Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Graphical Computer Analysis of the Ciphertext frequency relative to the Plaintext frequency.
15 views
Skip to first unread message
adacrypt
Nov 7, 2011, 11:33:06 AM11/7/11
to
This is a graphic computer analysis that was carried out by me to see
how well the design objective of tailoring the encryption data to
produce ciphertext that would be as perfectly disparate as possible
was achieved. Let the reader be the judge of how successful that has
been by viewing the frequency distribution of both the ciphertext and
the corresponding plaintext of my test model being published here (see
PDF) and then making a qualitative assessment of the resistance of the
ciphertext to all statistical attacks based on the frequency of
occurrence of repeated data in the ciphertext that you see – there are
virtually zero repeats (see appendix), a fact that augers well for a
total resistance to statistical attack by any adversary.
The results are to be seen on
http://www.adacrypt.com/downloads/Website%20New%20Page%20-%20Screen%20Grab%20of%20Frequency%20Demo.pdf
Note.
*The monochrome PDF needs to be scrolled down to see the beginning of
the text.
The published test file in the experiment was comprised of just 500
characters.
Tests on Files of 10000 characters and 32000 characters gave the same
result.
This is a very satisfactory result and is the one I have engineered
for.
The frequency of occurrence of the elements of ciphertext is clearly
so different to that of the corresponding plaintext in this cipher
that any attempt at mapping the ciphertext to the plaintext based on
statistical probability is immediately ruled out even by the most
cursory direct inspection.
This cryptography is engineered to always achieve that result – it is
a planned saturation therapy of studiously built in disparateness into
the elements of the ciphertext coupled with a one-way mathematical
function – a new ballgame like I said way back that needs
understanding. It is jointly a foil to two separate forms of attack,
1) a mathematical inversion attack on the individual elements of the
ciphertext string and 2) a statistical attack on the whole string
itself en masse being taken as the collected sample in hand by a
cryptanalyst who seeks to establish frequency relationships in the
ciphertext that enable him to do probabilistic mapping of the
ciphertext to the plaintext.
It cannot ever happen in any session of encryption that the effect of
encrypting strings of plaintext will reduce the disparateness of the
ensuing ciphertext to an extent whereby it will help a cryptanalyst to
break it. That cannot ever happen with this crypto system, the
possibility has been studiously designed out of the scheme.
Applying cryptanalysis methods that have been developed for open,
number-theoretic continuous scalar cryptography will not work with
this discontinuous vector cryptography. The recent fiasco of sending
sample files for cryptanalysis is an instance of its failure but if
anybody wants to try again then just say the word and I will oblige
once more (as I have done twice already) but you will not be allowed
inside my database this time. After all if you are simulating a real-
world adversary you cannot expect to be given that much help.
Remember also that this cryptography thrives on computer power and is
not like current scalar cryptography that is threatened by it.
Summary.
Random Ciphertext by Studious Design.
Because of the elements being made equally numerous the ciphertext
string may double in fact as a random key string. Each of its
elements has equal probability of being the one that represents any
plaintext. This is an extraordinary development in that a cipher
overall is being partly secured now by the ciphertext string being
made into a random keyset and therefore not being usable to launch a
statistical attack on itself in any ”known-plaintext” attack by the
fact of it being studiously made random in the basic cipher design.
The other part is that it is impossible to circumvent the proper
decryption algorithm by trying to mathematically invert single
elements of ciphertext by direct means.
I contend that making the ciphertext string random like this by
deliberate design is a ‘first’ in cryptography.
A random key set is one thing but a random ciphertext string is
unheard of as a means of providing extra security. The ciphertext
string itself is never known to be made random as a design target as I
have done here in this cipher design.
The results of this experiment are extensible to all future
encryptions by this cipher. The frequency spread of the ciphertext is
guaranteed at virtual unity every time i.e. virtually non-repeating –
it simply cannot be otherwise –it’s a package that cannot change – a
closed system in scientific terms. It can be likened to a standard
set of playing cards that cannot be changed by the entanglement of
shuffling and slicing only it is much larger in scope and if anything
may be improved as time goes by with more computer power – even a step
up to 64-bit computer architecture enhances everything in the cipher
described. Success is locked into the design.
As Mr Kerchoff stated it “As long as the entities keep the key (read
mutual databases for key here) safe then they can enjoy perfect
secrecy of communications.
PS. This not about internal random keys, it is instead about
ciphertext strings being made random by computations on internal keys
that may not necessarily be random themselves. Verbalising it as
succintly as possible it is a vector cipher that makes the ciphertext
random and is non-invertible by being the resultant of a one-way
mathematical function. It is a whole new exploration in cryptography.
Appendix.
