Good point. Do filesystems use a normal form to at least prevent having two files with effectively the same name?
I should point out the flip side though, that there’s no avoiding Unicode in filenames. Users in languages that don’t use the Latin alphabet (such as Japanese, Chinese, Korean, Hebrew, Arabic, Greek and Russian, and the list could go on) can reasonably expect to be able to give a file a name they can read and understand with no extra effort. All the software woes that come with it - too bad, software needs to deal with it.
I’m not sure. A few years ago I remember that OpenBSD expected ASCII for files, but I think Linux expects utf-8. I could be wrong though.
I’m assuming Unicode anyway, and UTF-8 is by far the most natural because most files will be in ASCII. A “normal form” (see link above), you might think of it as a canonical form, is a way to check if two strings are equivalent, even if they encoded the text differently. Like the example mentioned on Wikipedia:
For example, the distinct Unicode strings “U+212B” (the angstrom sign “Å”) and “U+00C5” (the Swedish letter “Å”) are both expanded by NFD (or NFKD) into the sequence “U+0041 U+030A” (Latin letter “A” and combining ring above “°”) which is then reduced by NFC (or NFKC) to “U+00C5” (the Swedish letter “Å”).