>>>>Whatever a view can do, CA does, plus twice as much with half the overhead.
>>>
>>>How is it half the overhead?
>>>
>>>If I have to see, say, the customer name associated with an invoice as well as the amount, why is a CA more efficient than using a view?
>>
>>1. don't need the dbc
>>2. can do it all in code (prgs), and you only need pass the handle. Don't need the named connection (aka DSN) to edit it.
>>2.b you can subclass your CA's at will, can't do that with views.
>
>1. I don't need a dbc to use a SQL view from VFP.
>2. I create a handle once when opening the program and use it for all SQL operations. I started with DSN's so I still use them but I've used DSNless connections (all .prg) in later apps.
>
>3. Never saw a need to subclass a SQL view.

Oh, I meant comparison with VFP-side views. As for SQL views, there's still the matter of what you use VFP-side to connect to them. Even there you still have the choice.