1994-01-26 - Re: Randomness of a bit string

Header Data

From: rjc@gnu.ai.mit.edu (Ray)
To: m5@vail.tivoli.com (Mike McNally)
Message Hash: 665f2ac3db0dee4f92ae599355a9bd924de6c2093a283a6ee30bcf8c5a0dfb06
Message ID: <9401261453.AA26814@churchy.gnu.ai.mit.edu>
Reply To: <9401261341.AA05996@vail.tivoli.com>
UTC Datetime: 1994-01-26 14:57:08 UTC
Raw Date: Wed, 26 Jan 94 06:57:08 PST

Raw message

From: rjc@gnu.ai.mit.edu (Ray)
Date: Wed, 26 Jan 94 06:57:08 PST
To: m5@vail.tivoli.com (Mike McNally)
Subject: Re: Randomness of a bit string
In-Reply-To: <9401261341.AA05996@vail.tivoli.com>
Message-ID: <9401261453.AA26814@churchy.gnu.ai.mit.edu>
MIME-Version: 1.0
Content-Type: text/plain


Mike McNally writes:
> 
> 
> Ray writes:
>  > All of this is meaningless anyway. Information theory was proven wrong
>  > by WEB technologies when they invented a compression program that can
>  > recursively compress any input data down to 64k. Harddrives are now
>  > obsolete.
> 
> Either I'm really dense in one of two ways (this is a joke I don't
> get, or it's really true), or my pegging bullshit meter is right.
> Could you go into a little more detail?

  It's a joke. WEB technologies is a company which announced a compression
product about a year ago. They claimed that they had "violated the laws
of information theory" (a simple counting argument proves what they claim
was impossible) by producing a compressor which can compress its own
output! Furthermore, the compression was ALWAYS 16:1, no matter what the
input. (that's right, they claimed ALL files of a certain size were
compressible by this ratio)

  Everyone knew it was BS, but BYTE magazine did a story on it which seemed
to enhance its credibility. People called them and tried to get
specs, or demo software but they were given the run around. Finally, WEB
claimed that their engineer had made a mistake and their software wouldn't
do what it claimed. A more accurate and detailed story can be found in
the comp.compression faq.



-- Ray Cromwell        |    Engineering is the implementation of science;    --
-- rjc@gnu.ai.mit.edu  |       politics is the implementation of faith.      --





Thread