The character sets that can be used with the MIF conform to one of the following standards:
If a MIF is created according to the Unicode specification, the first two octets of the MIF file must be X'0xFE' and X'0xFF'. If the service layer does not detect these values in the MIF file, the service layer treats the file as an ISO 8859-1 MIF.
There are four classes of tokens: keywords, integer constants, literal strings, and separators. Blanks, tabs, newlines, carriage returns and comments (known collectively as white space) are ignored except as they serve to separate tokens, such as adjacent keywords and constants.
When interpreting tokens, the MIF is not case sensitive, unless the token is a literal string enclosed within double quote (") characters. Case is retained for the contents of a literal string. Literal strings separated by white space are concatenated and stored as one literal string.