Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Customised Extra Entanglement in Vector Cryptography.
43 views
Skip to first unread message
adacrypt
Oct 22, 2011, 2:56:25 AM10/22/11
to
In the vector cipher to hand Alice uses a directed number line in
three-dimensional space as her encryption tool. The operative word
here is ‘directed’ because she is able to tweak the direction of this
line from within the source code of her encryption program in any
number of ways by incrementing a particular parameter called
“Increase” in steps of any size from 1 to 30000 (my home computer
crashed on integer overflow when I tried 50000 so I came down quickly
to 30000 but a 64-bit computer will do a lot more than this). The
upshot of changing this single parameter ‘increase’ in the program
source code of Alice's program by even one step say from increase = 4
to increase = 5 is that the program immediately outputs a different
set of ciphertext for the same message with each change withou
affecting the way it decrypts. Any message can be encrypted in 30000
very distinctive different ways therefore, without altering the way it
decrypts at Bob’s end.
Bob is unaffected by anything that Alice does in this respect and
doesn’t even need to know that she is doing it.
Clearly then, Alice has this powerful extra attribute in vector
cryptography of being able to formulate her ciphertext in any one of
30000 ways (and more of course with better computers) without changing
the eventual decrypted message-text at Bob’s end. The implications to
stronger security are self-evident. For instance if Alice has any
reason to believe that repeated passages of standard jargon like say
legal jargon or military jargon could lead to a known plaintext attack
by an adversary then she can obviate this at the touch of a key in
30000 ways (historical stories from Bletchley Park – UK, during WW2
are redolent of this repeated standard messages happening often).
I find it amusing that this can go over Bob’s head without his knowing
it or even needing to know. It is after all Alice’ secure
communications loop. She calls the shots in any way she wants.
Nothing she does however affects Bob.
The implications of this powerful attribute to multiple sendings of
the same message to lots of trusted third parties are very evident.
It might even be useful as an extra tool in non-repudiation schemes
(relax - I am not claiming that - not yet anyway)– the possibilities
are legion in different extra ways of securing infra structures I
think.
A look at the “Talking Machine” will help to demonstrate how Bob is
unaffected by whatever Alice does in private while corresponding with
him. He only needs to know her Pn vector i.e. the position vector
(Pn) of her transformed plaintext into a number (deducible from the
ciphertext sent to him by her) to proceed with the decryption. How
she created that position vector of her particular ‘n’ in the first
place is immaterial to him.
See Talking Machine - http://www.adacrypt.com/downloads/The%20Talking%20Machine.pdf
.
Consider now the brash claim by one reader that he can cryptanalyse
ciphertext from a provided set of plaintext and the corresponding set
of ciphertext in this cryptography using standard old fashioned
methods ! - what a hope when there are 30000 possibilities of cipher
text sets to send him for the one set of plaintext being sent to him
concurrently. That is to say nothing of the different mathematics and
algorithm - its like suggesting that model T Ford experience is
sufficient to service a modern racing machine like "Red Bull"
formula_1 cars or "Ferraris"
He hasn’t taken the trouble to understand this cryptography and
instead argues he can just use outdated One-Time Pad logic.
I can’t get him to come out of the ladies toilet to confront him with
some samples of plaintext and cipher text to see him demonstrate his
claims.
Future users of this cryptography, that is going to be free software
shortly under the GNU General Public License, will be able to
experiment with this useful encryption ploy themselves when they
download the software.
The cipher implementation called “Skew Line Encryptions” will be
downloadable from several sources (“Popular Cryptography Magazine” for
one has agreed to make it available – there will be others also).
Enjoy - adacrypt
three-dimensional space as her encryption tool. The operative word
here is ‘directed’ because she is able to tweak the direction of this
line from within the source code of her encryption program in any
number of ways by incrementing a particular parameter called
“Increase” in steps of any size from 1 to 30000 (my home computer
crashed on integer overflow when I tried 50000 so I came down quickly
to 30000 but a 64-bit computer will do a lot more than this). The
upshot of changing this single parameter ‘increase’ in the program
source code of Alice's program by even one step say from increase = 4
to increase = 5 is that the program immediately outputs a different
set of ciphertext for the same message with each change withou
affecting the way it decrypts. Any message can be encrypted in 30000
very distinctive different ways therefore, without altering the way it
decrypts at Bob’s end.
Bob is unaffected by anything that Alice does in this respect and
doesn’t even need to know that she is doing it.
Clearly then, Alice has this powerful extra attribute in vector
cryptography of being able to formulate her ciphertext in any one of
30000 ways (and more of course with better computers) without changing
the eventual decrypted message-text at Bob’s end. The implications to
stronger security are self-evident. For instance if Alice has any
reason to believe that repeated passages of standard jargon like say
legal jargon or military jargon could lead to a known plaintext attack
by an adversary then she can obviate this at the touch of a key in
30000 ways (historical stories from Bletchley Park – UK, during WW2
are redolent of this repeated standard messages happening often).
I find it amusing that this can go over Bob’s head without his knowing
it or even needing to know. It is after all Alice’ secure
communications loop. She calls the shots in any way she wants.
Nothing she does however affects Bob.
The implications of this powerful attribute to multiple sendings of
the same message to lots of trusted third parties are very evident.
It might even be useful as an extra tool in non-repudiation schemes
(relax - I am not claiming that - not yet anyway)– the possibilities
are legion in different extra ways of securing infra structures I
think.
A look at the “Talking Machine” will help to demonstrate how Bob is
unaffected by whatever Alice does in private while corresponding with
him. He only needs to know her Pn vector i.e. the position vector
(Pn) of her transformed plaintext into a number (deducible from the
ciphertext sent to him by her) to proceed with the decryption. How
she created that position vector of her particular ‘n’ in the first
place is immaterial to him.
See Talking Machine - http://www.adacrypt.com/downloads/The%20Talking%20Machine.pdf
.
Consider now the brash claim by one reader that he can cryptanalyse
ciphertext from a provided set of plaintext and the corresponding set
of ciphertext in this cryptography using standard old fashioned
methods ! - what a hope when there are 30000 possibilities of cipher
text sets to send him for the one set of plaintext being sent to him
concurrently. That is to say nothing of the different mathematics and
algorithm - its like suggesting that model T Ford experience is
sufficient to service a modern racing machine like "Red Bull"
formula_1 cars or "Ferraris"
He hasn’t taken the trouble to understand this cryptography and
instead argues he can just use outdated One-Time Pad logic.
I can’t get him to come out of the ladies toilet to confront him with
some samples of plaintext and cipher text to see him demonstrate his
claims.
Future users of this cryptography, that is going to be free software
shortly under the GNU General Public License, will be able to
experiment with this useful encryption ploy themselves when they
download the software.
The cipher implementation called “Skew Line Encryptions” will be
downloadable from several sources (“Popular Cryptography Magazine” for
one has agreed to make it available – there will be others also).
Enjoy - adacrypt
Mark Murray
Oct 22, 2011, 4:50:00 AM10/22/11
to
On 22/10/2011 07:56, adacrypt wrote:
> Consider now the brash claim by one reader that he can cryptanalyse
> ciphertext from a provided set of plaintext and the corresponding set
> of ciphertext in this cryptography using standard old fashioned
> methods ! - what a hope when there are 30000 possibilities of cipher
> text sets to send him for the one set of plaintext being sent to him
> concurrently. That is to say nothing of the different mathematics and
> algorithm - its like suggesting that model T Ford experience is
> sufficient to service a modern racing machine like "Red Bull"
> formula_1 cars or "Ferraris"
Send him the ciphertexts he has asked for. Put up or shut up.
> Consider now the brash claim by one reader that he can cryptanalyse
> ciphertext from a provided set of plaintext and the corresponding set
> of ciphertext in this cryptography using standard old fashioned
> methods ! - what a hope when there are 30000 possibilities of cipher
> text sets to send him for the one set of plaintext being sent to him
> concurrently. That is to say nothing of the different mathematics and
> algorithm - its like suggesting that model T Ford experience is
> sufficient to service a modern racing machine like "Red Bull"
> formula_1 cars or "Ferraris"
> He hasn’t taken the trouble to understand this cryptography and
> instead argues he can just use outdated One-Time Pad logic.
> I can’t get him to come out of the ladies toilet to confront him with
> some samples of plaintext and cipher text to see him demonstrate his
> claims.
him anything. You scared he might break your cipher?
Its easy; encrypt some files as if you were both Bob and Alice, and
publish the ciphertexts and cleartext parameters here on sci.crypt,
making all readers here Eve. If you cipher is good, it will stand up
to this.
M
--
Mark "No Nickname" Murray
Notable nebbish, extreme generalist.
adacrypt
Oct 22, 2011, 4:53:07 AM10/22/11
to
Does he want more? he only has to say so. - adacrypt
Mark Murray
Oct 22, 2011, 5:03:41 AM10/22/11
to
On 22/10/2011 09:53, adacrypt wrote:
> Does he want more? he only has to say so. - adacrypt
He already did.
