On Thu, May 14, 2009 at 12:48 PM, Roger Hui <[email protected]> wrote:

> The question really is, can you code it recursively
> without paying severe performance penalty?
> (Why else would you avoid recursion if a recursive
> solution is shorter, simpler, more robust, etc.)
>
> Answer: yes.
>
> fib=: 3 : 0 M.
>  if. 2>y do. y else. (fib y-1) + (fib y-2) end.
> )
>
> fib1=: 3 : 0
>  if. 2>y do. y else. (fib1 y-1) + (fib1 y-2) end.
> )
>
BTW, TCO would not help in this case even it is implemented.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to