1

Topic: The chetvergovyj registrar

Greetings friends.
Dima, the Owl, Ilya, Ivan-FHC, -128, Ehp98, Barlon, to Zjama and other forgive whom forgot.
Throughout topics:
I set a subject.
Experiments with   , Numeration systems, codecs and archivers . Breadboard models. Proo-of-conept. Working
Application. Here we will run in our ideas in the form of implementation.
Some points.
1) Base type . For the formal proof of our hypotheses, bits are not necessary to us.
Lines suffices us. std:: string, String, e.t.c. An example:

string s = "00101010010101010";

Any can easily do on them the Boolean
Operations also that the main thing - concatenation and a scrap. Storages suffices us. The overhead projector makes
The order of 8-16 times well and .
And bits go then. When the hypothesis flies up.
2) Visualization. . I am the big fan of a scientific drawing and visualization of ideas. Frequently
Happens that given presented by a picture carry more than valuable convergence than a heap
Clever verbal theses. Therefore black-and-white (bi-tonal) the picture gives more
Visualization for entropy and dominance of any properties in a binary file.
For those to whom laziness  graphic codecs - take format PPM. It is a matrix  in the text.
Or draw in the browser on . Examples - are trivial.
3) Language. any of the popular. It is possible even JavaScript, the main thing that we could  somehow .
4) Sortsy. . As always I lift .  here.
https://sourceforge.net/p/psevdo-random … HEAD/tree/
5) Indirect libraries . It is desirable that were  to understand idea.
Go-go !
Theorists - smoke aside.

2

Re: The chetvergovyj registrar

Throughout topics

3

Re: The chetvergovyj registrar

The idea of the archiver with  not the worker, will not be compression. Wrote here and here
If it is short: the length of substring of the initial data which coincides with sequence from  most likely will be less or is equal to length  demanded for initialization . I.e. we change M bit on M bit.
Variants of longer coincidence are possible, but everyone . coincidence reduces twice probability that the following bit coincides.
PS Own exotic ideas on this point in question at me are not present.

4

Re: The chetvergovyj registrar

mayton wrote:

archivers
2) Visualization.

Forgive for a few not in a subject.
About last - everyone is engaged in what to it )
Archiving, item 2):
Suddenly recalled for a long time  variants. To the beginning 90 (640 , 10 or  only )  programs did a disk with an exterior help-file (it and a resource file). For the sake of saving of a place and-or protection against independent addition the elementary variant of compression by repeating parts of words was used. It was simply visible, as in process of file paging in the beginning there were bole-mene  nearly whole words, then turning in  pieces of letters. Here and all visualization.
Implementation primitive, compression strong, however completely not maximum. I hope with idea all clearly, source codes, excuse, far.
-----------
in an ex-ale:  I will find a picture...
Section " the MSEC Office "
a subject generator-sluchaynyh-chisel-v-excel
The message
On  simply fragment of the theorem that m. . Equally arranged . has norm.raspr. All life trusted the theory, recently took, and looked at a slice.

5

Re: The chetvergovyj registrar

To a question on applications .
Still the old recalled. Was, not me, it suggested to model model with a view of a location a cross-country terrain a variant iterated 2-dimensional  models with parameters, on the basis of . There interesting pictures , various vulgar regions of type of ovals, hoses, amoebas etc., each region with the . Characteristics. In any case arose "treug-to Serpinsky". Well and as a subtask was to improve  .
It is necessary to tell, as well as in a case with natural fractals, under the given picture the model should be selected (read search).
For for the sake of model and source codes it is necessary long and to rummage physically if it was saved.

6

Re: The chetvergovyj registrar

Dima T wrote:

Idea of the archiver with  not the worker, will not be compression.

Ugu...
Very much for a long time, when archivers were feeble, and I absolutely green, at me originated similar idea. But she was killed by simple experiment. Took I a file (the big text),  its archiver ( RAR) with the maximum adjustments, then XOR it  (it is remembered, made already 100 copies, everyone is processed with the seed) and tried to press out the same archiver with the same adjustments... Looked at result - and  to produce bad ideas.

