This is a subjective discussion so we can express our dreams and hopes here. A "topic" must be like a task with point to have as end result a software poduct. A "topic" must be mainly about "Software engineering", "Algorithm and data structure concepts" and perhaps "Design patterns".

I mean let us try to look what is not already there? What can be developed in fiew month and give a breakthrue / start a new leap / show somethig not realized before in science of f multiple computers programming?

What i see is already there:

• LAN / wire and other infrastractural programms for connecting on device level
• MPI/ Bit torrent/Jabber protocols / APIs / servers for messaging on top
• Boost and analogs on evry OS in most languages for multithreading
• there are lots of CUDA like on computer frameworks for fast calculating on computers GPUs

What I personally do not see out there is a crossplatform framework for multiple processes interaction. Meaning one that would allow easy creation of multyple processes running in paralell inside one hoster app on one machine. In level not harder than needed for threads creation (so no seprate server apps - just one lib doing it all) Is there ny such lib and what can you propose for research topic?