It would be useful to have a portable way to markup non-ASCII characters in comments. Some possibilities: &#xhhhh; or &#dddd; (as in SGML) \xhhhh (possibly with \&, as in Haskell strings) \uhhhh (as formerly in the Haskell Report) These could be converted to Char by the lexer, so there would be no direct connection to the output format (the HTML backend renders them as &#xhhhh;).