Under C1 accelerators and some other USB HID activity require the cursor to be on a graphics window. An exception would be the mouse or keyboard interacting with other gui windows such as button pushes or entering file names. For C2 it would be good to have a dedicated input channel independent of the cursor position like the 3dconnexion. The general trend I see with researchers is dual display. One handles graphics and the other gui controls. This gets tricky when s3d is used and the cursor. Is not exactly what. you think (2 visible cursors one active the other inactive) I manage this now with a dedicated chimera extension user gui buttons that rearrange the screens; and with controller mate I can have preconfigured-window-gui-location accelerator commands sent to activate the buttons, at the same time controller mate can set cursor positions and do button pushes. Works for Mac, not an option for PC or Linux. I guess I could set a frame trigger and check a dedicated tcp port, but I am hesitant creating a kludge, because of the underlying strategic design issue and messing with the event loop. Matthew Dougherty National Center for Macromolecular Imaging Baylor College of Medicine ================================================= =================================================