On 4 Mar 2004 at 20:52, h_urlaf wrote:

>> If I were you, I'd put each translation into separate database.
>> this would solve problems with connection charset and sorting.
>
>Wouldn't using UNICODE_FSS accomplish that?

Yes, UNICODE_FSS is quite suitable if you need only keep messages
and not to use UPPER() and language-depended sort order.

SY, Dimitry Sibiryakov.