[plt-scheme] guido on tail recursion

From: Shriram Krishnamurthi (sk at cs.brown.edu)
Date: Thu Apr 23 10:29:06 EDT 2009

I was disappointed that he posted this well past April 1.

The sad thing is, he's roughly right about the technical stuff.  It's
the big picture he's missing.  The whole "let's count %ages of text in
Wikipedia" thing entirely obfuscates the real issues; the commentors
have confused it further by talking about mutation; and nobody seems
to have realized that it's just another kind of garbage collection,
which is indeed one of those things you don't want to put in a
language for fear programmers will come to rely on it, and your
overall (space) complexity can be impacted in a measurable way --
except Python already has GC.  But given that GC was botched
initially, it seems hardly surprising that tail calls are botched,

Ah well.  Stuff like this is good for us.


Posted on the users mailing list.