My point was that if you make the notion of subroutine so broad as to
include the people in the model I proposed, you have robbed its explanatory
power.  It becomes no simpler and clearer than the notions of methods and
messaging that it was supposed to explain.


>> >A message send is a plymorphic selection, followed by a subroutine
>> >callto the appropriate method.
>>
>> Does this mean that you couldn't have a distributed implementation of
>> Squeak where objects were people, methods were instructions for them
>> to follow, and messages were emails?
>
>No, it wouldn't. After all, we can implement subroutine calls and
>polymorphic selection with these means as well.
>That is, as long as people are willing to act as subroutine invocations