Let us say that the i-coefficient, the j-Coefficient and the k-
coefficient of the ciphertexts are boxes to be ticked in each case in
this experiment. It is only necessary to run the test on one
coefficient to establish the ciphertexts that are different to each
other and therefore not repeats. If one box is ticked (see the
solitary 2 in the output) as a repeat it means to say that the ‘j’ Box
has been ticked for two ciphertexts, but to be a useful repeat to a
cryptananlyst the i and k boxes of the same two ciphertexts must also
tick simultaneously for the two ciphertexts to be equal vectors and
therefore genuine repeats. The odds against this happening are in the
billions of trillions and even then it does not materially help an
adversary to have say one or two repeats when he’s hoping for groups
of at least fifty repeats and upwards.
adacrypt
how well the design objective of tailoring the encryption data to
produce ciphertext that would be as perfectly disparate as possible
was achieved. Let the reader be the judge of how successful that has
been by viewing the frequency distribution of both the ciphertext and
the corresponding plaintext of my test model being published here (see
PDF) and then making a qualitative assessment of the resistance of the
ciphertext to all statistical attacks based on the frequency of
occurrence of repeated data in the ciphertext that you see – there are
virtually zero repeats (see appendix), a fact that augers well for a
total resistance to statistical attack by any adversary.
The results are to be seen on
http://www.adacrypt.com/downloads/Website%20New%20Page%20-%20Screen%20Grab%20of%20Frequency%20Demo.pdf
Note.
*The monochrome PDF needs to be scrolled down to see the beginning of
the text.
The published test file in the experiment was comprised of just 500
characters.
Tests on Files of 10000 characters and 32000 characters gave the same
result.
This is a very satisfactory result and is the one I have engineered
for.
The frequency of occurrence of the elements of ciphertext is clearly
so different to that of the corresponding plaintext in this cipher
that any attempt at mapping the ciphertext to the plaintext based on
statistical probability is immediately ruled out even by the most
cursory direct inspection.
This cryptography is engineered to always achieve that result – it is
a planned saturation therapy of studiously built in disparateness into
the elements of the ciphertext coupled with a one-way mathematical
function – a new ballgame like I said way back that needs
understanding. It is jointly a foil to two separate forms of attack,
1) a mathematical inversion attack on the individual elements of the
ciphertext string and 2) a statistical attack on the whole string
itself en masse being taken as the collected sample in hand by a
cryptanalyst who seeks to establish frequency relationships in the
ciphertext that enable him to do probabilistic mapping of the
ciphertext to the plaintext.
It cannot ever happen in any session of encryption that the effect of
encrypting strings of plaintext will reduce the disparateness of the
ensuing ciphertext to an extent whereby it will help a cryptanalyst to
break it. That cannot ever happen with this crypto system, the
possibility has been studiously designed out of the scheme.
Applying cryptanalysis methods that have been developed for open,
number-theoretic continuous scalar cryptography will not work with
this discontinuous vector cryptography. The recent fiasco of sending
sample files for cryptanalysis is an instance of its failure but if
anybody wants to try again then just say the word and I will oblige
once more (as I have done twice already) but you will not be allowed
inside my database this time. After all if you are simulating a real-
world adversary you cannot expect to be given that much help.
Remember also that this cryptography thrives on computer power and is
not like current scalar cryptography that is threatened by it.
Summary.
Random Ciphertext by Studious Design.
Because of the elements being made equally numerous the ciphertext
string may double in fact as a random key string. Each of its
elements has equal probability of being the one that represents any
plaintext. This is an extraordinary development in that a cipher
overall is being partly secured now by the ciphertext string being
made into a random keyset and therefore not being usable to launch a
statistical attack on itself in any ”known-plaintext” attack by the
fact of it being studiously made random in the basic cipher design.
The other part is that it is impossible to circumvent the proper
decryption algorithm by trying to mathematically invert single
elements of ciphertext by direct means.
I contend that making the ciphertext string random like this by
deliberate design is a ‘first’ in cryptography.
A random key set is one thing but a random ciphertext string is
unheard of as a means of providing extra security. The ciphertext
string itself is never known to be made random as a design target as I
have done here in this cipher design.
The results of this experiment are extensible to all future
encryptions by this cipher. The frequency spread of the ciphertext is
guaranteed at virtual unity every time i.e. virtually non-repeating –
it simply cannot be otherwise –it’s a package that cannot change – a
closed system in scientific terms. It can be likened to a standard
set of playing cards that cannot be changed by the entanglement of
shuffling and slicing only it is much larger in scope and if anything
may be improved as time goes by with more computer power – even a step
up to 64-bit computer architecture enhances everything in the cipher
described. Success is locked into the design.
As Mr Kerchoff stated it “As long as the entities keep the key (read
mutual databases for key here) safe then they can enjoy perfect
secrecy of communications.
PS. This not about internal random keys, it is instead about
ciphertext strings being made random by computations on internal keys
that may not necessarily be random themselves. Verbalising it as
succintly as possible it is a vector cipher that makes the ciphertext
random and is non-invertible by being the resultant of a one-way
mathematical function. It is a whole new exploration in cryptography.
Appendix.
