Utf Seven

Defined in RFC 2152, this is yet another UniCode encoding, this one targeted at mail communication over 7-bit SMTP. ASCII equivalent for 'direct characters' -- alphanumerics plus some punctuation. Depending on the profile, some subset of the other printable ASCII characters except '+' ('optional direct characters') may also be ASCII equivalent.

'+' is expressed as '+-'. All other non-direct code points are UTF-16 encoded and then Base64 encoded (MIME profile minus the '=' padding) and put between a '+' and any non-Base64 character. If the terminating character is a '-', it is ignored by the decoder, otherwise, the character is preserved.

Example:

Native: Vässarö, scoutön

UTF-7: V+AOQ-ssar+APY, scout+APY-n

Slightly less efficient than QuotedPrintable?-encoded latin-1 for typical Western European texts, with the bonus of being able to cover all UniCode code points. Excellent for ASCII text with the occasional math, wingding or CJK character.