One option that hasn't been said is to code a custom filter that spits out a case-preserved token and a lower-cased token --per word. The common case is to return just one token since most words are all lower case anyway. This option is better than the other ideas I've heard here so far because the size of the index would be smaller than the dual-index and dual-field strategies suggested.
~ Dave Smiley -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
