smh (04/13/83)
Several of us at MIT have recently been giving this problem some thought, but no real solutions yet. For now, you might think about the following additional complexities: 1) It would be nice if checkpointing of large jobs could happen regularly, not just when actively commissioned by a system shutdown. 2) Remember that it is idiomatic un Unix for several processes to communicate via pipes. It would be nice if the state of such a collection of processes, pipes, and perhaps even signals could be restarted as well as single `simple' processes. I fear it ain't gonna be simple. - Steve Haflich MIT Experimental Music Studio ...!genrad!mit-eddie!mit-ems!smh