Re:Rage Against the Finite-State Machines
by mwilcox on Sunday April 03, 2016 @07:57AM
(#11415813)
I'd say those are features of agent-based systems as opposed to problems. But it depends on the application
[ Parent
| on HN ]
Features, yes, provided one has time to actually think of the protocol and express it in terms of some formalism (CSP, TLA+) to at least informally sketch a proof of major global properties. Or do some extensive testing (https://aphyr.com/posts/291-jepsen-zookeeper https://aphyr.com/posts/316-jepsen-etcd-and-consul ).