Ask Paul: July 27 (Premium)

We just returned from Berlin, so greetings from heat wave-soaked Stockholm! Here's another round of Q & A to end the week.
A lightweight Windows alternative to Chrome
christian.hvid asks:
Given the success of Chromebooks, there's obviously a market for light-weight, zero-maintenance operating systems that are basically there just to get you online. Needless to say, it would be the easiest thing in the world for Microsoft to follow in Google's footsteps and whip up a Windows version consisting of just a minimal core and Edge. It would run PWA and pure UWP apps of course, since the APIs would be there anyway, but the browser would be front and center ... Of course, Microsoft must already have pondered this a hundred times and dismissed the idea for whatever reason. Why is it, in your opinion, that Microsoft rejects a model that has already proven its appeal among consumers, small businesses and schools - the very markets that Microsoft now stands to lose?
One of my favorite lines of all time is Donald Rumsfeld's "you go to war with the army you have, not the army you might want or wish to have at a later time." That sort of explains what's happening here.

Which is, the platform that Microsoft made and rode to great success is a complex legacy product that has, in recent years, been largely supplanted by simpler, more modern mobile platforms. And as I've noted in the past, taking something complex like Windows and making it simpler is hard, whereas taking something simple (like Android, Chrome OS, or iOS) and making it more sophisticated is, if not "easy," at least easier.

In the past, Microsoft was able to thwart existential threats to Windows like OS/2 (mid-1990s), Netscape/the web (late 1990's) and Linux on netbooks (late 2000s) for a variety of reasons, of course. But it's fair to say that Microsoft was able to leverage various advantages in each case and emerge victorious. But today's mobile/web platforms represent a new kind of challenge. And with people's usage requirements evolving, classic desktop systems like Windows (and the Mac) are less necessary for most people.

Microsoft has been working to simplify Windows for decades, and in a variety of ways. And componentization is one of the core repeated efforts along those lines: NT was designed to be more componentized than classic Windows on DOS, and there have been many, many efforts over the years to further componentize the NT-based versions of Windows we've been using since XP. (Most recently, you've heard of the latest go-round in the form of things like Windows Core and Polaris.)

Microsoft idea is that you can build a scalable platform that will run on anything from small embedded device to datacenter servers, and everything in between. It's the "One Windows" thing. And it works, sort of. Windows 10 is the basis for a wide range of non-PC platforms today, from the Xbox One to Surface Hub.

But it also doesn't work in some key ways. For example, Microsoft had...