>Thanks for a very informative answer.
>If Table Buffering is locking the table (like FLOCK()), during TABLEUPDATE, does this mean that I cannot issue a TABLUPDATE if a second person is also using the table in a shared mode?
>Or does TABLEUPDATE try to lock each record to be updated one at a time like RLOCK() while performing the update?

If you use optimistic table buffering (5) then other users can still access all records. Changes are written in buffers not in actual table.
Locking is happening on saving (by tableupdate()).

In case of colision (someone else already updated one of the records you tackled) VFP will flush error 1585 which you can trap and act accordingly.

You might not even need transaction to revert table to its previous state
since you are amending single table - tablerevert() might be just enough for what you need.

However, It pays back big time to hv saving process well organised, having tableupdates always wrapped together into transaction fold.