If I remember correctly, John was using pessimistic record buffering. If I'm not mistaken, pessimistic table buffering locks the entire table while the user is editing. So pessimistic record buffering or optimistic table buffering are probably the best options.

>John,
>
>If you TABLE buffer the file you can move the record pointer without the blank record getting commiteed until tableupdate() time.
>
>>You hit the nail on the head... I was moving the record pointer in a refresh method... Now if I am in the add mode it's smart enought (because I told it what to do) to not move the record pointer until it's time...