This is a very annoying Windows 2008 R2 bug which was reported way back in 2010 but still there is no fix from Microsoft. Here’s an example from the educational sector.
Here’s the symptom: You have a bunch of users and want to keep their “my documents” folders on a network share so that it can be easily backed up, and so if their computer dies, it’s no big deal.
You create the share, we’ll call it “UserData” and share it out to all the workstations. Then, you go to each windows 7 workstation in turn and start redirecting each user’s “my documents” to a folder with their name in that share. So “Stan” would get a “Stan” folder inside “UserData”. Nice and tidy. All of your users now have their own personal sub-directory and it’s easy to see who owns what data. It’s also easy to back up and restore.
At some time later on (not sure what the trigger is, but it seems it has to do with when the user logs off), this folder is mysteriously renamed to “my documents”. Now, you have a ton of folders in “UserData” which apparently have the same name (“My Documents”), and you don’t know who owns what data!!! The links from the Windows 7 PCs seem to stay intact, but it’s impossible to manage at the server level.
Sure, one solution is to have “UserData/Stan” and put a “my documents” folder in there. But why should I do that? When you do the “my documents” redirect, it allows you to pick a folder as the redirect target. If the folder name you’ve picked is not acceptable, you shouldn’t be able to pick it. But there is no such error or warning. Everything seems fine until some random time when the system decides to “fix” it for you.
This is what I refer to as “arrogant code”. Somebody at Microsoft made a conscious decision that they know better than you do, and regardless what you wanted to rename that folder, they will impose their iron will and rename it the way THEY want it. This isn’t a bug where somebody made a typo in the code. This is a very poor consciously made behavioral decision. One might wonder… why? Why is some engineer at Microsoft overriding the customer’s decision?
So which is it, a bug or "da man" writing arrogant code? No competant network administrator would put collections of data folders under a root share. You put them under /share/username/… If so, this is a minor annoyance at best (and I don't even get that point). /share/username/documents can be whatever name it wants, as long as no links break (which you said they don't).
Since this worked in 2003, but doesn't work in Server 2008, you could call it a bug. But a bug is typically some piece of code that went wrong. This is an intentional "rename" call which renames my folder. So I'd say it was intentional, and given that Microsoft has layers of code review and design review, multiple people thought this was a good idea.
Putting data folders or not in a root share is totally irrelevant. The whole point is that the operating system should act as a base for applications, and as such should be predictable and solid. Randomly renaming subdirectories is absolutely ridiculous behavoir for an operating system.