David,

I basically agree with your arguments, yet all we've said in this respect is mostly speculation. However, to claim that the OLE-DB provider for Fox renders results that much more different than Fox itself is a far stretch, to say the least. A demonstration to the contrary would be most interesting nonetheless.
Actually, if that were true then perhaps we should access the data through the Fox OLE-DB from VFP itself :)


>Sorin,
>
>>That would be really surprising, and I don't think it is true! Optimizations, if any, used by OLE-DB would have been incorporated into VFP by now.
>
>I don't think we can conclude that without further testing in various scenarios (and perhaps by getting more info from the devs who did the work). Remember that the OLE-DB driver is a stripped-down VFP engine that is quite different in many ways, including the removal of many, many commands and functions.
>
>It also does not just sit there running itself with all the overhead of a VFP app, but is instantiated by a call from another process somewhere for a limited set of possible actions.
>
>It is not unreasonable that the OLE-DB driver might be more efficient in how it handles disk access, memory management, or a whole host of other factors. It will take some time, but I suspect that more info will become available over time about this, as more people have a chance to test it.
>
>Drawing a definite, sweeping conclusion based solely on Kevin's comparison is not enough to establish the facts yet, but his scenario is a good starting point for examining the possible differences.