Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!spool.mu.edu!olivea!mintaka!ai-lab!life.ai.mit.edu!mycroft
From: mycroft@kropotki.gnu.ai.mit.edu (Charles Hannum)
Newsgroups: comp.compression
Subject: Re: Graph of performance of data compression programs.
Message-ID: <MYCROFT.91Jun21141746@kropotki.gnu.ai.mit.edu>
Date: 21 Jun 91 18:17:46 GMT
References: <867@spam.ua.oz>
Sender: news@ai.mit.edu
Followup-To: comp.compression
Organization: None.  Utter chaos.
Lines: 15
In-reply-to: ross@spam.ua.oz.au's message of 21 Jun 91 14:33:48 GMT

In article <867@spam.ua.oz> ross@spam.ua.oz.au (Ross Williams) writes:

   I am involved in designing high-performance algorithms but am finding
   it difficult to work out what the state of the art is. Some algorithms
   are written in C, some in machine code. Algorithms are run on PCs,
   Macs, Suns, and other machines. It would be nice to standardize
   somehow.

I hate to pick nits, but the paragraph above confused the hell out of
me.  An 'algorithm' is written in pseudecode or a natural language;
never in a programming language.  I think we need to differentiate
between an 'algorithm' and an 'implementation of an algorithm'.  When
we start thinking of an algorithm as inseparable from its
implementation, then we get such meaningless benchmark results as the
ones you referenced.