Okay, I found the problem: length of code. It takes a minimum of 17 bytes of XPL to avoid this problem.This encodes fine: 12345678901234567 But this creates the going-past-EOF problem: 1234567890123456 Hardly something needing correction.So, the clear application of Mill's Methods here is to add bits of the encodeable file's code to the unencodeable file, and see where the change occurs. I'll do that when I have more time.Harry Binswanger hb@xxxxxxxx