An index an you descible would not be reliable. I might have a firstname + lastname that when concantenated look like another firstname + lastname, since there is no seperation of the fields.
Jimmy Evans = Jim Myevans--not that it's likely.
I still think the best advice is never TRIM in index expressions. If you have Values from the interface to match, PADR() them to the proper spacing.

>You should only use TRIM in an index expression if you are concatenating 2 fields. Like Index on trim(FirstName) + LastName. Otherwise, TRIM is not needed. You should also check settings for SET EXACT, SET COLLATE [among others] because these are scoped to each datasession. See SET DATASESSION in VFP Help for all SETs that are scoped to the DS.
>
>Alos, on character data, you may consider using the UPPER or LOWER function to make SEEKs and LOCATEs case insensitive. Of course the parameter you are pasing needs to be converted to UPPER or LOWER, depending on the index.