On Thu, Aug 2, 2012 at 7:00 PM, Henry Baker <hbaker1@pipeline.com> wrote:
Carl Hewitt & I discussed this eta-tail recursion connection with Steele & Sussman
I'll have to ask him about it; I see him at a weekly meeting about object capabilities from time to time.
re Scheme back in the 1970's when Steele & Sussman were pushing to make tail recursion a requirement for Scheme & other Lisp's (Maclisp at the time was _not_ tail recursive, nor was the Lisp Machine Lisp). I haven't checked recently, but I think that even Steele wasn't able to get tail-recursion into the Common Lisp standard.
Apparently not: http://www.lispworks.com/documentation/HyperSpec/Body/03_bbc.htm
I think this discussion was prompted by Steve Ward's undergraduate course at MIT on lambda calculus circa 1975??
BTW, I found it very interesting that someone (Church??) had discovered that axiom eta was _independent_ of the other lambda calculus axioms. That alerted me that something important was going on.
I don't know about *independence*, but eta is *adjoint* to beta: http://www.paultaylor.eu/ASD/foufct/cattype.html -- Mike Stay - metaweta@gmail.com http://www.cs.auckland.ac.nz/~mike http://reperiendi.wordpress.com