2440 defines a User ID (section 5.11) as:
A User ID packet consists of data that is intended to represent the
name and email address of the key holder. By convention, it
includes an RFC 822 mail name, but there are no restrictions on its
content. The packet length in the header specifies the length of
the user id. If it is text, it is encoded in UTF-8.
The various 2440bis drafts keep this same text.
Suggestion: change the current specification of contents (which is
mostly unspecified) to UTF8 text only.
Rationale: The current language was written before the OpenPGP
standard got attribute IDs. Attribute IDs are a significantly better
way to incorporate arbitrary binary blobs into user IDs.
A simple way to look at this is to note that any user ID that isn't
UTF8 text is going to be implementation-dependent. Since it is
implementation dependent anyway, implementors may as well use
attribute packets which won't gum up the programs (all OpenPGP
programs I've seen) that more or less do expect user IDs to be text.
Suggested language:
A User ID packet consists of UTF-8 text that is intended to
represent the name and email address of the key holder. By
convention, it includes an RFC 822 mail name, but there are no
restrictions on its content. The packet length in the header
specifies the length of the user id.
David