On Wed, 29 Mar 2006 13:13:40 +0000 (UTC), Martin Gerner
<martin.gerner@nospam.com> wrote, quoted or indirectly quoted someone
who said :

Exactly. It is a messy problem.

On Wed, 29 Mar 2006 13:13:40 +0000 (UTC), Martin Gerner
<martin.gerner@nospam.com> wrote, quoted or indirectly quoted someone
who said :

If you have control over the creating of these files, you could put
the encoding on the front of the file  followed by a  \n. That would
make your job much easier. Or you could tell everyone to use UTF-8
which would make the problem disappear.

You might also do it by tracking the source of the file. You figure
out manually which encoding each source uses over which date range.

The habit of not recording the encoding goes way back. The idea was
documents were local and all encoded the same way. You did not
exchange documents with others, of if you did, you exchanged a whole
tape full all the same, so again the problem of identification did not
come up.