Thank you for the feedback. I've added a command-line qualifier to coerce encoding to the feature request list.

To detect encoding, BC checks for a UTF-8 byte order mark. It also checks for an XML or HTML encoding tag at the beginning of a file. If none of those are found, then it uses a third party library to detect the encoding of the beginning of the file. I don't know the specifics of the algorithm in the third party library.