Hi [ I'm making this a separate thread as it was previously in the 'mdr20 value changed' thread which is a mostly unrelated bug. ] Currently ascii and latin1 are special in that any character that can not be represented in those character sets is transliterated to ascii/latin1. When using any other code page then any character that cannot be represented is replaced with a question mark. I've created a patch that is currently in the 'transliteration' branch which makes it possible to use any code page and still get transliteration to ascii for any other character that is not in the code-page. I'll merge it in a couple of days if no one finds any problems. ..Steve