> Does he want more? he only has to say so. - adacrypt
adacrypt
Oct 22, 2011, 6:29:46 AM10/22/11
to
Austin O'Byrne
Austin O'Byrne
austin...@hotmail.com
Change picture
View profile
To pmar...@grupopie.com
From: Austin O'Byrne (austin...@hotmail.com)
Sent: 22 October 2011 10:17:58
To: pmar...@grupopie.com
Hotmail Active View
5 attachments (total 580.5 KB)
PlainText...dat
Download(4.7 KB)
CipherTex...dat
Download(151.6 KB)
PlainText...dat
Download(10.9 KB)
CipherTex...dat
Download(350.1 KB)
CipherTex...dat
Download(63.2 KB)
Download all as zip
Hi,
Herewith the files you asked for again. Each File has the information
"Option 5" accompanying it - this is the only information that Bob
will get from Alice by way of parameter passing instructions - that
means that the adversary that you are simulating gets this information
also and no more.
Recapping: You said two plaintextfiles and the associated ciphertext
files + a cipheretext file which you will then cryptanalyse.
Herewith,
1) PlaintextFile_4000.dat
CipherTextFile_4000.dat
2) PlainTextFile_10000.dat
CipherTextFile_10000.dat
3) PlainTextFile_2000.dat
- adacrypt
New
|
Reply
Reply all
Forward
|
Delete
Mark this folder as read
Empty this folder
Mark
Done again - adacrypt
Mark Murray
Oct 22, 2011, 8:22:17 AM10/22/11
to
I suspect you are cheating (again).
This "Option 5" business is not part of the original
cipher and challenge; in fact you clearly stated, on
numerous occasions that the parameters were public
information and that changing them on different
invocations was sufficient to safeguard the cypher.
M
This "Option 5" business is not part of the original
cipher and challenge; in fact you clearly stated, on
numerous occasions that the parameters were public
information and that changing them on different
invocations was sufficient to safeguard the cypher.
M
adacrypt
Oct 22, 2011, 8:50:09 AM10/22/11
to
On Oct 22, 1:22 pm, Mark Murray <w.h.o...@example.com> wrote:
> I suspect you are cheating (again).
>
> This "Option 5" business is not part of the original
> cipher and challenge; in fact you clearly stated, on
> numerous occasions that the parameters were public
> information and that changing them on different
> invocations was sufficient to safeguard the cypher.
>
> M
>
> On 22/10/2011 11:29, adacrypt wrote:
>
>
>
>
>
> > On Oct 22, 10:03 am, Mark Murray<w.h.o...@example.com> wrote:
> >> On 22/10/2011 09:53, adacrypt wrote:
>
> >>> Does he want more? he only has to say so. - adacrypt
>
> >> He already did.
>
> >> M
> >> --
> >> Mark "No Nickname" Murray
> >> Notable nebbish, extreme generalist.
>
> > Files you asked for.
>
> > Austin O'Byrne
>
> > Austin O'Byrne
> > austin.oby...@hotmail.com
> I suspect you are cheating (again).
>
> This "Option 5" business is not part of the original
> cipher and challenge; in fact you clearly stated, on
> numerous occasions that the parameters were public
> information and that changing them on different
> invocations was sufficient to safeguard the cypher.
>
> M
>
> On 22/10/2011 11:29, adacrypt wrote:
>
>
>
>
>
> > On Oct 22, 10:03 am, Mark Murray<w.h.o...@example.com> wrote:
> >> On 22/10/2011 09:53, adacrypt wrote:
>
> >>> Does he want more? he only has to say so. - adacrypt
>
> >> He already did.
>
> >> M
> >> --
> >> Mark "No Nickname" Murray
> >> Notable nebbish, extreme generalist.
>
> > Files you asked for.
>
> > Austin O'Byrne
>
> > Austin O'Byrne
>
> > Change picture
> > View profile
> > To pmarq...@grupopie.com
> > From: Austin O'Byrne (austin.oby...@hotmail.com)
> > Sent: 22 October 2011 10:17:58
> > To: pmarq...@grupopie.com
>
> - Show quoted text -
It makes no difference whatever - PM has more going for him than any
adversary in the real world would ever have - I can change my cipher
at will to stop inane arguments any time I like which is what I am
doing.
This kind of remark indicates beyond all doubt that you haven't a clue
in hell about the cipher that I am promoting but you want to cast
doubt by throwing in ridiculous charges of deceit on my part - I
wouldn't have wasted the last fifteen years on something I wasn't so
sure about if that was the case - also I am a stark realist and
scrupilously conscientious on matters of science especially when they
are so mathematically intensive.
Keep on rowing like WT said a long time ago.
- adacrypt
PS - You'll lose your place in the herd if you are caught
communicating with me - this Irish pariah loves the solitude - You've
had your instructions from PM not to answer me on any any account -
Isn't life grand !
Mark Murray
Oct 22, 2011, 9:16:32 AM10/22/11
to
On 22/10/2011 13:50, adacrypt wrote:
> This kind of remark indicates beyond all doubt that you haven't a clue
> in hell about the cipher that I am promoting but you want to cast
> doubt by throwing in ridiculous charges of deceit on my part
What preposterous logic!
> This kind of remark indicates beyond all doubt that you haven't a clue
> in hell about the cipher that I am promoting but you want to cast
> doubt by throwing in ridiculous charges of deceit on my part
> - I
> wouldn't have wasted the last fifteen years on something I wasn't so
> sure about if that was the case - also I am a stark realist and
> scrupilously conscientious on matters of science especially when they
> are so mathematically intensive.
Being sure of something is distinct from being correct.
As for conscientiousness, your routine ignoring of evidence,
interspersed with wild scrambles to contain successful attacks
speaks for itself; you have little clue about what you are doing;
you substitute with bluster.
adacrypt
Oct 22, 2011, 9:23:05 AM10/22/11
to
rossum
Oct 22, 2011, 5:41:10 PM10/22/11
to
On Sat, 22 Oct 2011 05:50:09 -0700 (PDT), adacrypt
<austin...@hotmail.com> wrote:
>It makes no difference whatever - PM has more going for him than any
>adversary in the real world would ever have - I can change my cipher
>at will to stop inane arguments any time I like which is what I am
>doing.
You can indeed change your cypher at will, but then Bob won't be able
<austin...@hotmail.com> wrote:
>It makes no difference whatever - PM has more going for him than any
>adversary in the real world would ever have - I can change my cipher
>at will to stop inane arguments any time I like which is what I am
>doing.
to read anything you send, because Bob still has the old, unchanged,
version of the cypher.
How are you going to tell Bob about the changes you make? If you do
it in clear then Eve also knows the changes. If you use the old
cypher, then Eve may well have broken it (otherwise why change?).
How do you tell Bob about the changes you have made?
rossum
Mark Murray
Oct 22, 2011, 6:01:03 PM10/22/11
to
He doesn't have "a cipher" anymore. He has a rapidly morphing
program that he fervently believes is "a cipher" because he says
its so.
adacrypt
Oct 22, 2011, 7:47:03 PM10/22/11
to
On Oct 22, 10:41 pm, rossum <rossu...@coldmail.com> wrote:
> On Sat, 22 Oct 2011 05:50:09 -0700 (PDT), adacrypt
>
> On Sat, 22 Oct 2011 05:50:09 -0700 (PDT), adacrypt
>
Everything is being confused at the moment because PM is inside my
database - this is a position that he will never again enjoy and he is
only there now because he abused my generosity when I presented him
with the free download - I went out on a limb to be kind and it broke
but even at that i.e. the worst imaginable scenario in a real world
atttack by a real world adversary he still hasn't managed to break the
sample ciphertext he asked for and was given by me. No adversary could
get this far without a complete breach of the all-important database -
PM has had it handed to him and could not capitalise on the advantage
put to him on a plate.
The original design will be reverted to when the dust has settled on
this unplanned fiasco (maybe - its optional) In any case referencing
the scrambling parameters implicitly from arrays in a new change to
the design is quite workable and then Alice need only point to the
concealed option in Bob's database for him to know the parameter
passing option. This travels inside the initial secure delivery and
it can be changed by secure means thereafter if that is decided by
Alice - its a design possibility at present only.
Asking for two files of plaintext as he did on top of being inside the
database is laughable. He would make a worse thief than he is as a
cryptanalyst.
I'm pleased with the outcome of this fiasco at the end of the day - it
has become a tad embarrasing to me to have to put up with it all when
I generously offered a program that was intended to be a demonstration
and therefore was totally unprotected - talk about biting the
hand.....
