>>Are SQLCONNECT and SQLEXEC xBase? One passthrough yields a cursor don't "you" usually xbase the cursor?
>
No, they are not xBase commands at all. You're getting a dataset in the form of a cursor. Then you manipulate this using xBase at the client side. This is what I'm saying you should do.
I do that all the time - did my remarks leave you belive that I did not work with SQL or passthrough cursors - been doing it for years - since S6, anyway!?

In the same way, instead of "using" a table, you can call a web service (for example) to get an XML resultset that you can convert into a cursor, then manipulate it locally. But there is a BIG difference.

The niche markets I look at are buying web servcies. They want bean counters that offer 'automeation' with other desktop apps or third party objects (like propietary financial feeds or oil storage tank "sonar" [analog to digital] readers!:). Some have web services - but the bean counters and accounting are something they "hide" from the web. Nobody's business but theirs!

I work with XML a bunch too! Not all XMLs are tabular. Some will break them XMLTOCURSOR engines.!

Please, give me example of who "all the NET guys" are. What several VFP people working on .NET said (including me, sometimes) is that if you have some very stable data access (or even domain logic) components, and you need to expose that in ASP.NET, for example, interoperate is easy and works "quite" well.
There were posts a while back - and articles 'integrating VFP with NET', etc, where it was suggested that a VFP COM. selecting a passthrough cursor and then converting it to ADO was one method of collecting data in NET! I don't open the UT Mags that have "NET" all over the cover - but I have seen some. In fact, I have to chase down a lot of fox stuff on other newsgroups because UT has become so NET or seems to push the views of certain consultants that have their banners on the start page! UT aint as foxy as it used to be!:)