I'm not sure what you're arguing here. Yes, I described a very specific type of optimistic concurrency control that can be layered on top of Calvin to provide s

I'm not sure what you're arguing here. Yes, I described a very specific type of optimistic concurrency control that can be layered on top of Calvin to provide support for dependent transactions. (The "optimistic" is in the name of OLLP, after all.) But the underlying Calvin protocol is not optimistic, as it acquires locks. The result is a hybrid scheme that is neither fully optimistic nor fully pessimistic.