Hi George,

>I would do so, however, at a time when the overall system was being re-designed. If, however, I found quantifiable reasons for not doing it, then I wouldn't. This is certainly analogous to what's happening with VFP now. Porting to a 64 bit OS involves more than simply adding new header files and re-compiling.

And you would so if it pays off. I have no doubt that Microsoft would rewrite the kernel, if they gain more by doing so than they would gain by doing something different (as long as there are no political issues involved). It's just that I can't see what Microsoft would gain. I don't have the source code of VFP, maybe it is really impossible (or difficult) to take the existing code over to a 64 bit OS system and I really hope so, because then there's a business reason to re-write the kernel. But if it's easier to take over existing code, Microsoft would probably do so, no matter how great a re-write would be.

>While n-tier certainly has its place, there will still be solutions requiring a single tier.

Don't tell me, I mainly write single-tier applications without SQL-Server backend. But that's not the way Microsoft seems to see the future of software. Unfortunately.

>Lastly, while SP3 does bring new potential for the middle tier, I wouldn't say it does so exclusively. The Datasession object can be utilized effectively in a single tier solution.

Sure, especially because it doesn't have the bugs of a formset and the limitations of a form when you want to create a private datasession. But even in the description Microsoft said that the new session class is intended to be used in COM servers to create non-visual objects with a private datasession.

Christof