Martin Monsorno writes: > � > > doesn't seem like an 'ü' to me Neither to me. The three bytes I received were ef, bf and bd, displayed as: "LATIN_SMALL_LETTER_I_WITH_DIAERESIS" "INVERTED_QUESTION_MARK" "VULGAR_FRACTION_ONE_HALF" in the iso-8859-1 encoding of the message. If you want to send us unicode character you'd better use utf-8! $ od -t x1 /tmp/a 0000000 ef bf bd 0a 0000004 $ cat /tmp/a # I've got a utf-8 terminal � and this is a unicode "REPLACEMENT_CHARACTER". > the umlauts are not de-/encoded correctly, don't know whom to blame > (however, most of the times it's eclipse). and i also don't know how > to handle this problem. If you were expecting a ü, then it's not utf-8. You have to know what encoding it is exactly. What's sure, is that it's not a ü in any known encoding: $ for encoding in $(iconv -l|sed -e '1,/^$/d' -e /:/d -e s-//--) ; do printf "%24s " "$encoding" ; ch=$( iconv -f $encoding -t utf-8