Ken,

>since I got sick of this and objectively <bg> stepped back and determined that I won the debate

Huh? oops, wait, I see the <bg> now. I am glad your objectivity is so wonderful, and also glad that you are fortunate to have won the debate *smile*

>but - it's like acknowledging that continueing to use 2.6 style gather/scatter stuff is "OK in some circumstances" when those circumstances are based purely on time/money/etc. Sure it works. Sure it may be necessary in those circumstances. Is it the "best way" - really? I don't think so.

If one is the type of person who wants to always use a single methodology, then I would probably capitulate (this is just my opinion...) and say that SPs are the "can do everything" answer. But for "best way" methodologies, each scenario is going to have different needs, and different tools may be called for. We have all agreed that for small apps or apps with multiple backends (usually VFP native and a server-based solution), RVs have an obvious advantage in that you don't need to write any data-logic twice, and RVs are very quick to get up and running (especially with tools like eView). Yes, you can argue that you have written various and sundry classes that handle it, and once done, they work forever. But by that argument I could say, hey, why don't you write everything in assembly code -- sure, it takes a while, but once its done its done and it will be fast. I know that is a crazy scenario, but I find the craziest scenarios to be the most illuminating.

We are fortunate to have the RV vs SP choice. I propose we all simply use that choice wisely.

JoeK