>Another good way is to have the index be UPPER(field1). Then when seeking just use the UPPER function on what you are seeking with. This ensures that your index is not going to be case sensitive.
True, but you have overhead twice: first, your index is based on a function, second, you must use the function in each SQL select and each Seek if you want to use this tag. The overhead of this is probably bigger than the overhead of a native VFP's way to take care of non-machine collating sequences. Of course, this should be measured, before we can make any educated guess at the speed issues.