7

Re: The chetvergovyj registrar

Akina;
And remained in your storage something more specific about result of that experiment: type - any of 100 copies was not compressed on one bit? Whether we after all consider a question "it is possible to compress a casual line", whether instead of "it is possible to compress it on tens percent"...

8

Re: The chetvergovyj registrar

FXS wrote:

and remained in your storage something more specific about result of that experiment: type - any of 100 copies was not compressed on one bit?

The source file - the formatted ASCII-text in the size of the order of 1 Mb, in an oblate type of the order of 80 Kb.
Coding environment (and ) - Borland Turbo BASIC 1.0.
The maximum additional compression - the order of 1.6 % (on one file).
Amount of the files, which size decreased - the order of 70 %.
Average reduction of file size, on all array - the order of 0,2 %.
Average reduction of file size, only on in addition compressed - the order of 0,3 %.

9

Re: The chetvergovyj registrar

Dima T wrote:

Idea of the archiver with  not the worker, will not be compression. Wrote here and here
If it is short: the length of substring of the initial data which coincides with sequence from  most likely will be less or is equal to length  demanded for initialization . I.e. we change M bit on M bit.
Variants of longer coincidence are possible, but everyone . coincidence reduces twice probability that the following bit coincides.
PS Own exotic ideas on this point in question at me are not present.

+1

exp98 wrote:

To a question on applications .
Still the old recalled. Was, not me, it suggested to model model with a view of a location a cross-country terrain a variant iterated 2-dimensional  models with parameters, on the basis of . There interesting pictures , various vulgar regions of type of ovals, hoses, amoebas etc., each region with the . Characteristics. In any case arose "treug-to Serpinsky". Well and as a subtask was to improve  .
It is necessary to tell, as well as in a case with natural fractals , under the given picture the model should be selected (read search).
For for the sake of model and source codes it is necessary long and to rummage physically if it was saved.

+1
Excuse mayton, but the task-idea to write on this subject the code - attempt
To deliver a cart ahead of a horse.
Fundamentally  it is necessary to approach, look on the other hand
On the general-purpose algorithms of compression
Not losing an information essence, but reducing detailing without recovery possibility.
If very deeply to dig, a subject, transition point of quality of compression,
In quality  information
Is where that by this technique .
Also 2 passes on all information volume will be necessary a minimum,
What to catch fractal dependences.

10

Re: The chetvergovyj registrar

If there is a question  lossless compressions;
That is other area of mathematics - multidimentional spaces.
Here it is even more stolen to us.
Search for a calculation floor-mat on algorithm ADSL and other algorithms of laying down
Full-spheres in multidimentional spaces.
To remain sitting simultaneously on 2 chairs if it turns out,
That the code any  and
Practical interest will not represent.

11

Re: The chetvergovyj registrar

0k wrote:

If there is a question  lossless compressions;
That is other area of mathematics - multidimentional spaces.
Here it is even more stolen to us.
Search for a calculation floor-mat on algorithm ADSL and other algorithms of laying down
Full-spheres in multidimentional spaces.
To remain sitting simultaneously on 2 chairs if it turns out,
That the code any  and
Practical interest will not represent.

Dock, I think you for a long time already outgrew it a forum and it is impossible to surprise you with anything.
I think as to me with you to compete heavily. In you there is a tenacious mathematical mind.
I simply engineer and  the person. I have no anything against yours
Theses. But I am afraid that multidimentional spaces is yet my level.
The matter is that we here occupy the niche. Also we are engaged not in the mathematical
Discoveries and engineering of algorithms. Like designer Lego, we collect
From primitives something. It not a science. This adaptation. We ""
For itself splinters of knowledge just for fun. An evolutionary way. Selection. Iterated
The approach. Best practices. Scrum/Agile. This all different titles of this process.
We do not claim even for the rigorous proof of fidelity of our way.
Here there are no formulas. The congestion an expert is simple.
And I as well as all others  too am here for fun. Besides it
I solve parallely 3-4 more others tasks which I did not sound but they
Interleave with considered hashes-TAGS and sooner or later I will share them.
Here is how that in such here aspect.

