>I haven't tried closing the source tables in a real working environment. I've only done it on an experimental basis to learn what was going on.
Same here.

>I would look into what's causing the network to slow down, rather than trying to close the tables. It sounds like that's more of a problem.

Sound advice - but I have not always the budget or the net to put my hands on. Having an external switch to modify closing behaviour would be the first step to try "via phone", and could also be viewed as "data-gathering".

>If you try the technique of closing the tables, you might find yourself fighting with how VFP wants to work.
One of my first toughts as well. The time difference in opening tables was very great, but as long as I don't get overboard and try to keep a couple of hundred tables open, vfp runs better with the tables open [especially in single user scenarios <g>].

Still, from another project I was told that nearly every operation took 2 to 3 times longer if the table was opened from more than 1 WS concurrently, even if one WS had no action at all. And, as a side benefit, the closing approach would minimize the window of opportunity to create a problem on shared tables, since the tables are open only for very short times.

One of these days I'll try these waters

rgds

thomas