
The original version of MMA was written to use Python 2.x. In those
 good old days, Python didn't care much about the encodings used in
 a text file. It just read everything as ASCII text (well, sort of).
 Actually, the whole thing was a mess.

In Python 3.x a complete revamp of all this has been done. Now we have
 to decide what the "encoding" of a file is BEFORE opening it. This is
 required, mostly, for characters outside of the ascii range 32 to 127.

MMA does accept some "extended" characters $B0 (176) and $F8 (248) which
 are used by the roman numeral module for Diminished and Half-Diminishned
 notation. 

Now, this will all work only if we open the file in the right encoding. To
 make things "pretty much work" we use the encoding of "cp1252". If this
 is a problem, please let us know!

