they decided to forego commercial support because they felt they could do a better, and less expensive, job in-house. Turns out, in hindsight, they were wrong. o sell product.
No. That's Microsoft propaganda.
The reasoning was not that felt they could do better and cheaper. But as it is, they are doing cheaper and better.
What we have here is not a critique of GNU/Linux systems so much as a critique of the Munich bureaucracy
True. It is mostly an example of what you get when you have 17 independent IT departmenst working on the same intranet, trying to find a consensus. Which had been the case before and during the migration phase. Plus, what you get if you dive into such a huge project saying "yeah, no problem, we'll figure as we go".
If you read the project's site (in German only, unluckily) you believe it's a joke. They had 17 IT departments support 22 distinct groups with some 14k machines of vastly different hardware, communicating via two incompatible file sharing networks and using half a dozen of MS Office versions. Half of the "special functionality" for running the registry office work was written as ActiveX controls, and most of the "office special stuff" would be Visual Basic macros that did some stuff nobody could understand or maintain.
What they realized was that the longer they stuck with this, the more they would get locked in and dependent on Microsoft-proprietary technology, and the more expensive it would get.
So they moved to OpenOffice expecting that it would "just work" and immediately got complaints that OpenOffice and Word wouldn't play together seamlessly, and Office would format things differently. Surprise.
After that, they eventually moved over everything, and now after nearly a decade, they have one consistent system. Was it a shit approach? Probably. But the result isn't really that bad.