>2. How to 'distract' the user for a few seconds while large files open, so they don't think diddly is happening, and try opening the exe again.

A common method is to use a "splash screen' that shows company logo, or some sort of identification that the program. It is the *first* thing that is done *before* any actual processing. Such a thing isn't unique to PCs -- I'd seen similar sort of "loader" software on videogame consoles. On one game, used a copy of Galaxians as a "loader" -- funny bit is that it was actually playable. If you simply sat there and waited, the actual game would load and start. If you started playing Galaxians, it would let you continue to play until you quit before it continued with the actual game software (which by that time was completely loaded in memory and was ready to start)..

>4. The app I am supporting runs on a server - each PC is mapped to it, and runs the exe on the server. Would it be a lot faster if the exe were to run on the local PC, and just have the data mapped? And what are the demons to look for in that?

Are you sure that the EXE loads on the server? Just because the EXE is hosted on the server doesn't mean that it runs on the server. On the other hand, if you're using a configuration where users are running the application through some sort of remote desktop, then it is likely running on the server.

Putting the EXE on a local drive does save the amount of time it takes to access the EXE, but if I understand the situation you describe, the problem is more to do with the time it takes to pen large tables -- which putting the EXE locally on the workstation does NOT address.