About the GUI, since I would like to run this video phone application on set-top box (Linux based) in the future (i.e. do video chatting via on big screen TV using set-top box via HDMI interface) so the interface/GUI would look like "appliance-like" interface rather than standard windowing interface like Qt, wxWidgets & Gtk+. Good example is Mirial Softphone which seems to use OpenGL/Direct3D for user interface with it's own simple custom written widgets (i.e. contact list, call history) on top of OpenGL/Direct3D. The only part of Mirial Softphone that use standard Windows window and it's standard widgets is the "Configurations" window. In my case, I would use one of windowing library (Qt, wx, Gtk) for that window. So I think I should go this route by using OpenGL directly (nice that OpenGL is supported in all three platforms (using Mesa 3D library?) So it will be some extra work to write some custom GUI widgets on top of OpenGL (maybe someone has written some of them?) Any better way of doing this? Since I would want this video phone application to have "unlimited" (as much as network bandwidth, memory, CPU load allows) "lines" (like the two line phone with "A and "B" buttons), should this Haskell program run other copy of itself, or should it spwan/run other thread? For example, someone do video chat call and I answer it, then someone else also call and I answer that one and I would be talking with two (or more) people at the same time using shared webcam input and shared network connection. What is ideal way to deal with this in Haskell? Thanks! Jamie