On Wednesday, 10 September 2014 at 11:11:08 UTC, Marc Schütz wrote:
I think there should be an explicit guarantee for those protocols that can provide it. Otherwise it would just be an implementation detail that cannot be relied upon, and people always have to assume the worst.
Agreed. I absolutely don't want to get into a situation where people are trying to create an ordered protocol on top of an unordered one, like how people occasionally end up trying to implement TCP via UDP.
