A Reversible Semantics for Erlang

A theory of reversibility for Erlang

“…Finally, as mentioned in the Introduction, this paper extends and improves [27] in different ways. Firstly, [27] only presents a rollback semantics.…”

“…Finally, as mentioned in the Introduction, this paper extends and improves [27] in different ways. Firstly, [27] only presents a rollback semantics. Here, we have introduced an uncontrolled reversible semantics and have proved a number of fundamental theoretical properties, including its causal consistency (no proofs of technical results are provided in [27]).…”

“…Moreover, the process' histories also include items for the applications of rule Sched, which makes the underlying notion of concurrency unnecessarily restrictive. As for the rollback semantics of [27], besides the points mentioned above, it only considered one rollback for each process, while sets of rollbacks are accepted in this work. Consequently, we have now reduced the number of rules required to undo the sending of a message or to undo the introduction of a checkpoint, so that the rollback semantics is simpler.…”

“…This paper is an extended version of [27]. Compared to [27], we introduce an uncontrolled reversible semantics and prove a number of fundamental theoretical properties, including its causal consistency.…”

Playing with Bisimulation in Erlang

“…In practice, this amounts to consider that each process is potentially run in a different node. An alternative definition ensuring this restriction can be found in [27].…”