Re: universal constants in cryptography?

Bryan Olson <fakeaddress@xxxxxxxxxxx> writes:

de_magnete@xxxxxxxx wrote:
As you known the physical ssciences have a set of what one might call
universal constants. Example "c" the speed of light, that turn up on
manay places. As far as I can understand these constants are useful as
a means of deciding whether a proposed theory or construction might be
correct. For example, A theory of electomagnetism might show that "c'"
is a consequence of the theory.

c is a units conversion "constant". Historically, people, not knowing
relativity, measured time and space with different units. It would be like
measuring horizontal distances in kilometres and vertical in inches. Since
they are really the same thing, one needs a conversion constant. In the
example that conversion constant is inches/km. In the case of time and
space it is m/sec. Similarly hbar converts metres to inverse momentum, or time
to inverse energy.
And G/c^2 converts mass to distance. Ie, they have no physical
significance, but are purely a correction factor to account for historical
accident.

Now there are some other constants. 1/137... for the charge^2 of the
electron,4x10^(-23) for the mass of an electron (in natural units), pi for
the ratio of circumference to diameter of a circle, etc. which are natural
constants.

What you are describing is dimensional analysis, where one guesses for a
physical process that its depenence is on mass, legths and time (in
theories where their separate specification makes sense) and sticks in
appropriate conversion factors to take care of the mismatch in dimensions,
and assumes that in "absolute" units, the answer is approximately 1.

No, no, no. Scientific theories are tested by comparing their
predictions against real-world observations. Arbitrary use of familiar
constants is a common feature of crank theories.

Now I wondered if cryptographic/information theory had such univeral
constants that can eb used in a similar manner. that is constants that
one might look derivable from a new proposal, in order to add weight
to the correctess of the new idea.

For example: might a deivation of an entropy of 2.58 have any
significance?
Does a probability of an event of 2/3 ~= 0.65 have any universal
significance?
Does a value of 0.38 when measuring the number of say the bits
matching between Item 1 and item two have any universal signifiicance?

Do you generally see what I mean by my question?

I think so, and it sounds like numerology.

I mean the derivation of such universal constants from a proposal
would be of great use in theory validation?

I do hope you see what I am getting at.

I just do not know if parameters equivalent to "c' are considered to
exist for information and cryptographic theory.

Certain constants have been exhausted beyond rational justification, and
getting rid of them has proven difficult.


--
--Bryan
.

Relevant Pages

  • Re: Boeing and metrcication question
    ... conversion) mass to force. ... clumsy derived units. ... head that are far abstracted from what you're actually measuring, ... I only have to look at the domain-dependant units of pound, pondal, pound ...
    (rec.crafts.metalworking)
  • Re: News: 10 Questions, and Answers, About Evolution.
    ... that itself is measuring the voltage on a strain-gauge or temperature ... of individual facts. ... that your senses very often lie to you, you cannot sense mass. ...
    (talk.origins)
  • Re: News: 10 Questions, and Answers, About Evolution.
    ... that itself is measuring the voltage on a strain-gauge or temperature ... of individual facts. ... that your senses very often lie to you, you cannot sense mass. ...
    (talk.origins)
  • Re: Why so artificial?
    ... The whole idea of measuring a single aspect, is to put all entities on ... Imagine that our topic of interest was "mass". ... Are humans more intelligent than chimps? ... Machines could easily surpass human performance in intelligence some day, ...
    (comp.ai.philosophy)
  • Re: Why so artificial?
    ... The whole idea of measuring a single aspect, is to put all entities on ... Imagine that our topic of interest was "mass". ...  Are humans more intelligent than chimps? ... Machines could easily surpass human performance in intelligence some day, ...
    (comp.ai.philosophy)