- adacrypt
Gene and Debbie Styer
Oct 22, 2011, 11:24:43 PM10/22/11
to
> Asking for two files of plaintext as he did on top of being inside the
> database is laughable. He would make a worse thief than he is as a
> cryptanalyst.
the recent WikiLeaks mess - Historically you cannot assume that
everybody will be honest, theft will not occur, or that nobody will
(accidentally or deliberately) release sensitive information to the
enemy. If the release of one person's information means that other
people cannot communicate securely, then your system is not usable in
a practical situation.
adacrypt
Oct 23, 2011, 12:52:35 AM10/23/11
to
On Oct 23, 4:24 am, Gene and Debbie Styer <Eugene.St...@eku.edu>
wrote:
People is plural - there is only one person being jeopardised, that's
Alice - This fiasco has spawned more red herrings than enough
That might be the only thing to be claimed out of this faisco - Pm
simulated that situation and my cipher survived.
Because something appeared in wikipedia that doesnt make it a theorem.
I treat everything I read there as just having only temporary
scientific credibility.
What you say implies that AES and RSA may have already been broken in
isolated cases that we know nothing about.
This unfortunate episode should never have happened the way it did - I
cannot ascribe any credibility to it as an experiment and it has been
a badly concieved parody of an adversary at work that has been a waste
of time.
I think it should be conducted all over again in a few weeks when I
can table a new database that PM will not be handed on a plate this
time to exploit as he did (not).
That has become a waste of time that is worthless to me or to anybody
- I am not claiming anything from it a a test because it was
worthless.
Let someone in authority stake out what is an acceptable simulation of
an attack by an adversary and I will provide the files for the model.
There must be some criteria about this in the crypto establishment
somewhere that can be used as a standard test reference.
It is truly in my interests to cooperate in such a test because it
will be a small piece of credility to others.
I await the next development with eager interest from anyone.
- adacrypt
wrote:
Alice - This fiasco has spawned more red herrings than enough
That might be the only thing to be claimed out of this faisco - Pm
simulated that situation and my cipher survived.
Because something appeared in wikipedia that doesnt make it a theorem.
I treat everything I read there as just having only temporary
scientific credibility.
What you say implies that AES and RSA may have already been broken in
isolated cases that we know nothing about.
This unfortunate episode should never have happened the way it did - I
cannot ascribe any credibility to it as an experiment and it has been
a badly concieved parody of an adversary at work that has been a waste
of time.
I think it should be conducted all over again in a few weeks when I
can table a new database that PM will not be handed on a plate this
time to exploit as he did (not).
That has become a waste of time that is worthless to me or to anybody
- I am not claiming anything from it a a test because it was
worthless.
Let someone in authority stake out what is an acceptable simulation of
an attack by an adversary and I will provide the files for the model.
There must be some criteria about this in the crypto establishment
somewhere that can be used as a standard test reference.
It is truly in my interests to cooperate in such a test because it
will be a small piece of credility to others.
I await the next development with eager interest from anyone.
- adacrypt
Mark Murray
Oct 23, 2011, 3:36:38 AM10/23/11
to
On 23/10/2011 05:52, adacrypt wrote:
> People is plural - there is only one person being jeopardised, that's
> Alice - This fiasco has spawned more red herrings than enough
Rubbish. If Alice is compromised, then so is Bob. Once their
> People is plural - there is only one person being jeopardised, that's
> Alice - This fiasco has spawned more red herrings than enough
key material is exposed, it is exposed forever.
Do this honestly this time; send PM the files he asked for BUT WITH
SECRET KEY MATERIAL!!!!
Your parameters aren't secret; you have said many times that they
may be sent in the clear; only the vector tables are secret.
> That might be the only thing to be claimed out of this faisco - Pm
> simulated that situation and my cipher survived.
explained how it failed.
> Because something appeared in wikipedia that doesnt make it a theorem.
> I treat everything I read there as just having only temporary
> scientific credibility.
> What you say implies that AES and RSA may have already been broken in
> isolated cases that we know nothing about.
NSA or GCHQ have broken these. They ain't telling.
> This unfortunate episode should never have happened the way it did - I
> cannot ascribe any credibility to it as an experiment and it has been
> a badly concieved parody of an adversary at work that has been a waste
> of time.
fair.
HERE IT IS:
1) Regenerate your secret vector tables. Use a computer to do it so
you can do it again. Ask someone to do this if you can't; it is
REALLY easy.
2) Generate the files that PM asked for INCLUDING the parameters that
he asked for (you swear that they are not relevant to the cipher
anyway!)
3) Send this AND NOT THE SECRET STUFF to PM.
4) Stop whining about PM being "in your database" when thats all
you'll give him!!
> I think it should be conducted all over again in a few weeks when I
> can table a new database that PM will not be handed on a plate this
> time to exploit as he did (not).
supposed to be easy to use? Are you doing it by hand? Can't you
program? Or are you just chicken?
A half-competent newbie ought to be able to do this in under a day.
> That has become a waste of time that is worthless to me or to anybody
> - I am not claiming anything from it a a test because it was
> worthless.
don't like. PM didn't require access to your numbers to formulate his
attack; he just needed what you did with them, and this was easily
broken.
> Let someone in authority stake out what is an acceptable simulation of
> an attack by an adversary and I will provide the files for the model.
a realistic "dialogue" between Alice and Bob. Give a few cribs (they
will always be available, and your cipher MUST be able to survive them)
and stop trying to rig the result.
> There must be some criteria about this in the crypto establishment
> somewhere that can be used as a standard test reference.
> It is truly in my interests to cooperate in such a test because it
> will be a small piece of credility to others.
> I await the next development with eager interest from anyone.
Supply an honest challenge for a start!
adacrypt
Oct 23, 2011, 3:57:26 AM10/23/11
to
On Oct 22, 10:41 pm, rossum <rossu...@coldmail.com> wrote:
> On Sat, 22 Oct 2011 05:50:09 -0700 (PDT), adacrypt
>
>
as one would in a classroom is that the technical points being
exchanged become seemingly weak, evasive, lame and garrulous.
At the risk of being mistakenly taken for that can I table my eventual
idea of how this crypto scheme migh be realised in the long run.
Anybody who has taken the trouble to download my Vector Cipher- 2 will
see that I write several versions of the same program.
Mark_0 is usually a diagnostic / tutorial version that is slow because
it outputs to the screen a full commentary in a blow-by-blow
demonstartion of the internal computations of the cipher as it runs at
work. I like this a lot as a teaching aid and I plan to to continue
it when I go public soon.
Mark_1 is a working version that has no commentary at all. It is very
easily achieved because it just simply means deleting the lines of the
source code that output the commentary and saving under a new name of
Mark_1.
Mark_2 is even easier still because it just means opening Mark_1 ,
renaming it and saving under its new name of Mark_2'
It is ditto for any number of more copies of Mark_1 up to even
Mark_100.
Changing the scrambling parameters within each of these programs means
Alice can call any one of them at any time to encrypt a message and
use it straight off as it stands or change some scarmbling parameters
as she wishes and then use it as she thinks fit.
Finally, I think Kerchoff's desiderata that says the security of the
cipher should depend only on keeping the key set secure is the only
rational goal post that can be met.
Postulating extraordinary exceptions is going beyond the bounds of
reason.
good to hear - adacrypt
Mark_100 scheme is very, very feasible - Alice then prefaces her
ciphertext message with a Mark_x advice note to Bob as the only
necessary parameter passing info that he needs.
adacrypt
Oct 23, 2011, 4:04:17 AM10/23/11
to
him any more information that you suggest he will be able to key in
the correct scrambling parameters, run the program and decrypt the
secret message that I sent him. - how stupid would that be?
Its time to end this farce now and wait for a new appraisal in a fresh
test that does not hand him my secure database.
- adacrypt
Mark Murray
Oct 23, 2011, 4:08:27 AM10/23/11
to
On 23/10/2011 08:57, adacrypt wrote:
>> How do you tell Bob about the changes you have made?
>>
>> rossum
>
:
>> How do you tell Bob about the changes you have made?
>>
>> rossum
>
:
> Mark_100 scheme is very, very feasible - Alice then prefaces her
> ciphertext message with a Mark_x advice note to Bob as the only
> necessary parameter passing info that he needs.
So they abandon the old and move onto a new, meaning effectively
> ciphertext message with a Mark_x advice note to Bob as the only
> necessary parameter passing info that he needs.
they have to keep hundreds of these "Mark_x" things?
How many?
Is the cipher so weak that multiple keys (Oh wait - whole programs!)
need to be kept in order to communicate?
What next? Use each one only once? (That will likely be secure, if
the key material generation is any good). OTP! OTP!!
Mark Murray
Oct 23, 2011, 4:10:36 AM10/23/11
to
On 23/10/2011 09:04, adacrypt wrote:
> I can't - he is inside my database and can run my program - if I give
> him any more information that you suggest he will be able to key in
> the correct scrambling parameters, run the program and decrypt the
> secret message that I sent him. - how stupid would that be?
