Say I have a script which at some point serves a plain text file as a content (right after "\n\n"). These files are provided by users, but I can expect they will be UTF-8. So I hard-wire Content-Type: text/plain; charset=UTF-8.

But while I can teach users to save everything in UTF-8, I can't be very sure that the files will be without BOM ("\xEE\xBB\xBF"), as at least on Windows, this is not very clearly distinguished in common plain text editors and not every one of them uses the same default.

So what about these files created on Windows, where they may/may not start with BOM? Should/will server or UA get rid of this debris for me? Or is it my task to prepare clean UTF-8, i.e. open each file and check whether BOM needs to be removed?