A misguided attempt to encode Unicode
into bytes in a way that avoids any characters outside the range 0x21..0x7e (and also avoids some characters inside that range), to make it safe to use on computer systems built in the 1960's.
If you want to see just how ugly this is, look here: ftp://sunsite.doc.ic.ac.uk/packages/rfc/rfc2152.txt
Don't even think about this. Use UTF-8 instead.