>
> Its time to end this farce now and wait for a new appraisal in a fresh
> test that does not hand him my secure database.
How dense can you get??!
> I can't - he is inside my database and can run my program - if I give
> him any more information that you suggest he will be able to key in
> the correct scrambling parameters, run the program and decrypt the
> secret message that I sent him. - how stupid would that be?
>
> Its time to end this farce now and wait for a new appraisal in a fresh
> test that does not hand him my secure database.
START AGAIN.
1) NEW DATABASE.
2) NEW MESSAGES.
3) SEND THEM.
How can PM be inside that??
Mark Murray
Oct 23, 2011, 4:18:01 AM10/23/11
to
On 23/10/2011 09:10, Mark Murray wrote:
> On 23/10/2011 09:04, adacrypt wrote:
>> I can't - he is inside my database and can run my program - if I give
>> him any more information that you suggest he will be able to key in
>> the correct scrambling parameters, run the program and decrypt the
>> secret message that I sent him. - how stupid would that be?
>>
>> Its time to end this farce now and wait for a new appraisal in a fresh
>> test that does not hand him my secure database.
>
> How dense can you get??!
>
> START AGAIN.
>
> 1) NEW DATABASE.
>
> 2) NEW MESSAGES.
>
> 3) SEND THEM.
>
> How can PM be inside that??
To be UTTERLY clear:
> On 23/10/2011 09:04, adacrypt wrote:
>> I can't - he is inside my database and can run my program - if I give
>> him any more information that you suggest he will be able to key in
>> the correct scrambling parameters, run the program and decrypt the
>> secret message that I sent him. - how stupid would that be?
>>
>> Its time to end this farce now and wait for a new appraisal in a fresh
>> test that does not hand him my secure database.
>
> How dense can you get??!
>
> START AGAIN.
>
> 1) NEW DATABASE.
>
> 2) NEW MESSAGES.
>
> 3) SEND THEM.
>
> How can PM be inside that??
1) NEW DATABASE
2) NEW MESSAGES
3) ENCRYPT MESSAGES
4) SEND WHAT PM ASKED FOR. DON'T SEND NEW DATABASE.
adacrypt
Oct 23, 2011, 6:24:23 AM10/23/11
to
> Notable nebbish, extreme generalist.- Hide quoted text -
>
> - Show quoted text -
Why can't PM ask for this himself and stake out what files he needs -
>
> - Show quoted text -
this should emulate a typical ciphertext-only attack. It won't have
any scrambling data with it as any other adversary will not have
either - it is a level playing field for everyone then - he shouldn't
ask for extra files that a real world adversay cannot expect.
He's gone very silent - remember he is the protagonist who is issuing
the challeneges - let's not lose sight of that fact.
He was very vociferous recently when admonishing every body not to
answer me on any account.
I felt like changing my nom-de-plume to the "Irish Pariah".
- adacrypt
Mark Murray
Oct 23, 2011, 6:33:16 AM10/23/11
to
On 23/10/2011 11:24, adacrypt wrote:
> Why can't PM ask for this himself and stake out what files he needs -
HE DID!! YOU HAVEN'T BLOODY GIVEN THIS TO HIM!!!
> Why can't PM ask for this himself and stake out what files he needs -
> this should emulate a typical ciphertext-only attack. It won't have
> any scrambling data with it as any other adversary will not have
> either - it is a level playing field for everyone then - he shouldn't
> ask for extra files that a real world adversay cannot expect.
know what a crib is you have no business writing a cryptosystem.
> He's gone very silent - remember he is the protagonist who is issuing
> the challeneges - let's not lose sight of that fact.
busy with real work and is sick of your unco-operative nonsense.
> He was very vociferous recently when admonishing every body not to
> answer me on any account.
toss rather han help folks help you.
> I felt like changing my nom-de-plume to the "Irish Pariah".
Mark Murray
Oct 23, 2011, 6:48:52 AM10/23/11
to
On 23/10/2011 11:24, adacrypt wrote:
> Why can't PM ask for this himself and stake out what files he needs -
> this should emulate a typical ciphertext-only attack. It won't have
> any scrambling data with it as any other adversary will not have
> either - it is a level playing field for everyone then - he shouldn't
> ask for extra files that a real world adversay cannot expect.
You cannot say to an adversary (and therefore a cryptanalyst)
> this should emulate a typical ciphertext-only attack. It won't have
> any scrambling data with it as any other adversary will not have
> either - it is a level playing field for everyone then - he shouldn't
> ask for extra files that a real world adversay cannot expect.
"You may only attack my cipher in ways that I approve of". The
attacker may attack any way (s)he likes. In the real world, cribs
WILL be available, and for your cipher to be any good, you need
to be resilient to them, and you can show good faith and confidence
in your own work by providing them up front. You are doing the
opposite by being devious in your behaviour.
Right now, the way you are behaving, you are tacitly admitting
that if a sloppy operator were to leak (even partial) plaintext
that this would permanently compromise the "mutual database".
If this is not the case, let folks try-and-fail.
adacrypt
Oct 23, 2011, 6:58:19 AM10/23/11
to
to him yesterday. Your'e talking about a repeat set - What's wrong
with him that he can't give me the results himself of the first
experiment that started yesterday - common decency demands that he
should reply.
Other Readers.
What PM is trying to do is to launch a Kasiski - Babbage attack on my
ciphertext - it might also double as a numerical attack and indeed a
statistical attack also but essentially it is an attack that was
developed on scalar data. It does'nt work on vector data although the
Kasiski-Babbage attack arguably does have a bearing.
My Contention.
My vector cryptography is crafted to defeat all of these scalar-
related attacks - the methodology requires a thorough understanding of
vector methods and plane geometry. PM seems to believe that he can
circumvent these convoluted decryption methods and instead demonstrate
relationships between the vector ciphertext and the scalar plaintext
that will enable him to map the ciphertext directly and unerringly to
the corresponding plaintext.
I say good luck to him.
I want it from the horse's mouth what the conditions of the next test
will be - he has failed this one despite having a huge amount of extra
advatage going for him.
MM stop doing his talking for him.
- adacrypt
Mark Murray
Oct 23, 2011, 7:42:49 AM10/23/11
to
On 23/10/2011 11:58, adacrypt wrote:
> No way - he still hasn't acknowledge the results of the first lot sent
> to him yesterday. Your'e talking about a repeat set - What's wrong
> with him that he can't give me the results himself of the first
> experiment that started yesterday - common decency demands that he
> should reply.
Its weekend. Maybe he's not doing work? Get real.
> No way - he still hasn't acknowledge the results of the first lot sent
> to him yesterday. Your'e talking about a repeat set - What's wrong
> with him that he can't give me the results himself of the first
> experiment that started yesterday - common decency demands that he
> should reply.
> Other Readers.
>
> What PM is trying to do is to launch a Kasiski - Babbage attack on my
> ciphertext - it might also double as a numerical attack and indeed a
> statistical attack also but essentially it is an attack that was
> developed on scalar data. It does'nt work on vector data although the
> Kasiski-Babbage attack arguably does have a bearing.
> My Contention.
>
> My vector cryptography is crafted to defeat all of these scalar-
> related attacks - the methodology requires a thorough understanding of
> vector methods and plane geometry. PM seems to believe that he can
> circumvent these convoluted decryption methods and instead demonstrate
> relationships between the vector ciphertext and the scalar plaintext
> that will enable him to map the ciphertext directly and unerringly to
> the corresponding plaintext.
you appear not to understand or appreciate.
> I say good luck to him.
>
> I want it from the horse's mouth what the conditions of the next test
> will be - he has failed this one despite having a huge amount of extra
> advatage going for him.
already!! You seem to dead scared of providing anything other than
tightly controlled tests with rigged results.
> MM stop doing his talking for him.
want public responses, don't make public statements.
Bruce Stephens
Oct 23, 2011, 9:57:40 AM10/23/11
to
adacrypt <austin...@hotmail.com> writes:
[...]
> My Contention.
>
> My vector cryptography is crafted to defeat all of these scalar-
> related attacks - the methodology requires a thorough understanding of
> vector methods and plane geometry. PM seems to believe that he can
> circumvent these convoluted decryption methods and instead demonstrate
> relationships between the vector ciphertext and the scalar plaintext
> that will enable him to map the ciphertext directly and unerringly to
> the corresponding plaintext.
Alternatively, your cryptography relies on having a large enough key for
its security---if the key is large enough that you can encrypt all
messages without reusing any of it, then the scheme has perfect
security.
The vector stuff is mere distraction, adding some obfuscation while
making the system more complex and much less efficient.
[...]
[...]
> My Contention.
