There is an infrastructure beyond the usual technical sort, an infrastructure of skills, habits, and assumptions. Windows matches this infrastructure extremely

There is an infrastructure beyond the usual technical sort, an infrastructure of skills, habits, and assumptions. Windows matches this infrastructure extremely well, because this infrastructure was almost entirely built on Windows over the past 30 years. People have grown accustomed to Windows, the Windows version of Office, and domain-specific custom apps that only run on Windows. Even Mac users struggle to integrate into this infrastructure, and that difficulty, plus the higher cost of Macs, is why you rarely hear a city or government agency declaring it is going all Mac. For (desktop) Linux, the situation is even worse. Unlike Mac, Linux isn't usually pre-installed, so you never seem to get things working right. Unlike Mac, there is no version--not even an old version--of MS-Office. The flip side of Linux's freedom to build anything you need is that you almost have to build everything you need, beyond the browser. Three years ago I was struggling to deal with my kids' (Silicon Valley) school system, because teachers would always send me important information in the form of MS-Word docs, and I no longer owned a copy of Office. By last year, almost all of them had switched to Google docs, and the problem is almost gone. What I expect to happen over the next decade is that the skill/habit/assumption infrastructure will turn away from desktop Windows toward cloud services usable through any browser from any platform (desktop, phone, tablet, wall screen, wearable....) Munich might need some new Windows machines during the transition, but they should probably also work on a transition to online services at the same time, after which desktop Linux will be more practical than it has been in the past.