@stilkov If I say "network type system" what do you think of? Trying to do a comparison for a talk.
2
Thanks! I want to consider whether session types (static checking of state transitions) fall prey to the same problems as CORBA et al.
1
Interesting, so you’re considering them for the distributed app’s state as a whole, not as internal to e.g. the server?
1
Session types let you define a global protocol, and project to types that check each participant's transitions.
3
Then I’d argue it suffers from the IDL problem, too, unless checks are completely local to each participant in definition and implementation
1
The problem I have is with metadata that describes an ideal that may be (and will be) out of sync with reality
1