>
> My vector cryptography is crafted to defeat all of these scalar-
> related attacks - the methodology requires a thorough understanding of
> vector methods and plane geometry. PM seems to believe that he can
> circumvent these convoluted decryption methods and instead demonstrate
> relationships between the vector ciphertext and the scalar plaintext
> that will enable him to map the ciphertext directly and unerringly to
> the corresponding plaintext.
its security---if the key is large enough that you can encrypt all
messages without reusing any of it, then the scheme has perfect
security.
The vector stuff is mere distraction, adding some obfuscation while
making the system more complex and much less efficient.
[...]
adacrypt
Oct 23, 2011, 11:19:09 AM10/23/11
to
On Oct 23, 2:57 pm, Bruce Stephens <bruce+use...@cenderis.demon.co.uk>
wrote:
Naw,
It is quite hard to see how vector ciphertext is better for
cryptography than scalars. - another time - it requires a very good
understanding of vectors.
- adacrypt
wrote:
It is quite hard to see how vector ciphertext is better for
cryptography than scalars. - another time - it requires a very good
understanding of vectors.
- adacrypt
Mark Murray
Oct 23, 2011, 11:24:46 AM10/23/11
to
On 23/10/2011 16:19, adacrypt wrote:
> It is quite hard to see how vector ciphertext is better for
> cryptography than scalars. - another time - it requires a very good
> understanding of vectors.
You keep returning to this.
> It is quite hard to see how vector ciphertext is better for
> cryptography than scalars. - another time - it requires a very good
> understanding of vectors.
Folks understanding of vectors is fine; its you who seems to
think they add anything to the cipher, in spite of PM showing
how they can be removed.
This is not a matter of difficulty, it is a matter of disagreement.
Bruce Stephens
Oct 23, 2011, 11:37:21 AM10/23/11
to
adacrypt <austin...@hotmail.com> writes:
[...]
> Naw,
>
> It is quite hard to see how vector ciphertext is better for
> cryptography than scalars. - another time - it requires a very good
> understanding of vectors.
So separate the issues: construct a cipher that only uses vector
[...]
> Naw,
>
> It is quite hard to see how vector ciphertext is better for
> cryptography than scalars. - another time - it requires a very good
> understanding of vectors.
cryptography and uses a small key (of fixed size, much smaller than the
plaintext).
adacrypt
Oct 23, 2011, 12:08:18 PM10/23/11
to
On Oct 23, 4:37 pm, Bruce Stephens <bruce+use...@cenderis.demon.co.uk>
wrote:
Why Vector Ciphertext is Better.
If the domain of a key set is made up of n elements of non-repeating
vectors then there are n of the ‘i’ unit vectors, n of the ‘j’ unit
vectors and n of the ‘k’ unit vectors. These can be arranged to give
n x n x n or n^3 different larger vectors that may be used as the
primary transformation of a single plaintext character. Each of the
larger keyset of n cubed vectors is totally disparate and independent
of any other element in the same set of n^3 so that there is no
mathematical induction assumptions whatever unlike scalar numbers that
can be made of a long string of ciphertext by a cryptanalyst.
Fpr instance, given say the rth element he cannot deduce the (r + p)
element because individual elements when known, are only related by
irrelevant geometry in cryptography – there is no number theory as in
scalar data.
In scalar cryptography on the other hand there is some functional
relationship between the plaintext and its transformation into
ciphertext usually and the separate elements of the string of
ciphertext is only ‘n’ in number compared to n^3 in vector
cryptography. The elements of the ciphertext string are always
mathematically related to each other in some way however difficult it
may be to find this. The fact is that numerical methods are there for
the finding of this relationship together with statistical and other
direct inspection methods like Kasiski-Babbage that cryptanalysts may
use.
The situation is fraught with what I call ‘structure’ that a
cryptanalyst can exploit.
Pound for pound therefore vector data is more effective in capacity
and is much more secure against cryptanalysis.
That’s my view.
- adacrypt
wrote:
If the domain of a key set is made up of n elements of non-repeating
vectors then there are n of the ‘i’ unit vectors, n of the ‘j’ unit
vectors and n of the ‘k’ unit vectors. These can be arranged to give
n x n x n or n^3 different larger vectors that may be used as the
primary transformation of a single plaintext character. Each of the
larger keyset of n cubed vectors is totally disparate and independent
of any other element in the same set of n^3 so that there is no
mathematical induction assumptions whatever unlike scalar numbers that
can be made of a long string of ciphertext by a cryptanalyst.
Fpr instance, given say the rth element he cannot deduce the (r + p)
element because individual elements when known, are only related by
irrelevant geometry in cryptography – there is no number theory as in
scalar data.
In scalar cryptography on the other hand there is some functional
relationship between the plaintext and its transformation into
ciphertext usually and the separate elements of the string of
ciphertext is only ‘n’ in number compared to n^3 in vector
cryptography. The elements of the ciphertext string are always
mathematically related to each other in some way however difficult it
may be to find this. The fact is that numerical methods are there for
the finding of this relationship together with statistical and other
direct inspection methods like Kasiski-Babbage that cryptanalysts may
use.
The situation is fraught with what I call ‘structure’ that a
cryptanalyst can exploit.
Pound for pound therefore vector data is more effective in capacity
and is much more secure against cryptanalysis.
That’s my view.
- adacrypt
Bruce Stephens
Oct 23, 2011, 12:14:56 PM10/23/11
to
adacrypt <austin...@hotmail.com> writes:
[...]
> Why Vector Ciphertext is Better.
>
> If the domain of a key set is made up of n elements of non-repeating
> vectors then there are n of the ‘i’ unit vectors, n of the ‘j’ unit
> vectors and n of the ‘k’ unit vectors. These can be arranged to give
> n x n x n or n^3 different larger vectors that may be used as the
> primary transformation of a single plaintext character. Each of the
> larger keyset of n cubed vectors is totally disparate and independent
> of any other element in the same set of n^3 so that there is no
> mathematical induction assumptions whatever unlike scalar numbers that
> can be made of a long string of ciphertext by a cryptanalyst.
What's n? Must it be at least the size of the plaintext, or might it be
1, say? Or 5?
[...]
> The situation is fraught with what I call ‘structure’ that a
> cryptanalyst can exploit.
But you must be wrong, since we know the one-time-pad has perfect
security.
[...]
[...]
> Why Vector Ciphertext is Better.
>
> If the domain of a key set is made up of n elements of non-repeating
> vectors then there are n of the ‘i’ unit vectors, n of the ‘j’ unit
> vectors and n of the ‘k’ unit vectors. These can be arranged to give
> n x n x n or n^3 different larger vectors that may be used as the
> primary transformation of a single plaintext character. Each of the
> larger keyset of n cubed vectors is totally disparate and independent
> of any other element in the same set of n^3 so that there is no
> mathematical induction assumptions whatever unlike scalar numbers that
> can be made of a long string of ciphertext by a cryptanalyst.
1, say? Or 5?
[...]
> The situation is fraught with what I call ‘structure’ that a
> cryptanalyst can exploit.
security.
[...]
Mark Murray
Oct 23, 2011, 12:18:28 PM10/23/11
to
On 23/10/2011 17:08, adacrypt wrote:
> That’s my view.
He asked you do do something, not to explain (yet again) why you think
your cipher is secure. Its reasonable to assume that people understand
your idea. Now try to understand theirs.
Here it the request:
<quote author="Bruce Stephens">
This is not a hard task; it is a demonstation of what you claim to be
the essential security of of your cipher.
Let the excuses roll.
> That’s my view.
He asked you do do something, not to explain (yet again) why you think
your cipher is secure. Its reasonable to assume that people understand
your idea. Now try to understand theirs.
Here it the request:
<quote author="Bruce Stephens">
So separate the issues: construct a cipher that only uses vector
cryptography and uses a small key (of fixed size, much smaller than the
plaintext).
</quote>
cryptography and uses a small key (of fixed size, much smaller than the
plaintext).
This is not a hard task; it is a demonstation of what you claim to be
the essential security of of your cipher.
Let the excuses roll.
rossum
Oct 23, 2011, 12:27:44 PM10/23/11
to
On Sat, 22 Oct 2011 16:47:03 -0700 (PDT), adacrypt
<austin...@hotmail.com> wrote:
>Asking for two files of plaintext as he did on top of being inside the
>database is laughable. He would make a worse thief than he is as a
>cryptanalyst.
You really need to read some history. The attacker can *always* find
<austin...@hotmail.com> wrote:
>Asking for two files of plaintext as he did on top of being inside the
>database is laughable. He would make a worse thief than he is as a
>cryptanalyst.
some plaintext/cyphertext pairs. It is so common that there is even a
name for this sort of attack: the "known plaintext attack". If your
cypher cannot withstand a known plaintext attack then it is too weak
to be of any use.
Now some history. The day before the Pearl Harbour attack, Tokyo sent
a long cyphertext to its embassy in Washington. The next day the
Japanese Ambassador handed a long plaintext statement to the US
Government. A perfect plaintext/cyphertext pair.
