A distributed type system would be a cool idea though.
1
2
@ctford no, it would be an awful idea, see: XML schema.
1
@jimwebber That's certainly a good counterexample. Do you think the failure is inherent to the idea?
1
@ctford yes: it increases coupling across boundaries that should be explicitly honoured. See also: CORBA
1
Replying to @jimwebber
@jimwebber @ctford +1. See Mr. CORBA @stevevinoski explain this most lucidly here: infoq.com/interviews/vinoski… (Q7)

Sep 21, 2013 · 4:35 PM UTC

1
1
Replying to @stilkov
@stilkov @jimwebber @stevevinoski I think that's an argument against a centralised type system in a distributed system. Decentralised types?
1
@stilkov @jimwebber @stevevinoski e.g. CDCs are a way of decentralising expectations (though they aren't universally quantified).