>I found the paragraph that you're quoting from:
>
>
> However, identity columns are no panacea. Retrieving the last value is normally handled by checking the value of the @@IDENTITY global variable. But it may be difficult to get the correct value of this variable, as insert triggers might have caused other identity columns to be incremented and thus would return the wrong value for @@IDENTITY. Identity columns can also cause a problem when databases are replicated. If you have the choice, you should avoid identity columns when you design a database. Learn to use them, though, because you may find yourself working on databases that use them.
>
>When this book was written, SQL Server 2000 had not been released. SQL Server 2000 has added a couple of functions that make using identity columns a little easier to use. Specifically I'm referring to the SCOPE_IDENTITY() function. Now, I think that you'll find that this function works based when used from within a stored procedure.
>
>And speaking stored procedures, I belong to the "stored procedure" camp. All data access is done using stored procedures, no ad-hoc queries.
>

Thank you very much for the explanation and for suggestion with "stored procedures".