>Most products that implement optimistic locking do so by checking to see if the row values have changed since the editing began. When the row is read, the values are stored twice. One set is used for editing and the other is used to determine if something has changed.
>
>When the product sends an update it will look something like this:
>
>UPDATE table
>SET column1 = new_value1, column2 = new_value2, ... , columnn = new_valuen
>WHERE column1 = old_value1, column2 = old_value2, ...
>
>You can emulate this behavior with a stored procedure but you'll have to carry the original values since stored procedures don't have state.
>
>-Mike

Thanks for your help. A co-worker of mine has a question on returning an ADO recordset from a stored procedure. Is it as easy as setting a recordset object equal to the return value of a command object that calls the stored procedure?