On Tuesday, May 21, 2013 11:45:54 AM UTC+1, Austin Obyrne wrote:
> On Tuesday, May 21, 2013 9:22:37 AM UTC+1,
robert...@yahoo.com wrote: > On Mon, 20 May 2013 06:36:49 -0700 (PDT), Daniel Kruyt <
daniel...@gmail.com> wrote: >Austin, rounding error is not to be over looked. Saying that there is an infinite number of vectors that can be represented is incorrect when using finite data sets, no? > >I plan on doing a short cryptanalysis of your cipher, I'll post it on this thread in a little while if you'll agree to it. > >A "break" shall be defined as something requiring less time/memory than a brute-force, are you okay with that? Why bother improving on brute force? Last I heard the secret part of the key was a single five (decimal) digit number. At other times it's been two four digit numbers, two integers with a product less than 14250, and a few other things. Hi, It’s impossible for me to comment on isolated fragments of the whole like this one from you without muddying the waters even more. A lot of readers are still thinking and arguing with a scalar mentality. I can assure you that this crypto has hugely geometric connotations that are implemented by a collection of random keys that must all be known and used in what is essentially vector arithmetic. Unless a reader is totally au fait with vector methods and plane geometry (this is what provides the essential substitution methodoly) there is no hope of me informing him by posts here in sci crypt and even then the academic people (outside of this group) that I have presented this stuff to are having difficulty getting their heads around it. It needs a lot of one-to-one. I am not being evasive when I say that it is impossible for me to convey a proper understanding by way of posts here – hence these supplements - but even that must stop now because a lot of people are simply not into vectors sufficiently to absorb what comes next and are still arguing along scalar lines which they understand but wrongly think will suffice. Just out of interest - Are *you* totally au fait with vector methods applied to plane geometry?? – would like to know. - adacrypt
Supplement - 4
This supplement is included here (as well as elsewhere) because it says a lot.
Number-theoretic cryptography took off in earnest when computers came into vogue in the 1970’s.
People thoughtlessly (why shouldn’t they? one may well ask) used the traditional data systems as the selection domain for raw data in cryptography because after all, there was no need to question its suitability. The fact that it had worked for them in billions of other cases since the year dot was probably so entrenched that it just didn’t bear questioning.
However, with hindsight, it can be seen today that this was a mistake for several reasons.
The raw data used for encryption transformations in cryptography should ideally come with a lot of *manageable disorder as an intrinsic property (call this innate ‘starting’ entropy if you like) such that a design cryptographer has only to invent minimal extra 'make-up' entropy to add to it by means of an encryption algorithm that will produce an unbreakable cipher.
That is not the case with the popular data sources that have been used however as the selection domains by cryptographers, namely, our number system , our natural alphabets and even the ASCII code that was created specially with computers in mind.
This data is so perfectly ordered that it has *zero entropy (as disorder) which means the cryptographer has a mountain to climb in providing the necessary extra entropy that will barely make it into ‘strong’ cryptography with little chance of making it totally unbreakable cryptography.
Nobody seems to have noticed this and cryptographers have instinctively filled the void with studiously contrived compensating complexity. Their entire modus operandi has been spent on thinking up (complexity-intensive) difficult-to-invert algorithms. These algorithms are horrendously difficult to invent and so far there has not been even one successful cipher that is totally immune to inversion by brute force (even AES isn’t completely safe). This is disregarding the impractical OTP.
Going down this wrong road by cryptographers in the 70's can be easily understood given that our number system has served mathematics perfectly for thousands of years even to the extent of a moon landing. Why should anybody stop to think that it might not be suitable for some purpose like cryptography and require changing – but that is what is required now – cryptographers must stop using these highly ordered data systems, to wit, the integer set, natural alphabets and information codes like ASCII are all taboo to cryptography in their unscrambled state.
Not mentioned so far and hugely important also is the fact that cryptanalysts have been given the same access to these same data systems that were being mutually used by both cryptographers and cryptanalysats all along, which is tantamount to giving the latter a head start in their nefarious trade of cryptanalysis. This is key information available to adversaries that is there simply for the picking up, an absolute and total folly by cryptographers is the only way to describe it but what was the alternative one may well ask.
Summarising.
The use of integers and alphanumeric data taken directly (i.e. without reversible substitution into any other form) from the highly ordered traditional sources as the selection domain by cryptographers for encryption transformations must stop forthwith because clearly, it is unsuitable to cryptography. Being used as the selection domain for the raw data that will be used in encryption transformations is tantamount to directly handing ‘giveaway’ information on a plate to adversaries.
It is amazing that this has gone on so long unnoticed.
The huge list of papers, past, present and future still being read in the establishment to day makes one wonder how much longer this fruitless search , i.e. for the essential compensating complexity that is required to mask the natural transparency of numbers and alphanumeric data, will go on when the only and proper solution is to stop using numbers from ordered number lines altogether.
Vector cryptography puts all of this right and guarantees unbreakable cryptography. No amount of computer power can break it.
There is no triumphalism or egoism on my part in what I say here – this is a mathematical and scientific fact – a part of the Universe.
I think most readers will see the sense of what I am saying here.
- adacrypt