Encoding of user ID strings

[Robert J. Hansen]

> In the case of "all 8-bit characters, no 7-bit" you're dealing with
> either a practical joker or EBCDIC. Same thing really...

Or KOI-8R/Windows-1251.

> After that you're into heuristics. There are quite a few programs out
> there that attempt to detect encodings statistically, but with such a
> short string of data you might as well pick a number. ;-)

Yeah, that's what I'm afraid of.  It's not valid UTF-8 encodings that
trouble me: it's having to deal with unknown encodings.