From ...
From: Erik Naggum <erik@naggum.no>
Subject: Re: sorting...
Date: 1998/12/15
Message-ID: <3122671946890928@naggum.no>#1/1
X-Deja-AN: 422316582
References: <cxjww3vp7s9.fsf@engc.bu.edu>
mail-copies-to: never
Organization: Naggum Software; +47 8800 8879; http://www.naggum.no
Newsgroups: comp.lang.lisp

* David Bakhash <cadet@bu.edu>
| suppose that you have a list.  You want to sort it.  suppose that you
| knew ahead of time that the list would be much closer to being sorted
| if you reversed it.  Would it make sense to call sort on the reversed
| list rather than the list itself?  would people think that this would
| save time, or take more time?

  this is all wrong as an approach to improving performance.  instead,
  study the available sort algorithms in the considerable literature and
  implement one that fits your data after you have discovered that the
  sorting time is significant to the rest of your work and worth fixing.

  I do (sort (cons <element> <sorted-list>) <test>) when adding one element
  at a time to a priority list.  profiling shows that it is irrelevant to
  the execution time of the program, although I'd bet that some form of
  programmer instinct would be triggered by this particular example and
  would try to rewrite it.

#:Erik
-- 
  man who cooks while hacking eats food that has died twice.