>To be honest, I think this depends on too much factors so there can't be a simple rule. It might be the first starting point, but what finally counts is the way you access data.
>
>Christof


Thanks for that perspective.

Peter