Thanks, Bob. I guess if I had thought it out to its conclusion (that is, that views are just SQL statements -- i.e., code) I would have realized that what I'm doing is fine. <g>

>>I am developing a networked app in VFP 5.0a. As I have been designing my databases, I've been operating on the assumption that while my tables will all reside on the machine acting as file server, the local views for data entry will all actually reside on the client machines. For this reason I have created a database called LOCALVIEWS that contains (you guessed it) nothing but views. When the network is set up, each client will get this LOCALVIEW database, and I will point the app to the file server.
>>
>>My question is: is this the best way to do this (or the worst)? Should the views also reside on the server? Are there any performance implications of one over the other?
>>
>>Thanks.
>
>Views do not really exist. The local machine will see them as tmp files in the temp directory but even these may not exist. A view is really just a sql statement that selects some subset of data. This subset becomes a temporary cursor that will reside on the local machine either in memory or as a temp file. With that said, it is very useful to program to views, one of the best reasons being, if you change to SQL Server or Oracle data, you merely have to change your view definitions (from local to remote) and your system should still work (more or less!)
>
>You cannot make your view reside in any particular place, unless you moved the users temp files to the network which would be a very BAD thing!