12

Re: The chetvergovyj registrar

Dima T wrote:

Idea of the archiver with  not the worker, will not be compression. Wrote here and here
If it is short: the length of substring of the initial data which coincides with sequence from  most likely will be less or is equal to length  demanded for initialization . I.e. we change M bit on M bit.

I am not going "to break about a knee" the archiver with . It quite to itself
Can exist as "" for its comparing with others.
I set other task. Not to refute Ivan's theses and is faster
To give it new ideas and theses. Again. In practice.

13

Re: The chetvergovyj registrar

exp98 wrote:

It is necessary to tell, as well as in a case with natural fractals, under the given picture the model should be selected (read search). for the sake of model and source codes it is necessary long and to rummage physically if it was saved.

Screenshots were saved?

14

Re: The chetvergovyj registrar

0 wrote:

I as though that it is not enough sense to try to collect puzzles .
Even for the sake of .
General-purpose Algorithms of lossy compression has huge practical sense.
It is basic  the abstract and figurative thinking  intelligence.
Therefore I would move a task priority in this side.

Give in couple of weeks. I will lift a subject with lossy compression. I there have more ideas...

15

Re: The chetvergovyj registrar

- I speak. Everyone here finds the.

16

Re: The chetvergovyj registrar

mayton wrote:

it is passed...
I am not going "to break about a knee" the archiver with . It quite to itself
Can exist as "" for its comparing with others.
I set other task. Not to refute Ivan's theses and is faster
To give it new ideas and theses. Again-. In practice.

One fool sets so much questions.... ()
It is already answered

17

Re: The chetvergovyj registrar

mayton wrote:

Screenshots were saved?

it is necessary long and to rummage physically if it was saved . As soon as and... Well 95th, about something speaks? Though in the report 2-3 certainly were, only it was custom research) And so, it is visually remembered as is above described.  - the key moment was, basically and the it is possible to recreate.
And yes, image compression by "fractal" methods, is similar that advantages on Kszh before  not especially. In , during comparable time. Though the idea to steal up to "a fixed point of compressing display" in the full Banach space very much inspired at first. It is a hint.

18

Re: The chetvergovyj registrar

0, let Ivan itself does outputs with the registration or without our judgements. To study late never and to anybody. Besides its idea not in the archiver, and in application of this.

19

Re: The chetvergovyj registrar

Fibonacci tried on "paper" numeration system with scales. Funny. At it is
Possibility to recover the traffic after failure.  even if there was a damage
That is available possibility "to catch " traffics as network protocols and television.
Unfortunately there are initially difficulties with coding of negative numbers, zero and unit.
The negative demand additional bit, and the zero and unit - are indiscernible with the traffic and therefore
It is impossible to encode them.
If to go on cunning and to add +2 that the last artifact it is eliminated.

20

Re: The chetvergovyj registrar

mayton wrote:

Unfortunately there are initially difficulties with coding of negative numbers, zero and unit.

Set of integer numbers to homeomorphic set of the natural. Always it is possible to use.

21

Re: The chetvergovyj registrar

softwarer wrote:

the Set of integer numbers is homeomorphic to set of the natural. Always it is possible to use.

Plz, is accurater with standard .! Only identical capacity that speaks only about . Even "from-" demands "operation", without speaking about "gomeo-".
Well or operation then it is necessary to define (rather exotic method).

22

Re: The chetvergovyj registrar

exp98;
Operation is defined easily and simply at the first year. For example;
[CSV =;] Natural; Whole
1; 0
2; 1
3;-1
4; 2
5;-2
6; 3
...;...
And as a whole , is accepted.