It is not impossible, but a lot of work. And if you want to do it correctly you would have to support UTF-16 (BE of LE) and UTF-32 (BE of LE) as well. You can't expect someone to start writing utf encoders and decoders every time he needs a fast parser.