What when you have names that do not
translate very well into english code?
Or when your company and your
customers are (and always will be)
local and not international, like
governments? And what about using
special characters like people in
almost all countries have?
All of those characters are available in UTF-8. The only drawback is that UTF-8 uses 2 or 3 bytes for characters that ISO-8859-n represents in 1 byte, but this is a very minor problem for code because most code uses ASCII-only identifiers, even in languages where Unicode ones are allowed.
If you use a single-byte encoding like ISO-8859-n, you restrict your character repertoire to 256 characters. And what if you need characters (like “—”) that are not in ISO-8859-15?
What if one of your tools misidentifies your files as being encoded in ISO-8859-1, and you lose your €'s?
What if your locale-specific encoding is one of the rarer ones like VISCII that isn't supported by all charset libraries?