All editing errors are my own, I re-formatted the original text to better address the points. Forwarded message:
Date: Sun, 18 Jan 1998 19:17:32 -0800 From: Steve Schear <schear@lvdi.net> Subject: Re: NTY compression proposal
Compression efficiency depends upon
1) the entropy of the input data
We are looking at selecting 100 64-bit keys out of a set of 2^64, that is not a lot of opportunity there, given a random key generation process, for a lot of commen patterns in the key set. If we don't have a lot of commen patterns in the text we don't get a lot of compression.
2) the allowable losses
In this case we can't afford any loss.
3) finding an efficient algorithm to code for that entropy.
Which if we do the math requires something in the neighborhood of 33:1, for a loss-less algorithm that's a pretty hurky boundary condition.
I think I'm on fairly firm ground assuming that key data entropy is very= high and therefore little or no compressible is feasible.
My point exactly, which would lead us to the conclusion that our requirement of a 33:1 compression factor is not realizable. ____________________________________________________________________ | | | The most powerful passion in life is not love or hate, | | but the desire to edit somebody elses words. | | | | Sign in Ed Barsis' office | | | | _____ The Armadillo Group | | ,::////;::-. Austin, Tx. USA | | /:'///// ``::>/|/ http://www.ssz.com/ | | .', |||| `/( e\ | | -====~~mm-'`-```-mm --'- Jim Choate | | ravage@ssz.com | | 512-451-7087 | |____________________________________________________________________|