>
>My problem was with VFP being released without a framework and with Tastrade being offered as "all you gotta do is". It was all great demo-ware but I don't think the whole ORM issue was thought through at all for real-world application development. I think that has always been the elephant in the room in the OO - Normalized table model and still is a mystery to a lot of developers - even to the extent of understanding the issue.
>
>That said, VFP definitely offered a brilliant underpinning for serious tool builders like Feltmans, Strahl, McNeish, Black, Hennig et al to build the stuff VFP needed. But the mindset in the community was to never spend money on anything beyond the basic product so most VFP development was done without Stonefield, DBCX2 or a serious professionally built framework. Had Microsoft subsidized a lot of those folks to build the product there could have been something special in the box.

I thought at the times that VFP was actually overcrowded with (quiet complex) frameworks back in mid nineties. Most of frameworks back then were trying to handle way to many things, becaming very heavy, while leaving very little space for customisation by end developer.
That left some fairly common problems/questions very dificult to resolve, on more or less every single framework back then.

For most ordinary folks, frameworks were way to complex to understand and way to restrictive. They were trying to reach holly grail of perfect OOP/multitier design rather then help people deal with more ordinary daily problems and tasks. Back in 1995-7 majority of existing applications were fat-client apps (financials, heavy reporting etc) whereas frameworks were mostly offering complex multibackend/tier CS with dbf being almost routinely undermined . Aloha! Most of dos apps WERE using dbfs back then.

Instead of focusing on reaching easy, corruption free DBF handling (btw viable with buffering/transactions) and simplest possible app design which would help people stand on their two feet (in back then new VFP environment) masses were pushed way ahead of the time in some perfect OOP / multitiers etc. Here at least, this was prety much detached from reality. It is IMO one of the main reasons why frameworks never made huge sales, NOT because FoxPro community was 'cheap' as some like to claim.

I worked for very small company which bought every single framework out there, and still ended up rolling their own at the end.
Why was that ? Cheapness ? Lask of skill ? Hell no! Simply, owner could not resurrect his own old dos app look & feel in any of them. (You know very well by now how much users tend to stick with their little comfourts & habbits) So he rolled his own at the end that ressembled quiet well behaviour of his own old applications. Not everybody could do that very well, so many ordinary folks ended up building very lowsy unstable applications with their own little quasimodo frameworks, crushing, corrupting dbfs etc. (Hence VFP/DBF reputation )

* * *

Now regarding foxpro being positioned as middle tier, I share your view quiet a bit. Back then I was (naively) thinking that VFP will be able to natively use MSSQL one day, and this way really become mainstream development tool. I don't believe that there is any other
technology out there that was so close to achieveing perfect database access capabilitis.
Yet all this was simply thrown down the drain, for some perceived threat to MSSQL sales which never existed, and some perceived higher profits that were to be made by pushing NET/MSSQL down everybody's throat.

It worked for now, but I doubt it will work for very long time. Market which is now trapped into heavy price/low benefit technology bubbles&clouds (almost by force) will eventually correct itself. It always does.