Lists Home |
Date Index |
Mark Baker wrote:
> Take it from a protocol weenie, you're in a for a big surprise. If
> you've got no a priori agreement about the task at hand, you're not
> going to accomplish much.
> RPC is a good example of this principle, but it's not difficult to
> replicate the problem for other interaction styles. What you describe,
> if I understand you correctly, is a good way to do it with messaging.
Just like RPC, that strategy will work at first. In fact it may never
stop working. Other protocols that do more will tend to become more
popular and will tend to be more important but if you could definitively
prove that this strategy would NOT work then we wouldn't be dealing with
the eleventh incarnation of RPC (which we are). But re-inventing the
transport layer is fun, like writing parsers. So there will always be
people inventing new transfer protocols, application protocols and