Let us say that the i-coefficient, the j-Coefficient and the k-
coefficient of the ciphertexts are boxes to be ticked in each case in
this experiment. It is only necessary to run the test on one
coefficient to establish the ciphertexts that are different to each
other and therefore not repeats. If one box is ticked (see the
solitary 2 in the output) as a repeat it means to say that the ‘j’ Box
has been ticked for two ciphertexts, but to be a useful repeat to a
cryptananlyst the i and k boxes of the same two ciphertexts must also
tick simultaneously for the two ciphertexts to be equal vectors and
therefore genuine repeats. The odds against this happening are in the
billions of trillions and even then it does not materially help an
adversary to have say one or two repeats when he’s hoping for groups
of at least fifty repeats and upwards.
adacrypt
WTShaw
Nov 8, 2011, 10:32:18 AM11/8/11
to
On Nov 7, 10:33 am, adacrypt <austin.oby...@hotmail.com> wrote:
> This is a graphic computer analysis that was carried out by me to see
> how well the design objective of tailoring the encryption data to
> produce ciphertext that would be as perfectly disparate as possible
> was achieved....
> This is a graphic computer analysis that was carried out by me to see
> how well the design objective of tailoring the encryption data to
> produce ciphertext that would be as perfectly disparate as possible
Viewing the file is dependent on having version 8 of Adobe PDF Reader.
That could be requiring too much since you have a website and could
use a less demanding format to display your results; keep it simple.
For my purposes, I get some clues from my program:
http://cryptopopcorn.com/Crypto/js/Utilities/CounterHistogram.html
adacrypt
Nov 8, 2011, 12:49:22 PM11/8/11
to
from adobe - is version 8 of this reader not generally available any
more? - more please for any reader who may have to acquire it -
thanks - adacrypt
adacrypt
Nov 8, 2011, 1:50:50 PM11/8/11
to
On Nov 8, 3:32 pm, WTShaw <lure...@gmail.com> wrote:
lot - almost half of it before you come to the text proper - Try that
- adacrypt
Mark Murray
Nov 8, 2011, 2:37:55 PM11/8/11
to
On 08/11/2011 18:50, adacrypt wrote:
> Just remembered - the black screen grab needs to be scrolled down a
> lot - almost half of it before you come to the text proper - Try that
> - adacrypt
You did say that (sort of) in the original posting.
> Just remembered - the black screen grab needs to be scrolled down a
> lot - almost half of it before you come to the text proper - Try that
> - adacrypt
You've managed to create a PDF that contains nothing graphical and with
no analysis, called (in the subject of this thread) "Graphical Computer
Analysis ...".
The PDF was unusual (to say the least); why not just post the text
output? Its trivial to obtain!
A subject of "Statistics of the Ciphertext frequency relative to the
Plaintext frequency" or some such would have been more honest.
Now try for statistical significance; a plaintext some orders of
magnitude larger than your vector table size, for example.
M
--
Mark "No Nickname" Murray
Notable nebbish, extreme generalist.
Globemaker
Nov 8, 2011, 3:11:44 PM11/8/11
to
> I have to admit I am not au fait with new downloads that are available
> from adobe - is version 8 of this reader not generally available any
> more? - more please for any reader who may have to acquire it -
> thanks - adacrypt
I tried to see the .pdf file, but it calls for version 8 of Adobe
> from adobe - is version 8 of this reader not generally available any
> more? - more please for any reader who may have to acquire it -
> thanks - adacrypt
Acrobat Reader. For me to get that software, a 60 page legal agreement
is presented to me. There are two things about the legal agreement
with Adobe that prevent me from seeing you .pdf file :
1 The 60 pages are longer than I care to read
2 The 60 pages cannot be searched for the words like "lawyer, fee,
attorney, suit, pay"
In summary, when I have spent time reading long contracts, they often
have a section where they want me to agree to pay any legal costs they
incur when they try to sue me. I refuse to pay attorneys to sue me. I
refuse to read a 60 page legal contract just to read your one page
graph.
Mark Murray
Nov 8, 2011, 4:08:19 PM11/8/11
to
On 08/11/2011 20:11, Globemaker wrote:
> In summary, when I have spent time reading long contracts, they often
> have a section where they want me to agree to pay any legal costs they
> incur when they try to sue me. I refuse to pay attorneys to sue me. I
> refuse to read a 60 page legal contract just to read your one page
> graph.
Its not even a graph.
> In summary, when I have spent time reading long contracts, they often
> have a section where they want me to agree to pay any legal costs they
> incur when they try to sue me. I refuse to pay attorneys to sue me. I
> refuse to read a 60 page legal contract just to read your one page
> graph.
Its a Windows CMD (in *n*x this would be "shell") text dump, done up
into a really weird PDF for no reason when the text[*] could have been
captured directly. (Foxit reads this PDF, as do *n*x readers, but
not before whining about attached files). The bloody thing is HUGE,
but the text that it represents is a couple of K. With proper data
presentation and removal of junk[+] the salient statistics could have
been less than a page, including pretty pictures, leaving plently of
space for the utterly absent analysis.)
[*] Notice: "text". Its a bunch of numbers, not a graph. It a very
large list of numbers, mostly the number "1".
[+] For some reason the ASCII set is partially listed, and labeled
(something like) "The ASCII Code". What this is for is unclear.
0 new messages