The charset identifer for UTF-8 is defined by the IANA as "UTF-8"; case is supposed to be insensitive. There are no aliases defined. In particular "utf8" is not officially recognized, though some software does implement that identifier as an alias.
In order to ensure maximum compatibility, I think we are best off using the official forms "UTF-8" or "utf-8" (not "UTF8" or "utf8"). I fixed the cases of this that I could find. They were mostly in Atom-related code. --a.
