Michael L Cole

Page 2 - Generalized recursive data compression using a recursive structure

Page 3 - Analysis of the Recursive data compression method

Page 4 - More Analytical results and Recursive data compression claims

Page 5 - Multimedia Search

Page 6 - News Headlines and News Today Search

Page 7 - Newsroom Automation Search

Page 9 - Everything Random Search

Page 10 - Data Search

Appendix A - Patents which reference Recursive Compression Patent 5,488,364

Appendix C - ANALYSIS OF COMPUTABLE REDUCIBILITY MODELS ANALOGOUS TO THAT DESCRIBED IN EXPIRED* U.S. PATENT 5,488,364 (Long version)

Appendix C -(Short version)

One of the many subjects that this website covers is some of my private compression research since completing work on my most famous invention "Recursive data compression" which was described in expired* U.S. Patent 5,488,364. A little reverse engineering and computable reducibility models analogous to the method illustrate that it extended ranges of file permutations which can be *-zipped by up to 153,000 percent. In 1994, that method introduced the concept of recursively splitting binary segments (into a sequence of packets) as a transform mechanism that allows the resulting binary subdivisions to compress. The Recursive data compression method demonstrates (and is complemented by the accompanying detailed procedures) that created redundancy is in part the result of the lossless conversion of n symbols to n-1 (minus one or greater) symbols, without compensation for a one byte keyword which is the methods aggregate overhead. One symbol consists of an unique fixed-length string of binary digits.

Systematical Background: Forming a recursive structure by reversibly setting, splitting and re-configuring composite binary input. Illustration A - As in the Recursive data compression patent Series 2 (below) was formed by decomposing the compositional structure of the eight binary digits that make-up the eight-bit symbols of Series 1 (defined below). The new composition was formed by the distribution of the decomposed binary structure of Series 1 by being keyword configured positionally (bits 2^7 through 2^0) and sequentially (bits 1-8 each byte) in order of occurrence (high to Low). The newly formed symbology reduces the number of symbols by approximately 98% from 256 to five while retaining a completely lossless system of numerical notation.     Recursive compression and Intrinsic Order

The lossless "Recursive data compression" method described in U.S. Patent 5,488,364 uses a keyword (8 bits) to reconfigure composite binary information. The recursive configuration is accomplished by separating bits into pairs of blocks (a and b) each part having a zero to eight bit range, equaling 16 bits of bilateral distribution. The binary digit one in the keyword represents a logical operation, a zero represents the inverse or no operation (ignored).

Below is the mathematical logic behind a Keyword formulation (8 bits) as an agency for reversible decomposability (Table 1a-1b)

Intricate bit distributions are established based on the keyword configured patterns. "The patterns may be generated and compared in any suitable manner, and the method of counting through all possible numeric values is only one example of a suitable method" according to the patent.

Information which is input is systematically configured according to this keyword by being reduced to variable length clustered packets and then distributed into one or more distinct pairs and tracked by a reversible key mask notation.

According to the patent, the "reconfiguring method comprises the steps of receiving input data; manipulating the input data in a manner that ensures that a majority of bits in the input data have a predetermined binary state; forming one or more keywords; and forming one or more pairs of blocks."

By mapping eight bits of information into nine bits of space (example below) the resulting logical map has a different entropy (by means of a simplified binary composition which is achieved by recursively splitting binary segments as a transform mechanism) than that which existed preceding the use of the re-configuration method. According to the patent, an effort is then made to compress the reordered composition "using any suitable compression method known in the art." To recursively compress data "these steps are repeated until the desired compression ratio is attained or until no further increases are observed in the compression ratio." Also, the re-configuration may be performed "before any compression has been performed" and "need not utilize the same compression algorithm on each iteration."

1020 total A and B clustered bit states (columns 5 and 6 in Table 1a )

510 total A or B clustered bit states (column 5 or 6 in Table 1a )   Evaluation table Evaluation table shows 510 sets by extracting column 2 elements from column 1 elements. Demonstrates logical keyword function.

Total AB clustered bit states (column 7 total in Table 1a) divided by keyword variation (shown on top of this page and illustrated across in Table 1a) equals nine.

