From ...
From: Erik Naggum <erik@naggum.no>
Subject: Re: Please help!!
Date: 1999/02/09
Message-ID: <3127567131757704@naggum.no>#1/1
X-Deja-AN: 442467890
References: <36BCA5F9.F7B8AE61@hotmail.com> <sfw679e5fry.fsf@world.std.com> <3127421575197623@naggum.no> <79nb45$c15$1@news1-alterdial.uu.net> <VLHv2.2054$oD6.86307@burlma1-snr1.gtei.net> <8790e7c44w.fsf@2xtreme.net> <ekvg18fzm81.fsf@gnoll.pvv.ntnu.no>
mail-copies-to: never
Organization: Naggum Software; +47 8800 8879; http://www.naggum.no
Newsgroups: comp.lang.lisp

* Stig Hemmer <stig@pvv.ntnu.no>
| The problem is that you will spend a lot of effort making a
| optimalization routine that will only be used very very seldom.  Most
| programs do _not_ look like fibonacci.

  well...  apart from wanting to avoid the first recursive function call
  (which can be done with a simple iteration from a lower limit to the
  value in question, but this is not something I want the compiler to
  decide for me), optimizing fibonacci is mostly a matter of memoizing.  it
  _would_ be useful if a compiler could see that there is no way a function
  would return a different value if given the same arguments, i.e.,
  automated detection of side-effect-free functions and various means of
  propagation of this information.

#:Erik
-- 
  Y2K conversion simplified: Januark, Februark, March, April, Mak, June,
  Julk, August, September, October, November, December.