The fact that the US had already broken the Japanese code is
irrelevant; the attacker can *always* find a plaintext/cyphertext
pair. If your cypher cannot withstand a known plaintext attack, then
it is weaker than AES, which has been tested against such attacks.
rossum
adacrypt
Oct 23, 2011, 12:40:58 PM10/23/11
to
On Oct 23, 5:27 pm, rossum <rossu...@coldmail.com> wrote:
> On Sat, 22 Oct 2011 16:47:03 -0700 (PDT), adacrypt
>
> On Sat, 22 Oct 2011 16:47:03 -0700 (PDT), adacrypt
>
being prevaricated and misquoted in posts - time will tell all -
adacrypt
Mark Murray
Oct 23, 2011, 12:50:25 PM10/23/11
to
On 23/10/2011 17:40, adacrypt wrote:
> This needs a lot of chalk and talk - it is too much the way things are
> being prevaricated and misquoted in posts - time will tell all -
You quote that "time will tell all" as some kind of mantra.
> This needs a lot of chalk and talk - it is too much the way things are
> being prevaricated and misquoted in posts - time will tell all -
Time HAS told, and sufficiently too.
You are a joke if you can't see that.
Bruce Stephens
Oct 23, 2011, 1:27:10 PM10/23/11
to
rossum <ross...@coldmail.com> writes:
> On Sat, 22 Oct 2011 16:47:03 -0700 (PDT), adacrypt
> <austin...@hotmail.com> wrote:
>
>>Asking for two files of plaintext as he did on top of being inside the
>>database is laughable. He would make a worse thief than he is as a
>>cryptanalyst.
> You really need to read some history. The attacker can *always* find
> some plaintext/cyphertext pairs. It is so common that there is even a
> name for this sort of attack: the "known plaintext attack". If your
> cypher cannot withstand a known plaintext attack then it is too weak
> to be of any use.
IIUC his complaint was about the "database", i.e., the key. We don't
> On Sat, 22 Oct 2011 16:47:03 -0700 (PDT), adacrypt
> <austin...@hotmail.com> wrote:
>
>>Asking for two files of plaintext as he did on top of being inside the
>>database is laughable. He would make a worse thief than he is as a
>>cryptanalyst.
> You really need to read some history. The attacker can *always* find
> some plaintext/cyphertext pairs. It is so common that there is even a
> name for this sort of attack: the "known plaintext attack". If your
> cypher cannot withstand a known plaintext attack then it is too weak
> to be of any use.
(AFAIK) expect a scheme to survive known-key attacks.
The fact that he messed up in this way originally perhaps suggests that
there's some problem with embedding the key in the program (maybe if it
were separate then this mistake would be less likely). However, such
mistakes aren't unknown in the wild (people sending PKCS#12 files
including private key when asked to send their certificate, for
example).
[...]
Mark Murray
Oct 23, 2011, 1:41:11 PM10/23/11
to
On 23/10/2011 18:27, Bruce Stephens wrote:
> The fact that he messed up in this way originally perhaps suggests that
> there's some problem with embedding the key in the program (maybe if it
> were separate then this mistake would be less likely).
The OP has a very big conceptual problem in this regard; he doesn't seem
> The fact that he messed up in this way originally perhaps suggests that
> there's some problem with embedding the key in the program (maybe if it
> were separate then this mistake would be less likely).
to have a clear idea about which bits of his downloadable package are
secret, public, mutable or immutable, and at times, these
classifications have changed to suit the moment.
adacrypt
Oct 23, 2011, 2:06:16 PM10/23/11
to
On Oct 23, 5:14 pm, Bruce Stephens <bruce+use...@cenderis.demon.co.uk>
wrote:
Think of 'n' as a keyset that is in fact an array of many elemental
vectors - in the cipher that I am preparing for later, the domain of
change-of-origin vectors for instance has 14250 vectors in it. This
is one of several silmilar but smaller domains of other keysets also.
There are no 'single' scalar keys as such.
Elemental vectors are called sequentially from the array of 14250 to
be included in each computation of a single ciphertext that represents
the transformation of one plaintext into ciphertext. It follows that
there will be the same length of keyset (i.e. as a subset of the full
14250 array) as the message length - this is something that simply
happens as a fallout of the design - there is tecnical requirement
that this should be so - it just happens that way - there is no
special security significance to it.
Other elemental keys are called from other arrays also - its so on to
the completion of the encryption of the entire message.
- Cheers - adacrypt
wrote:
vectors - in the cipher that I am preparing for later, the domain of
change-of-origin vectors for instance has 14250 vectors in it. This
is one of several silmilar but smaller domains of other keysets also.
There are no 'single' scalar keys as such.
Elemental vectors are called sequentially from the array of 14250 to
be included in each computation of a single ciphertext that represents
the transformation of one plaintext into ciphertext. It follows that
there will be the same length of keyset (i.e. as a subset of the full
14250 array) as the message length - this is something that simply
happens as a fallout of the design - there is tecnical requirement
that this should be so - it just happens that way - there is no
special security significance to it.
Other elemental keys are called from other arrays also - its so on to
the completion of the encryption of the entire message.
- Cheers - adacrypt
Bruce Stephens
Oct 23, 2011, 2:26:44 PM10/23/11
to
adacrypt <austin...@hotmail.com> writes:
[...]
> Elemental vectors are called sequentially from the array of 14250 to
Why 14250? Could one create a cipher using instead just 1, and do you
think it would be similarly secure?
I suspect not, and that in fact the larger the number the more secure
until the number's the size of the message (and beyond that you don't
gain anything). So with 14250, if the message being encrypted is no
longer than 14250 I'd guess that your cipher has perfect security (I'm
not confident of that---you might have screwed it up), but for longer
messages I'd expect it not to have perfect security (since you're
reusing "elemental vectors").
[...]
[...]
> Elemental vectors are called sequentially from the array of 14250 to
think it would be similarly secure?
I suspect not, and that in fact the larger the number the more secure
until the number's the size of the message (and beyond that you don't
gain anything). So with 14250, if the message being encrypted is no
longer than 14250 I'd guess that your cipher has perfect security (I'm
not confident of that---you might have screwed it up), but for longer
messages I'd expect it not to have perfect security (since you're
reusing "elemental vectors").
[...]
adacrypt
Oct 23, 2011, 2:40:41 PM10/23/11
to
On Oct 23, 7:26 pm, Bruce Stephens <bruce+use...@cenderis.demon.co.uk>
wrote:
There is no significance to the number 14250 - This was old data that
I was refurbishing now for a different use - it just happened to be
handy to use it again. The security has nothing to do with the array
length of 14250 or indeed any size of number - Also I wrap back to the
beginning of the array with impunity if I want go higher than 14250
characters in messagelength - the security is all down to the one-way
function provided by the change-of-origin vector - (more on all of
this requires a lot of chalk n' talk).
This has to all from me for today.
Cheers - adacrypt
wrote:
I was refurbishing now for a different use - it just happened to be
handy to use it again. The security has nothing to do with the array
length of 14250 or indeed any size of number - Also I wrap back to the
beginning of the array with impunity if I want go higher than 14250
characters in messagelength - the security is all down to the one-way
function provided by the change-of-origin vector - (more on all of
this requires a lot of chalk n' talk).
This has to all from me for today.
Cheers - adacrypt
Bruce Stephens
Oct 23, 2011, 2:43:22 PM10/23/11
to
adacrypt <austin...@hotmail.com> writes:
[...]
> There is no significance to the number 14250 - This was old data that
> I was refurbishing now for a different use - it just happened to be
> handy to use it again. The security has nothing to do with the array
> length of 14250 or indeed any size of number
OK, so use 1 instead. Much simpler.
[...]
> There is no significance to the number 14250 - This was old data that
> I was refurbishing now for a different use - it just happened to be
> handy to use it again. The security has nothing to do with the array
> length of 14250 or indeed any size of number
[...]
Greg Rose
Oct 23, 2011, 3:42:00 PM10/23/11
to
David Eather
Oct 23, 2011, 4:51:25 PM10/23/11
to
On 23/10/2011 6:04 PM, adacrypt wrote:
> On Oct 23, 8:36 am, Mark Murray<w.h.o...@example.com> wrote:
>> On 23/10/2011 05:52, adacrypt wrote:
>>
>>> People is plural - there is only one person being jeopardised, that's
>>> Alice - This fiasco has spawned more red herrings than enough
>>
>> Rubbish. If Alice is compromised, then so is Bob. Once their
>> key material is exposed, it is exposed forever.
>>
>> Do this honestly this time; send PM the files he asked for BUT WITH
>> SECRET KEY MATERIAL!!!!
>>
>> Your parameters aren't secret; you have said many times that they
>> may be sent in the clear; only the vector tables are secret.
