Given a system that produces and consumes serialized protobuf messages, how can we be assured that our code is equipped to handle older or newer messages?
Let's talk about protobuf's strategy for backward and forward compatibility as well as pitfalls specific to this strategy that you should watch out for.
