Jean-Marc Valin wrote: > > What do you mean by *parallel*? Why distribute if you don't run stuff in > parallel? Or I just got a different meaning for the word? When I think of distributed parallel computing, I think of SETI at Home and similar projects. They automatically break up a single data set and send the pieces to identical processes. With Piper, the data and the processes can be different. I'm not saying that Piper can't do parallel processing, but it is meant to be much more flexible. For one, a user can create a sequential (serial) workflow path and not be limited to a parallel one. And I imagine Piper would be good at tying together parallel systems. Jeff