>>
>>> That might be the only thing to be claimed out of this faisco - Pm
>>> simulated that situation and my cipher survived.
>>
>> PM used your caipher as you told him to and it failed, and he
>> explained how it failed.
>>
>>> Because something appeared in wikipedia that doesnt make it a theorem.
>>
>> No, but its evidence.
>>
>>> I treat everything I read there as just having only temporary
>>> scientific credibility.
>>
>> Rubbish. You ignore facts.
>>
>>> What you say implies that AES and RSA may have already been broken in
>>> isolated cases that we know nothing about.
>>
>> Its (in theory) possible that large government organizations like
>> NSA or GCHQ have broken these. They ain't telling.
>>
>>> This unfortunate episode should never have happened the way it did - I
>>> cannot ascribe any credibility to it as an experiment and it has been
>>> a badly concieved parody of an adversary at work that has been a waste
>>> of time.
>>
>> The only parody of an adversary here is you and your refusal to play
>> fair.
>>
>> HERE IT IS:
>>
>> 1) Regenerate your secret vector tables. Use a computer to do it so
>> you can do it again. Ask someone to do this if you can't; it is
>> REALLY easy.
>>
>> 2) Generate the files that PM asked for INCLUDING the parameters that
>> he asked for (you swear that they are not relevant to the cipher
>> anyway!)
>>
>> 3) Send this AND NOT THE SECRET STUFF to PM.
>>
>> 4) Stop whining about PM being "in your database" when thats all
>> you'll give him!!
>>
>>> I think it should be conducted all over again in a few weeks when I
>>> can table a new database that PM will not be handed on a plate this
>>> time to exploit as he did (not).
>>
>> Why does it take a few weks to table a new database? This thing is
>> supposed to be easy to use? Are you doing it by hand? Can't you
>> program? Or are you just chicken?
>>
>> A half-competent newbie ought to be able to do this in under a day.
>>
>>> That has become a waste of time that is worthless to me or to anybody
>>> - I am not claiming anything from it a a test because it was
>>> worthless.
>>
>> I get the impression that you will declare orthless anything that you
>> don't like. PM didn't require access to your numbers to formulate his
>> attack; he just needed what you did with them, and this was easily
>> broken.
>>
>>> Let someone in authority stake out what is an acceptable simulation of
>>> an attack by an adversary and I will provide the files for the model.
>>
>> Read a book on cryptananlysis for crying out loud! Provide _IN_PUBLIC_
>> a realistic "dialogue" between Alice and Bob. Give a few cribs (they
>> will always be available, and your cipher MUST be able to survive them)
>> and stop trying to rig the result.
>>
>>> There must be some criteria about this in the crypto establishment
>>> somewhere that can be used as a standard test reference.
>>
>> YES!! Start reading books! Its all there!
>>
>>> It is truly in my interests to cooperate in such a test because it
>>> will be a small piece of credility to others.
>>
>> YES!! So start doing so!!
>>
>>> I await the next development with eager interest from anyone.
>>
>> NO!! YOU make the next development! It's YOU who aren't doing his bit!
>>
>> Supply an honest challenge for a start!
Perhaps it will help to think of this situation in an analogous real
world way. A cipher performs on data essentially the same function as a
lock performs on a door. Heaps of people in the street/suburb/country
use the exact same lock. Even the thief can buy one and study it. The
thief can pull it apart, examine and measure it in exact detail. The
thief can even practice breaking simpler versions of the lock to build
expertise. The thief knows absolutely everything about your lock and can
practice an infinite number of times, but your lock is still secure
because the *only* *thing* the thief does not know is the key. It is
exactly the same way with strong ciphers - especially if you intend for
it to be widely used.
You do realise that if you had used AES instead of your stuff PM would
have had the entire, complete and exact details as well. If this
identical attack, which you consider unfair, was inflicted on AES it
would have survived unscathed, where as your cipher is cracked wide
open. It is an elegant proof that AES is much stronger than your cipher,
no?
> On Oct 23, 8:36 am, Mark Murray<w.h.o...@example.com> wrote:
>> On 23/10/2011 05:52, adacrypt wrote:
>>
>>> People is plural - there is only one person being jeopardised, that's
>>> Alice - This fiasco has spawned more red herrings than enough
>>
>> Rubbish. If Alice is compromised, then so is Bob. Once their
>> key material is exposed, it is exposed forever.
>>
>> Do this honestly this time; send PM the files he asked for BUT WITH
>> SECRET KEY MATERIAL!!!!
>>
>> Your parameters aren't secret; you have said many times that they
>> may be sent in the clear; only the vector tables are secret.
>>
>>> That might be the only thing to be claimed out of this faisco - Pm
>>> simulated that situation and my cipher survived.
>>
>> PM used your caipher as you told him to and it failed, and he
>> explained how it failed.
>>
>>> Because something appeared in wikipedia that doesnt make it a theorem.
>>
>> No, but its evidence.
>>
>>> I treat everything I read there as just having only temporary
>>> scientific credibility.
>>
>> Rubbish. You ignore facts.
>>
>>> What you say implies that AES and RSA may have already been broken in
>>> isolated cases that we know nothing about.
>>
>> Its (in theory) possible that large government organizations like
>> NSA or GCHQ have broken these. They ain't telling.
>>
>>> This unfortunate episode should never have happened the way it did - I
>>> cannot ascribe any credibility to it as an experiment and it has been
>>> a badly concieved parody of an adversary at work that has been a waste
>>> of time.
>>
>> The only parody of an adversary here is you and your refusal to play
>> fair.
>>
>> HERE IT IS:
>>
>> 1) Regenerate your secret vector tables. Use a computer to do it so
>> you can do it again. Ask someone to do this if you can't; it is
>> REALLY easy.
>>
>> 2) Generate the files that PM asked for INCLUDING the parameters that
>> he asked for (you swear that they are not relevant to the cipher
>> anyway!)
>>
>> 3) Send this AND NOT THE SECRET STUFF to PM.
>>
>> 4) Stop whining about PM being "in your database" when thats all
>> you'll give him!!
>>
>>> I think it should be conducted all over again in a few weeks when I
>>> can table a new database that PM will not be handed on a plate this
>>> time to exploit as he did (not).
>>
>> Why does it take a few weks to table a new database? This thing is
>> supposed to be easy to use? Are you doing it by hand? Can't you
>> program? Or are you just chicken?
>>
>> A half-competent newbie ought to be able to do this in under a day.
>>
>>> That has become a waste of time that is worthless to me or to anybody
>>> - I am not claiming anything from it a a test because it was
>>> worthless.
>>
>> I get the impression that you will declare orthless anything that you
>> don't like. PM didn't require access to your numbers to formulate his
>> attack; he just needed what you did with them, and this was easily
>> broken.
>>
>>> Let someone in authority stake out what is an acceptable simulation of
>>> an attack by an adversary and I will provide the files for the model.
>>
>> Read a book on cryptananlysis for crying out loud! Provide _IN_PUBLIC_
>> a realistic "dialogue" between Alice and Bob. Give a few cribs (they
>> will always be available, and your cipher MUST be able to survive them)
>> and stop trying to rig the result.
>>
>>> There must be some criteria about this in the crypto establishment
>>> somewhere that can be used as a standard test reference.
>>
>> YES!! Start reading books! Its all there!
>>
>>> It is truly in my interests to cooperate in such a test because it
>>> will be a small piece of credility to others.
>>
>> YES!! So start doing so!!
>>
>>> I await the next development with eager interest from anyone.
>>
>> NO!! YOU make the next development! It's YOU who aren't doing his bit!
>>
>> Supply an honest challenge for a start!
>>
>> M
>> --
>> Mark "No Nickname" Murray
>> Notable nebbish, extreme generalist.
>
>> M
>> --
>> Mark "No Nickname" Murray
>> Notable nebbish, extreme generalist.
>
> I can't - he is inside my database and can run my program - if I give
> him any more information that you suggest he will be able to key in
> the correct scrambling parameters, run the program and decrypt the
> secret message that I sent him. - how stupid would that be?
>
> Its time to end this farce now and wait for a new appraisal in a fresh
> test that does not hand him my secure database.
>
> - adacrypt
> him any more information that you suggest he will be able to key in
> the correct scrambling parameters, run the program and decrypt the
> secret message that I sent him. - how stupid would that be?
>
> Its time to end this farce now and wait for a new appraisal in a fresh
> test that does not hand him my secure database.
>
Perhaps it will help to think of this situation in an analogous real
world way. A cipher performs on data essentially the same function as a
lock performs on a door. Heaps of people in the street/suburb/country
use the exact same lock. Even the thief can buy one and study it. The
thief can pull it apart, examine and measure it in exact detail. The
thief can even practice breaking simpler versions of the lock to build
expertise. The thief knows absolutely everything about your lock and can
practice an infinite number of times, but your lock is still secure
because the *only* *thing* the thief does not know is the key. It is
exactly the same way with strong ciphers - especially if you intend for
it to be widely used.
