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