While preferably this would be done in the data tier, we do it in the business tier because our app can have multiple backends. By having it handled via code at the business level the backend is not germain. It suits our purposes fine.

Of course you could create a separate 'logical' data tier in code to handle this type of thing. You could create your own data objects and place your appropriate code there depending on the backend type. Probably some would do that. It adds another level to maintain though but would work as well.


>Hi everybody,
>
>We just had a discussion between developers about how do we usually handle lookup codes. E.g.suppose you have a table, which has several coded fields, which could be ever empty (NULL) or valid codes. Do you create a trigger (restrict) for Insert/Update/Delete or do you simply handle this in the application itself. Also I'd like to hear about your experience with different back-ends. I see advantages and disadvantages of both approaches and would really appreciate your input on the matter. So, basically I'm conducting a pool.
>
>Thanks a lot in advance.