You do realise that if you had used AES instead of your stuff PM would
have had the entire, complete and exact details as well. If this
identical attack, which you consider unfair, was inflicted on AES it
would have survived unscathed, where as your cipher is cracked wide
open. It is an elegant proof that AES is much stronger than your cipher,
no?
David Eather
Oct 23, 2011, 4:59:43 PM10/23/11
to
course!) to a free file hosting facility like "the file factory", "file
dropper" or "rapid share" - there are heaps of others - and post the
link on sci.crypt. (file dropper is the easiest to use, but the file
factory is almost the same)
In this way there is absolute proof you sent the files and a valid
attack or failure can be confirmed by anyone who wants to check.
This solves any "I sent" , "no he didn't" problems, costs you nothing
and otherwise works well for you doesn't it?
David Eather
Oct 23, 2011, 5:09:12 PM10/23/11
to
Noob
Oct 24, 2011, 7:56:17 AM10/24/11
to
Mark Murray wrote:
> Let the excuses roll.
I thought it was the good times?
> Let the excuses roll.
I thought it was the good times?
Paulo Marques
Oct 24, 2011, 12:45:53 PM10/24/11
to
adacrypt wrote:
> [...]
> He hasn’t taken the trouble to understand this cryptography and
> instead argues he can just use outdated One-Time Pad logic.
>
> I can’t get him to come out of the ladies toilet to confront him with
> some samples of plaintext and cipher text to see him demonstrate his
> claims.
Man, I go "off the grid" for 3 days (2 of them being weekend, when I'm
usually off the grid anyway) and you get all cocky and smug. What kind
of insult is "come out of the ladies toilet" anyway?
As Mark suspected, the files you sent me are built from some new
algorithm that you haven't published yet and these discussions are
pointless.
To Mark and Bruce: I appreciate your support, but your efforts would
work only on someone capable of a rational discussion. Adacrypt has
already proven on more than one occasion that he doesn't belong in that
group of people.
There is only one alternative here: to write down the _exact_ rules of a
proper challenge and have you commit to it.
1 - the challenge must respect Kerckhoffs's principle. This means that
you have to publish the _exact_ algorithm that you are going to use to
generate the challenge files. You must also make very clear what is
"secret" and that you are allowed to change without disclosing, and what
is sent in each message. In a normal "block cypher" or "stream cypher",
the communications protocols built on top of those are already well
understood and are implicit in any test. In your case, you have to
clarify the communications protocol, too. You have to write out exactly
what Alice must do when she wants to send a few messages and what Bob
must do to receive them.
2 - the challenge files should be the same as requested before. You must
stop with the "known plaintext attacks are too hard" nonsense. Many
devices in the real world have to survive known plaintext attacks
(set-top boxes, satellite receivers, cell phones, etc., etc.) and all
the decent algorithms survive them perfectly well. Since we're making
everything cristal clear: you need to publish 2 plaintexts with their
respective ciphertexts and a ciphertext that you don't publish the
plaintext of. If we can recover the plaintext, you lose the challenge.
Each of the plaintexts should be at least 4kB in size.
3 - all the files (the program + challenge data, not secret data) must
be placed in a zip file and offered for download from your site, with
the title "challenge files" (or something like that) so that ANYONE can
download and see the files. No more of this "I've already sent the
files" and "PM says the files are so and so, but no one can confirm" crap.
4 - we need an incentive: you have to promise on your honor that if your
cypher doesn't withstand the challenge you'll stop posting your
ramblings about vector cryptography on sci.crypt. No more "You did that,
but that was just because of a small mistake. The underlying principle
is intact" crap.
If you feel that your current algorithm is not up to task and only your
future algorithm is ok, then work on the future algorithm instead of
wasting everybody's time with your lengthy posts. It's not worth it to
write so much about something that you can not back with real evidence.
Until you have real evidence, just go work on it and leave sci.crypt alone.
--
Paulo Marques - www.grupopie.com
"And folks, let's be honest. Sturgeon was an optimist.
Way more than 90% of code is crap."
Al Viro
> [...]
> He hasn’t taken the trouble to understand this cryptography and
> instead argues he can just use outdated One-Time Pad logic.
>
> I can’t get him to come out of the ladies toilet to confront him with
> some samples of plaintext and cipher text to see him demonstrate his
> claims.
Man, I go "off the grid" for 3 days (2 of them being weekend, when I'm
usually off the grid anyway) and you get all cocky and smug. What kind
of insult is "come out of the ladies toilet" anyway?
As Mark suspected, the files you sent me are built from some new
algorithm that you haven't published yet and these discussions are
pointless.
To Mark and Bruce: I appreciate your support, but your efforts would
work only on someone capable of a rational discussion. Adacrypt has
already proven on more than one occasion that he doesn't belong in that
group of people.
There is only one alternative here: to write down the _exact_ rules of a
proper challenge and have you commit to it.
1 - the challenge must respect Kerckhoffs's principle. This means that
you have to publish the _exact_ algorithm that you are going to use to
generate the challenge files. You must also make very clear what is
"secret" and that you are allowed to change without disclosing, and what
is sent in each message. In a normal "block cypher" or "stream cypher",
the communications protocols built on top of those are already well
understood and are implicit in any test. In your case, you have to
clarify the communications protocol, too. You have to write out exactly
what Alice must do when she wants to send a few messages and what Bob
must do to receive them.
2 - the challenge files should be the same as requested before. You must
stop with the "known plaintext attacks are too hard" nonsense. Many
devices in the real world have to survive known plaintext attacks
(set-top boxes, satellite receivers, cell phones, etc., etc.) and all
the decent algorithms survive them perfectly well. Since we're making
everything cristal clear: you need to publish 2 plaintexts with their
respective ciphertexts and a ciphertext that you don't publish the
plaintext of. If we can recover the plaintext, you lose the challenge.
Each of the plaintexts should be at least 4kB in size.
3 - all the files (the program + challenge data, not secret data) must
be placed in a zip file and offered for download from your site, with
the title "challenge files" (or something like that) so that ANYONE can
download and see the files. No more of this "I've already sent the
files" and "PM says the files are so and so, but no one can confirm" crap.
4 - we need an incentive: you have to promise on your honor that if your
cypher doesn't withstand the challenge you'll stop posting your
ramblings about vector cryptography on sci.crypt. No more "You did that,
but that was just because of a small mistake. The underlying principle
is intact" crap.
If you feel that your current algorithm is not up to task and only your
future algorithm is ok, then work on the future algorithm instead of
wasting everybody's time with your lengthy posts. It's not worth it to
write so much about something that you can not back with real evidence.
Until you have real evidence, just go work on it and leave sci.crypt alone.
--
Paulo Marques - www.grupopie.com
"And folks, let's be honest. Sturgeon was an optimist.
Way more than 90% of code is crap."
Al Viro
Gene and Debbie Styer
Oct 24, 2011, 4:06:22 PM10/24/11
to
On Oct 23, 12:52 am, adacrypt <austin.oby...@hotmail.com> wrote:
>
> People is plural - there is only one person being jeopardised, that's
> Alice - This fiasco has spawned more red herrings than enough
>
[text deleted].
>
> People is plural - there is only one person being jeopardised, that's
> Alice - This fiasco has spawned more red herrings than enough
>
>
> Because something appeared in wikipedia that doesnt make it a theorem.
>
[text deleted]
> Because something appeared in wikipedia that doesnt make it a theorem.
>
You may not have read my post carefully, but I was talking about
Wikileaks, not Wikipedia - Those are two separate entities.
Trying to make my point again:
Suppose Alice, Bob, Carol, and David are communicating with each other
using your system. At some point along the way, the others find out
the Alice is a traitor, and that she has revealed everything she knows
to the enemy. My question is, can the enemy use that information to
break the messages between Bob, Carol and David?
Gordon Burditt
Oct 24, 2011, 11:46:54 PM10/24/11
to
> I think it should be conducted all over again in a few weeks when I
> can table a new database that PM will not be handed on a plate this
> time to exploit as he did (not).
This is harsh condemnation indeed of a cipher where it is intended
> can table a new database that PM will not be handed on a plate this
> time to exploit as he did (not).
that every pair of people have their own database and each person
is expected to have as many databases as they do Facebook friends.
Setting up a new database should be easy and quick.
If it takes "a few weeks" to set up a new database, it will never
be usable as practical cryptography (even if the alternative is
posting all e-commerce transactions to alt.ecommerce.transactions
in plaintext to the world, *with a Chinese translation* to make
it easier for them to read it).
"Over 10 minutes" is probably a killer for an e-commerce site that
has thousands of new customers a day, and if this cipher is used
for this purpose, would require setting up thousands of databases
per day.
0 new messages