The case-insensitive collation option sounds the most seemless.  I might have 
to choose character sets more carefully - the DB I tested with was ISO8859_1 
for a default, and I couldn't find a CI collation for it (maybe I needed to 
look harder).  The second is basically a simpler version of what I did - would 
probably work, but I do harp on certain details, which is why mine looks like 
something concocted by a mad scientist.  The third option looks like a bit of a 
kludge at first, but it might be quite workable with a few refinements.  
Thanks.-MB

Reply via email to