You raise a point that I don't have a full answer for. If the database column permits NULL values, and the column currently contains data, but now needs to store a NULL, I'm not sure. There's a difference between a null by default and an explicit null. I'll have to think about that one. We ran into that in a survey application where "no response" was an entry - we went back and forth on whether to store a -1 or NULL.You'll run into this in any SQL Server table with a datetime field.
It would be a non-issue if you used change tracking and selective updates. That's my point. You can still have all your pessimistic functionality if customers truly believe that's best. IOW you can handle *any* imaginable scenario. That's one of the stated raisons d'etre of the typed dataset, isn't it?
Great discussion, but I'm going to have to vanish for a while. Thanks for helping get my mind around some of this stuff.
"... They ne'er cared for us
yet: suffer us to famish, and their store-houses
crammed with grain; make edicts for usury, to
support usurers; repeal daily any wholesome act
established against the rich, and provide more
piercing statutes daily, to chain up and restrain
the poor. If the wars eat us not up, they will; and
there's all the love they bear us."
-- Shakespeare: Coriolanus, Act 1, scene 1
