The implementation we have now is good for showing one Paxos run. I think we should prepare it for unsupervised interactions --- i.e., 1.auto-increment term number at proposer and 2. auto reset states for another consecutive run (it would be even better if we can handle consecutive run without resets)
(1.) is required to implement a live-lock scenario which is an essential part of Paxos.
(2.) is good for smoother interaction with the algorithm and it will look much better to supervisor and at the design fair (big marks)
Please leave comments about what you think about it and how you think they should be implemented. @rfairley @zzlyn @AbhiPat123