First of all, this specification is much more useful for byte streams
which are less than length 1000 as opposed to archives in the gigabyte
range. The reason is that this one bloats the input data stream by
about a factor of five to produce an output stream. I recommend that
this method be applied last of all the methods which I have specified
to be used to encrypt the data.
Consider that a character string forming a word can be converted into
a single byte code which is the least significant 8 bits of the
weighted sum of all the bytes in the string. A set of random number
tables associated with the particular library can be used to assign a
value by which to multiply each letter value, and a letter value
dependent upon the character and its position in the word. So a
library containing hundreds of thousands of real common English words
with their codes can be created. The library is indexed by the derived
codes and a word length organized from short words to long ones. The
input byte stream is processed one byte at a time to select a choice
word to be used to represent the value of the byte. An improved
version will perform the selection so that the words chosen form
sentences which more or less have meaning in the context of the total
byte stream. Since the values of the byte code for words in the output
can be deduced by anyone with sophisticated enough computers, it is
recommended that the techniques of the thread posted on Google group
"Lonnie Courtney Clay" thread "Merry Christmas 3" should be used to
make the information relatively worthless.
As an enhancement to the original library, an archive of text
documents can be used to generate coded sentences which decode to
commonly encountered input byte patterns. If such an option is
available, then documents can be pre-processed before the usage of the
library specified in "Merry Christmas" thread on Google group "Lonnie
Courtney Clay". By doing it first before the data compression stage,
the bloat might be reduced to a factor of two or three.
I see evidence in the content of the Sporger attack posts that someone
is already using something similar to this technique to flood the
internet with text of dubious value. I recommend that better filters
be developed to track down the real authors' ISP and shut their
operations down. Whatever they are saying to each other, it is likely
of a criminal nature...
It took one hour to prepare this document, but the method was thought
out years ago. Through serendipity it popped into my conscious mind
today.
Don't clap, throw money!
Send checks to 3395 Harrell Road Arlington TN 38002-4261
Lonnie Courtney Clay Laughing Crazy Coot TARZAN Chic Logo Guy
This garbage is from a sporger, not me. The sporger does not include a
signature such as follows.