Even if this is completely unrelated to ConTeXt, allow me to mention this since I find it so funny: The message to which I'm replying has the following weird subject: # Subject: Re: [NTG-context] Commands based on previously defined commands, # DA with fixed values for šrguments Notice the word "DA" at the beginning of the second line, and the word "šrguments"? This was obviously not in Vedran's original message: # Subject: [NTG-context] Commands based on previously defined commands, # with fixed values for arguments Let me venture an explanation about how this happened: when replying to the first message, René's e-mail transformed the subject as such: # Subject: [NTG-context] # =?UTF-8?Q?Re:__Commands_based_on_previously_defined_commands, # =D=A_with_fixed_values_for_=9arguments?= This is quoted-printable, where all the characters but the "basic" ones are quoted using their code point in the current encoding (here UTF-8): hence the end line characters (carriage return + line feed) have been rewritten as =D and =A respectively, in hexadecimal (it's 13 and 10 in decimal), and the tabulation character as =9 (for some reason, there was a tab before the word "arguments"). But this is actually not correct, the character codes should be formatted using two hexadecimal digits (hence =0D, =0A and =09). This is probably why when replying to René, Vedran's e-mail agent has interpreted the "=D" and "=A" sequences as plain 'D' and 'A', and, upon seeing the sequence "=9arguments", it has use the letter 'a' in "arguments" as an exadecimal digit! Which it has interpreted as a Windows-1252 character (Microsoft's extension of latin-1), where 0x9A is 'š', thus producing this rather weird result. This is a particularily disturbing example of mojibake (http://en.wikipedia.org/wiki/Mojibake). Arthur