As opposed to an admin who thinks he's above user preference?
Who are you to dictate file extensions? What if he didn't want to use them at all? What if he had never heard of those particular extensions. What if another program used the same extensions?
Okay, I'll bite. Your last assertion seems valid, but cannot validate the rest of the shenanigans in your post. If a user wants to name paper files which exist in a complete vacuum relative to standards, that is how someone wants to file their papers is generally an arbitrary choice, then that is fine. However, if users want to maintain data in an environment which consists of standards then it most certainly is the admin's responsibility to ensure the users stay withing those standards, or at least in most cases it should be safe to assume those standards are valid and followed.
Of course, we know how assumptions work, so it is also the responsibility of the admin the event that valid data is affected by such automated processes. As a matter of policy I do not delete data in customer use areas and leave that up to the users, even in times when space is low and I have to guide the user through the process I stay away from the liability.
I will agree that if the admin acts maliciously in a this-will-teach-them approach, without ever having taken the time to advise or guide the user, said admin is not meeting his responsibility, but certainly, yes, standards trump the user's preferences, especially for forward usability.
You remind me of the admin who blindly deleted someones file called "penis" that contained biological research data.
He didn't get away with it. I'm surprised that you did.
This is not even close to the same thing. If someone wants to put the name "penis" in their files, then so be it, even more so in a biological research environment. Stipulation this is a real even in the first place and your retelling of the tale is a 100% true representation of the event, the admin who did this sounds like a penis, himself.