Prev: Prime Number Sums reveal both a 2D and 3D pattern when stepped by the golden ratio logarithm!
Next: Public/Public
From: WTShaw on 21 Apr 2010 02:32
Chaining ciphers can increase overall strength, or not. A
bootstrapping procedure would be used to increase strength, should not
duplicate functions with the other algorithm, and should be efficient
in data conversions.

Even a poorer algorithm might be bootstrapped disproving the prejudice
that anything can only be as strong as its weakest link. Primitives
by definition may be poor but useful in combination. As it might be
said, some algorithms need all the help they can get.

Consider that improper bootstrapping might work against you, making
analysis easier, or increasing the operational protocol to something
unmanageable.

In one area of work, bootstraps might be internally combined with
another algorithm, perhaps with a few compromises Having all i one
package is convenient, otherwise some intermediate copying and pasting
might be required.






From: Mok-Kong Shen on 22 Apr 2010 06:01
WTShaw wrote:
> Chaining ciphers can increase overall strength, or not. A
> bootstrapping procedure would be used to increase strength, should not
> duplicate functions with the other algorithm, and should be efficient
> in data conversions.
[snip]

Unless we have different definitions of bootstrapping, I wouldn't
consider bootstrapping, which in CS is used in compiler constructions,
to be a useful and appropriate concept to explain the design of ciphers.
For instance, I suggested previously long ago to use a master PRNG
to create a sizable number of PRNGs to be pseudo-randomly activated
to generate outputs. I personally wouldn't deem that as a bootstrapping
process but consider the starting phase with the master PRNG "simply"
as an initialization. Yes, indeed one "could" interpret that as
bootstrapping, if one "wants" to. But on grounds of the Occam's Razor
I would like to strongly limit the number of basic design principles
of ciphers, which I consider in one dimension to be the well-known
confusion and diffusion and in another dimesion to be variability
(dynamics) and indirectness.

M. K. Shen


From: WTShaw on 23 Apr 2010 04:25
On Apr 22, 5:01 am, Mok-Kong Shen <mok-kong.s...(a)t-online.de> wrote:
> WTShaw wrote:
> > Chaining ciphers can increase overall strength, or not.  A
> > bootstrapping procedure would be used to increase strength, should not
> > duplicate functions with the other algorithm, and should be efficient
> > in data conversions.
>
> [snip]
>
> Unless we have different definitions of bootstrapping, I wouldn't
> consider bootstrapping, which in CS is used in compiler constructions,
> to be a useful and appropriate concept to explain the design of ciphers.
> For instance, I suggested previously long ago to use a master PRNG
> to create a sizable number of PRNGs to be pseudo-randomly activated
> to generate outputs. I personally wouldn't deem that as a bootstrapping
> process but consider the starting phase with the master PRNG "simply"
> as an initialization. Yes, indeed one "could" interpret that as
> bootstrapping, if one "wants" to. But on grounds of the Occam's Razor
> I would like to strongly limit the number of basic design principles
> of ciphers, which I consider in one dimension to be the well-known
> confusion and diffusion and in another dimesion to be variability
> (dynamics) and indirectness.
>
> M. K. Shen

All the concerns are addressable. Time means picking those of
interest. Occam's Razor is as much a call to find simpler means as
anything else, not to avoid them. No, not all easy improvements are
widely known. For examples, using language itself as a generator is an
improvement over adding many steps to do something that can be rather
simple, avoiding a primitive because of prejudice can mean doing
things a harder way, and not following logic seems irrational in a
field that depends so much on it, all of it.

The algorithm family loosely referred to here as varieties of Pizza is
based a handful of good ideas, not many, that cut to the chase. Even
by itself it is good, very good. With the "test vectors," nearest
name to the information to be at hand, necessary diagnostics I used in
development tell the whole story. I'm just going to look at it
another day to see if I can cause any difficulties I have not seen.

One problem with casual use of Occam's razor is that knowing what is
obvious means understanding why such and such is so and so, and that
relies of comprehensive knowledge which many blindly lack.
From: WTShaw on 24 Apr 2010 12:12
On Apr 23, 4:10 am, Mok-Kong Shen <mok-kong.s...(a)t-online.de> wrote:
> WTShaw wrote:
>
> [snip]...  Occam's Razor is as much a call to find simpler means as
>
> > anything else, not to avoid them.
>
> I personally understand the essential value of Occam's Razor in this:
> If one could explain something in terms of simpler (to understand)
> terms/concepts and ways, then do that instead of explanations more
> complicated or even obscure.
>
> M. K. Shen

OK, look at the following bar graphs and mind that the classes in the
second one are smaller. If the same thresholds in the second were as
the first, the second one would be fairly much a straight horizontal
line.

It's plain text and after running it through Small Pizza 26:
-------
Typical Text: Total Characters = 1143
Order is eniatosrhldfmcugwbpyvkjqxz
Highest Percentage is 12.237762237762238
1/20th Highest Value is 0.6118881118881119
ee
ee
ee
ee
ee
ee
eenn
eennii
eenniiaattoo
eenniiaattoo
eenniiaattooss
eenniiaattoossrr
eenniiaattoossrr
eenniiaattoossrrhh
eenniiaattoossrrhhlldd
eenniiaattoossrrhhllddff
eenniiaattoossrrhhllddff
eenniiaattoossrrhhllddffmmccuuggwwbbpp
eenniiaattoossrrhhllddffmmccuuggwwbbppyy
eenniiaattoossrrhhllddffmmccuuggwwbbppyyvvkk
eenniiaattoossrrhhllddffmmccuuggwwbbppyyvvkkjjqqxxzz
------
Encrypted with Pizza: Total Characters = 1143
Order is vuicgfomszpqrdjxtyablnwkhe
Highest Percentage is 4.895104895104895
1/20th Highest Value is 0.24475524475524474

vvuuii
vvuuii
vvuuiiccgg
vvuuiiccggffoommsszz
vvuuiiccggffoommsszzppqqrrddjjxx
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnn
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkk
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhh
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee
vvuuiiccggffoommsszzppqqrrddjjxxttyyaabbllnnwwkkhhee

Cheers
From: WTShaw on 24 Apr 2010 12:13
On Apr 23, 4:24 am, Earl_Colby_Pottinger
<earlcolby.pottin...(a)sympatico.ca> wrote:
> On Apr 23, 4:10 am, Mok-Kong Shen <mok-kong.s...(a)t-online.de> wrote:
>
> > WTShaw wrote:
>
> > [snip]...  Occam's Razor is as much a call to find simpler means as
>
> > > anything else, not to avoid them.
>
> > I personally understand the essential value of Occam's Razor in this:
> > If one could explain something in terms of simpler (to understand)
> > terms/concepts and ways, then do that instead of explanations more
> > complicated or even obscure.
>
> > M. K. Shen
>
> And this why I never bother reading anything that wtshaw posts, If
> there is a simple way and a complex way to describe anything, he will
> post an even *MORE* complex way using terms that he will not define
> (or define in a way that tells you nothing about the terms).  Basic
> words, his posts are a waste of electrons.

Maybe you are a visual learner. Please go back to flash cards.
 |  Next  |  Last
Pages: 1 2 3
Prev: Prime Number Sums reveal both a 2D and 3D pattern when stepped by the golden ratio logarithm!
Next: Public/Public