Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site ulysses.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gsp
From: gsp@ulysses.UUCP (Gary Perlman)
Newsgroups: net.sources,net.followup,net.flame
Subject: Re: A Simple Bubble Sort Function
Message-ID: <879@ulysses.UUCP>
Date: Mon, 11-Jun-84 20:12:26 EDT
Article-I.D.: ulysses.879
Posted: Mon Jun 11 20:12:26 1984
Date-Received: Tue, 12-Jun-84 01:28:52 EDT
References: <165@callan.UUCP>
Organization: AT&T Bell Laboratories, Murray Hill
Lines: 28



> Jack Purdum is rapidly getting a reputation as an idiot with me.  I thought
> that by now *EVERYBODY* knew that the bubble sort is for cretins.  If you
> want a quick simple sort, write a "selection sort":  search 0 thru n for the
> largest item and swap it with the item in slot n;  repeat with n=n-1 until
> done.  This is exactly the effect that the bubble sort achieves (think about
> it for a while if you aren't sure), but without all the unnecessary exchanges.

> Moral:	Don't waste your effort optimizing the wrong solution to the problem.

> 	Geoff Kuenning
> 	Callan Data Systems
> 	...!ihnp4!wlbr!callan!geoff

It is true that bubble sort is about as bad an algorithm as can be found.
Floor sort, used by me when I once dropped my FORTAN deck, is a bit worse.
Still, bubble sort is useful as a teaching tool, and its impracticality
makes it unlikely that teachers can find implementations of it in C.
I have trouble finding serious fault with anyone generous enough to
post their software to net.sources.  I realized that I was free to ignore it.
So while I can't cheer for the posting of some code that people should not use
for efficiency's sake, I find personal attacks in very poor taste.

For another non-optimal sorting routine, see the standard C text,
The C Programming Language (Prentice Hall) for a shell sort routine.
For small arrays, it gives remarkably good results for such a small procedure.
	Gary Perlman	BTL MH 5D-105	(201) 582-3624	ulysses!gsp