Dennis,

Okay, here's the deal. Anytime you open a table, VFP essentially reads some portion of that table (sometimes the whole table if it's small enough) into memory on the workstation. This is because VFP data engine is contained within VFP, not as a separate component. The ODBC driver is no different - it is essentially the data engine part of VFP. The ODBC driver must live on a client computer - therefore, the data must be read from it's location on the network and put in the memory of the client computer.

The VFP ODBC driver is different than most in this regard - if you were using SQL Server, Oracle, or DB2, then the processing would be happening on the server. The help pages that you have read are obviously talking about using something other than the VFP ODBC driver.

There is no issue with 100 users opening the table simultaneously from your server, as long as you SET EXCLUSIVE OFF or use the SHARED clause to the USE command. Yes, you would call this file sharing, but that's what VFP does with its own tables. If you want to use VFP data in a client/server fashion, you would have to write a separate application that would reside on the server - this application would take requests for records and return them to the caller in some way. That's how SQL Server (and others) work.

One more thing: about the only good reason to use the VFP ODBC driver from within a VFP application is for the situation where you want to always use ODBC, such as an app where the data could be stored in VFP tables OR in SQL Server tables. I don't think you have that as a requirement here, do you?