Mobile platform and technical debt

Editor's note: This article is a16z partner Benedict Evans. The article points out that the mobile operating system is based on some of the assumptions and compromises at the time, and the conditions change, and the platform changes. Change does not come, it means death. That's how Nokia, Palm and Rim were killed. So after a 7-year cycle, did Apple and Google, which killed them, face the same problem? What assumptions and trade-offs should be made next?

Why did Nokia, Palm and RIM get killed? One of the fundamental problems is that 2000 years or so, the design revolves around the assumptions and tradeoffs that were right at the time, but these assumptions and trade-offs are hard to compete with IOS and Android 5-10 years later. The assumption was that CPUs and networks were slow, with little memory, only resistive touch screens or even none, sacrificing performance and a richer experience to improve battery life. In the 2000, these assumptions and trade-offs were correct, but were not established by the year 2007. That means they have to replace the platform, and the replacement platform almost means a near-death experience, with Apple and Microsoft as proof.

So the question is: what assumptions did Apple and Google have that would hinder their future progress? From the Nokia S60 series/Symbian mobile phone to the 2007 IPhone launch has 5 years, but now 8 years have passed. Much of what we've seen recently can be seen as a change in the balance of the time.

The original iPhone resolution is fixed, does not run Third-party apps, and does not support multitasking. The reason is partly philosophical (believe it or not, jobs doesn't want apps), partly because the IPhone is MVP in many ways. But most of it has to do with a problem, that is, under that cost and power allocation, how to get a device to achieve the basic level of experience the user wants. The problem is exactly the same as Nokia, Palm and RIM faced in the early years. Secure, sandboxed, multitasking, third-party applications are not the first generation of hardware that can be managed, at least one battery doesn't last a day.

In other words, what Apple has been doing in the last few editions is to rebuild the operating system so that the compromises are adjusted to improve performance (about 50 times times more). So now, 8 years later, we do have a sandbox security multitasking and expansion, and the IPhone and the IPad are abandoning fixed resolutions. In the process, Apple has replaced the engine and has been somewhat lucky to succeed. It has changed all the assumptions at the time.

Android can see things like that, and the key trade-off is openness. Google developed/purchased a set of Open-source operating systems with very basic UX and little centralized control, which was the right trade-off at the time. As a result, it has been hugely successful in its unprecedented size-now Android devices are already more than the PCs they use, twice times more significant than the consumer PCs. Of course, this leads to fragmentation as well as differentiation and sub-standard user experience, and there is a constant attempt by third parties (Amazon, and perhaps Millet and Samsung) to take over the user experience.

It's right to make that trade-off in 2007-you can look at the alternate path to Windows Phone. But the trade-off is changing as time shifts. If you want to attract developers, then fragmentation has to be dealt with, and as the OS itself becomes the aggregation layer of the services in front of the browser, Google's control of UX becomes increasingly important as the user data becomes something as important as the original web search scope. So now you can see that Google is shutting down some of Android's "open" doors. By putting its own services and APIs into the GMS (that is, the Googlemobile service), make sure that a larger portion of the library has the latest "version" and work harder to maintain control of the UX, preventing OEMs from changing too much. It's not clear where the situation is going to stabilise, because the proliferation of Android in China and the turbulence of Android OEMs are clearly correcting those assumptions about openness.

How do you look at all this? One way to look at the overlap between IOS and Android, though, is that they started off at different ends, but now there is a convergence of abilities. Apple is loosening control, while Google is tightening. Of course, Google and Apple each have to add a lot of things to Android and IOS separately (to draw inspiration from each other), just as Apple has added cloud services and Google has redesigned the user interface.

But their underlying philosophy is still very different-for Apple, the device is smart, and the cloud is dumb to store, and Google is the opposite, the cloud is intelligent, and the device is dumb glasses. These assumptions and trade-offs are still deeply ingrained. At the same time, the next phase of the smartphone (will be the platform of the message app for the dominant interface of the watch)? Will test all of these assumptions again.