Re: This Improvisation Will Buy Time Alright But….

On Thu, 12 Jan 2012 21:53:04 +1000, adacrypt <austin.obyrne@xxxxxxxxxxx> wrote:

On Jan 12, 11:30 am, "David Eather" <eat...@xxxxxxxxxx> wrote:
On Thu, 12 Jan 2012 20:05:24 +1000, adacrypt <austin.oby...@xxxxxxxxxxx>

> On Jan 11, 6:19 pm, Mark Murray <w.h.o...@xxxxxxxxxxx> wrote:
>> On 11/01/2012 14:39, adacrypt wrote:

>> > I find it hard to believe that there can be so much complacency about
>> > the fact that all modern cryptography is still computer dependent and
>> > nobody seems to worry about it.

>> Please explain this. What is wrong with computer-based crypto that
>> makes this "computer dependence" problematical?

>> Are you saying that some form of non-computer crypto should replace it?
>> Of so, how do you imagine people using it? Pencil-and-paper? What would
>> they be encrypting?

>> > This is something that cannot be engineered out of the industry by
>> > progressive computer science but instead boils down to nothing more
>> > basic than human intervention by the entities of secure communications
>> > systems. They can collude via the arrays of their mutual database to
>> > withhold keys and release them ad hoc by calling them sequentially at
>> > the precise moment that they are needed by the decryption program.

>> If this is addressing the "computer dependance" problem, then I don't
>> see how.

>> > Nothing very clever about this but the sheer transparency of the
>> > universal number system must be overcome beforehand. This means not
>> > using the ordinary universal number lines that may have arbitrary
>> > direction and periodicity and instead customising unique number lines
>> > for each and every encryption of a single plaintext. Number lines
>> > having unique direction and unique periodicity enable this to happen
>> > are quite easily defined. It only means shifting he math goalposts a
>> > little.

>> What about nonlinear functions?

>> > After that its Mr Kerchoff’s admonition once more “ the cipher must
>> > depend only on keeping the key-set safe” which is truly the case that
>> > I am promoting with my crypto.

>> Thats pretty much what most folks try to do when designing ciphers..

>> > The algorithms are up and running with a good performance rate.

>> Disagree. Your example code is VERY slow.

>> M
>> --
>> Mark "No Nickname" Murray
>> Notable nebbish, extreme generalist.

> <Please explain this. What is wrong with computer-based crypto that
> <makes this "computer dependence" problematical?

> The crypto in question is prone to brute force by any computer that
> can do it in an acceptably short time - at the present time computers
> with that kind of power are are not to hand (yet) but they are on the
> way. The solution to this in my view is new cipher designs that are
> intellectually above computer capacity no matter how great i.e.
> independent of computer power.

> There may be a price to be paid for this in performance speed but that
> may have to be accepted and indeed the same increases in computer
> power that are posing a threat may alleviate this downside.

>> What about nonlinear functions?

> Nothing is safe from numerical attack on the current number system -
> this is what I mean by the vunerability of the universal number
> system because of its sheer traqnsparency in its straight-off-the
> cuff form - it needs to be customised specially for cryptography in
> order to thwart numerical attack by any of the several current methods
> that are available to cryptanalysts. I think everybody will agree
> with this.

>> Disagree. Your example code is VERY slow.
> In my view nothing will ever compete witht the encryption/decryption
> rate of AES. By comparison my algorithm is bound to be slow as indeed
> will also be the case for any other algorithm that comes up in the
> future. AES has it made in that respect but it is not wise to make
> comparisons because once computer power increases sufficient to knock
> out complexity-driven crypto like AES then that's the end of for all
> others also that are of similar design (albeit as yet unknown).

> Slowness can be overcome by the same technology that threatens the
> status quo with increasing power.

> I believe there is some concern at the higher levels like NSA at the
> possibilities being outlined here but nobody is saying anything.

> Even as a curio I think my stuff is very worthwhile - it is not meant
> to be beautiful - just strong crypto.

> 30000 characters per second on a very ordinary home computer bodes
> well for what a large purposely chosen computer might do in a real
> hands-on crypto scene.

> - adacrypt

Mark and Adacrypt,
Oops. I haven't done you any favors. I will take my intrusion into a different thread. If at all possible please get back to Q and A without rancor. I can do my bitching elsewhere.

Note to Mark,
Adacrypt's software is also being slowed down significantly by windows. It would perform somewhat better if it was done in C but if that would be quick enough to class as 'fast' I don't know.

Two things on the plus side: I have not seen an emails to a hijack address despite Adacrypt being quit active in the last few days. Perhaps the problem if fixed... but caution dictates waiting a bit longer to be sure.. Also a note to Adacrypt - it is not opening the email that is dangerous - it is clicking on the link provided that will cause the grief. So you can read or ignore however you like - just don't click.

Relevant Pages