[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: tail recursion, again

     wasn't worth the performance improvement of the small percentage
     of tail-recursed function calls. 

I'm not whining about tail recursion for the sake of a small
performance improvement. Our code simply will not run on the Symbolics
without blowing the stack. As a result, my 3645 is nothing but an
expensive terminal to a Sun. I would consider it well worth a
performance slowdown for the COMPILER if it detected the cases where
there are no stack-consed objects, unwind-protects, or other
esoterica, such as 

(defun my-member (x l)
   (if (eql x (car l))
       (my-member x (cdr l))))

Ah, well. Thanks for explaining the reason behind the decision. I am
glad to find out that a design choice Symbolics made a while back is
preventing the optimization, and not "we won't do it because we don't
like your programming style."

Larry Smith
Computational Logic, Inc