Re: Ann: delay pack

On 12/08/2013, at 7:07 PM, Jan Wielemaker wrote:
> The nice thing about
> explicit inclusion of delay is that it makes you more aware of
> the biggest risk: delaying over a cut changes the semantics.

I spent some time thinking about coroutining in the mid-to-late
1980s.  I had what I thought was an efficient scheme, which was
ridiculed by one of the founders of Quintus before I had a chance
to explain what the scheme actually _was_, and I've never since
been able to remember it.  I do recall that one of my key design
goals was that if you have
	p(...) :- q(...), !, r(...)
and you reach the cut with any part of q(...) still suspended,
the (!, r(...)) should suspend until q had finished.  It seemed
to me then, and still seems to me now, a fairly fundamental bug
to do otherwise.