Validating data in access
In N-Version programming several teams of programmers all write code based on the same specification, all versions are implemented and a decision algorithm checks if the different versions produce the same result in real time, signaling an error if they are in disagreement.
The method is similar to writing automated tests where you write a second program that checks the result of the first and raises an error if there is a discrepancy.
Many of us have also been taught that one should avoid duplicating logic but instead should re-use existing routines. Some safety critical systems like aircraft control systems are based on .Whereabouts in the application should the business logic of data-validation checks be made?The advantages of a layered approach to validation that includes database constraints, would seem to outweigh the disadvantages. As a database designer, I see the database as being the ultimate arbiter of the validity for data.If we just rely on the database (or the middleware for that matter) to apply validation then the user won’t know that a problem has occurred until they send the data over the network.This is not just user unfriendly, but may have performance implications too.
Every developer must remember to only update the database only over the designated routine.