> It may make sense in YOUR mind not to buffer the source table for an updateable view (ie "double buffer"), but. . .<

What does MY mind have to do with any of this?

>1) in a multi-table view, just where would the data come from?. . .

Irrelevant as only one base table can be updated.

>2) what makes you so sure that VFP doesn't do so???

>
>I've got to believe that it, in fact, DOES "double buffer" view data. More fuel to the need for more documentation on this topic.

I'm not so sure about that. It may be using OS or NETBIOS type lowlevel locks for base tables "behind" updateable views. Windows is read/write buffering for sure (and don't get me started on THAT topic). I was talking about the need to explicitly buffer base tables.

And you are 100% right about documentation. MS or somebody should do a "White Paper" about how that is handled internally; it'd sure come in handy when planning a topology.