Marcel --

That's right. Coroutines were a big thing in the 60s, and they were an
important part of Simula and my early experiments at designing a dynamic
OOPL. When  we did "Simpula" ("Simula" in Smalltalk-72) we discovered that
it was simpler and cleaner to use several methods for the different phases
of the coroutine. The scheduling set in Simpula carried a pointer to the
objects that were sleeping plus what message to send them when they were
awakened. This worked well and it was one of the motivations for not
worrying so much about the symmetrical send-receive at that time.

Cheers,

Alan

------

At 6:59 AM -0000 3/23/99, Marcel Weiher wrote:
>> The sender of a method can be obtained by sending sender to
>> thisContext. I am suggesting adding a new pseudovariable (something
>> like self, super, etc.) that would access the caller directly. I
>think that
>> this would provide the symmetry mentioned by Alan earlier.
>
>I can't speak for Alan, but my understanding was that this is not
>sufficient for the symmetry he was talking about.  The reason is that
>sending messages to the caller-object does not return the objects to
>the calling method.  What you need to get this sort of symmetry is
>(at the least) some sort of co-routine concept.
>
>Marcel