Re: New Generation Lossless Data Representations
- From: LawCounsels <LawCounsels@xxxxxxx>
- Date: Thu, 12 Aug 2010 21:13:00 -0700 (PDT)
On 11 Aug, 06:52, LawCounsels <LawCouns...@xxxxxxx> wrote:
here is link to download a 'mathematics structure' encoding of a
complete random 4,074 bits long file (a random chosen part of Mark
Nelson's AMillionRandomDigits.bin challenge)
http://www.box.net/shared/eyy2v28dbf
download link's new discovered 'mathematics structure' endoded file's
details
.. in a file with N bits (sufficient large like 8Kbits onwards to
1Mbits) , assume the distributions of unique prefix '0's or '10' or
'110' or '1110' ... or '111...10' (always ends with a '0') is
standard binomial power (ie random) , BUT with added constraints/
restriction that the maximum length of '111....10' at any time could
only be at most log(base2)[R] where R is the total # of bits from
present bit position to the end Least Significant Bit position [ eg
if
bits_string is 0 111110 11110 10 0 10 110 10 0 10 0 10 0 0 '
then here 110 is the 13th bits , there can be no unique prefix of
total# of bits length > log(base2)[R] at any time, where R is the bit
position # of the unique prefix counting from end Least Significant
position bit ....]
.....so this is not simple regular usual normal binomial power series/
random distributions [ usual normal binomial power series/ random
distributions is 'god gifted' not to be compressible] , but here
there
is added constraint/ restriction that eg '1110' ( of length 4 bits
long) could not occur at position R = 15 or smaller (since log(base2)
[R=15 or smaller value of R] will not allow unique prefix of total#
of
bits >= 4 to occur at position R < 16 ......
THIS IS IMMEDIATE APPARENT READY FURTHER VERY COMPRESSIBLE 4,073 bits
'constraint' 'mathematics structures' encoded (already 1 bit smaller)
file , from input 'universal thought uncompressable' random 4,074 bits
'random' file
.
- Follow-Ups:
- Re: New Generation Lossless Data Representations
- From: LawCounsels
- Re: New Generation Lossless Data Representations
- References:
- New Generation Lossless Data Representations
- From: LawCounsels
- New Generation Lossless Data Representations
- Prev by Date: Re: New Random Number Generator for encryption
- Next by Date: www.cntrade09.com -----*** 2010 brand shoes - Gucci shoes -shoes Clothes - Men's jeans - women's jeans - Men's Shorts - women's
- Previous by thread: New Generation Lossless Data Representations
- Next by thread: Re: New Generation Lossless Data Representations
- Index(es):