>>>could be untyped datasets for all I know. Are you using MM or nHibernate or anything else,
>
>It's even more primitive than untyped datasets (gasp). It's not datasets at all. It's datatables (shudder.)
>I treat the datatables that the ExecuteReader returns exactly the way I treated the cursors that SQL PT returned in VFP. The SP's do all the denormalization etc, so the datatable is really all I need, just as the cursor was in VFP.
>
>I didn't use anything like MM or nHibernate or anything like them with VFP and don't use them with C#. I've read about them and attended workshops on them but the marginal gain for the client compared to the marginal effort I'd have to put into learning and using them doesn't compute for me yet.

I agree. I keep it simple and maintainable w/o adding more layers.

>When you look at the huge data manipulation power we have at our disposal with SQL Server, it really makes me wonder why we have to keep adding more layers of technology over it, so there will have to be more compelling reasons for me to do that than I've seen so far.
>
>I've only deployed .NET 3.5 apps (it's great to come in later, when all the bad stuff is gone) and I have 3 in production at clients now and they slid in without a hitch.