>Even though you are using insert into, an implicit lock is performed on the table header. It is very fast so the odds of someone hitting it at the exact same time are remote, but they fall within the "probably will happen at some point when the VP wants to get into your system" time frame.
>
>Try locking recno(0) and then issuing the insert. This gives you something tangible to test for and you can control the behavior.

What would be the test command and the lock command? Since it is a free table is there a test for record locking or is the test a file locking test?