Block A and B divergence of clustered bits (columns 1 and 2 in Table 1a) total 72 bits per Recursive data keyword.

Illustration 2a This binary variation is derived from the 52-byte string (top line of example). It demonstrates the lossless conversion from 27 symbols (top line) to 12 symbols (binary block and below).   This single recursive data compression variation (representative, see Table 1a) has many opportunities to succeed in microcosm; on strings of 256 bytes in length, containing 256 different symbols forming random sequences. The probability of matching at least one bit per eight bits is .(point)996094 A 256 byte string comprised of 256 repeatable symbols each randomly selected at a random probability of 0.(point)996094 has a probability of not matching one bit (per eight bits) zero to three occurrences per string at probability of 0.(point)981252 or algebraically: Individual probability of the number of times that there is no match per 256 byte string comprised of the (as described) 256 repeatable symbols. The table below shows the random probabilities of the distribution of bytes from Column A and Column B. No match (as described) occurs each time a byte value from Column A (0-127) has an adjoining invertible match (i.e. 0, 255), this requires a byte value from Column B (128-255).     The input for the next multi-part illustration is the first twelve characters from Illustration 2a which demonstrated that redundancy can be formed as a result of the recursive conversion of n symbols to n-1 (one or greater) symbols.   Illustration 2b       In the example the conversion steps through a succession of recursive compression stages from seven eight bit (sub-division to seven 4 bit) symbols to nine eight bit (sub-division to five 4 bit) symbols. Illustration 2b-2 Example of sub-division to five (4 bit symbols) Same data as used in Illustration 2b     Illustration 2b-3 The patterns of succession can also be counted as they progress further through recursive sub-divisions and symbol stages.   Example of recursive structure conversion step to six (12 bit symbols) and recursive sub-division to seven (6 bit symbols) Same data as used in Illustration 2b             Analysis of VUSP Sub-Structure nets 153,000 percent more compressible recursive compression *-zip files.   Extensive information about the invention described in expired* Patent 5,488,364: Description information from the website of The World Intellectual Property Organization (WIPO) an agency of the United Nations. Information about the patent claims from The World Intellectual Property Organization (WIPO) website. Bibliographical information from The World Intellectual Property Organization (WIPO) website.   More on page 2       According to the US Patent Office, expired* Recursive compression Patent 5,488,364 references patent: 4,092,665   Expired* Patent 5,488,364, Recursive data compression, is referenced by the following patents:   1 - 8,117,173 Efficient chunking algorithm 2 - 8,112,619 Systems and methods for accelerated loading of operating systems and application programs 3 - 8,112,496 Efficient algorithm for finding candidate objects for remote differential compression 4 - 8,090,936 Systems and methods for accelerated loading of operating systems and application programs 5 - 8,073,926 Virtual machine image server 6 - 8,073,047 Bandwidth sensitive data compression and decompression 7 - 8,054,879 Bandwidth sensitive data compression and decompression 8 - 8,024,483 Selective compression for network connections 9 - 8,010,668, Selective compression for network connections 10 - 7,882,084, Compression of data transmitted over a network 11 - 7,873,065, Selectively enabling network packet concatenation based on metrics 12 - 7,849,462, Image server 13 - 7,783,781, Adaptive compression 14 - 7,777,651, System and method for data feed acceleration and encryption 15 - 7,714,747, Data compression systems and methods 16 - 7,613,787, Efficient algorithm for finding candidate objects for remote differential compression 17 - 7,555,531, Efficient algorithm and protocol for remote differential compression 18- 7,370,120, Method and system for reducing network latency in data communication 19- 7,321,952, System and method for data phase of memory and power efficient mechanism for fast table lookup 20- 7,296,114, Control of memory and power efficient mechanism for fast table lookup 21- 7,296,113, Memory and power efficient mechanism for fast table lookup 22- 7,292,162, Data coding system and method 23- 6,301,394, Method and apparatus for compressing data 24- 6,008,657, Method for inspecting the elements of piping systems by electromagnetic waves 25- 5,666,560, Storage method and hierarchical padding structure for direct access storage device (DASD) data compression 26- 5,627,534, Dual stage compression of bit mapped image data using refined run length and LZ compression             Continue