

Your PC shows up with a few (or a few dozen) third party apps pre-installed, apps by vendors who paid the OEM a fee for that placement — hoping to sell you a monthly subscription for features ranging from antivirus to backup to ink cartridges. Given the historical lack of application life cycle management in Windows (more on that later), these apps often step on each other’s toes. As expected, the end to end quality suffers.

With a MacBook or iPad, available only from Apple, you get the choice of one disk controller, one network controller, one graphics controller, etc. — whichever one Apple decides to offer. It wouldn’t be too much of a stretch to claim that, often, the only thing you get to pick is the color of the device: silver or gray. Apple intentionally reduces the hardware interoperability matrix which, in turn, improves quality.

There is no magic here: increased test combinatorics and distributed responsibility for end-to-end testing always result in lower end product quality. Windows (and Android) will always deliver a less polished product than anything Apple produces since someone else (the OEM) gets to pick and choose the hardware and software components that make up the overall system — and he gets to do so from among a dizzying array of options, most of which haven’t actually been tested together. This is exactly why Microsoft started designing its own “Surface” tablets and Google its own “Pixel” smartphones.