>I use a little different approach.
>
>I install the exe on the local computer. The exe consists of a splash screen with a timer which then fires up the application (compiled as an app) which is located on the network along with the data files. As people access the application I log them to a user table and clear the log entry when they exit. This way I will know if anyone is still acessing the application if I need to make a change to the app on the network.

I tried logging the users too, but it turned out to be too much trouble. What worked, in the end, is to create a timer that checks every 2 minutes if there is a file "maintenance.tmp" on the server.

If there is, the applications OnQuit() method is invoked. Of course, there is a slight chance that the user has unsaved data, and therefore sees the "Save changes? Yes / No / Cancel" dialog, instead of immediately quitting. In theory, the application won't quit if the user left his desk unattended under these circumstances. But this didn't turn out to be a serious problem in practice.

Hilmar.