1

Topic: About one method data compressions

There is data  - bit sequence of length N, - which are required to be compressed, and standard .
1. We find bit sequence of the minimum length M which never meets in D.Nazovyom its "flag" ().
2. In a cycle we sort out values seed (s), used for initialization :  (s).
3. By means of  (s) it is generated casual sequence (joint venture) of the same length N.
4. We compare   and the joint venture. All coinciding - in  and the joint venture - bit sequences, length of more M is replaced (in !) on F.Pri's flag it we track, that "on joints" an interposed flag and adjacent initial bits  did not arise "stray flags".
5. As a result of changeovers bit sequence  ' N ' (less N turns out lengths!).
6. We continue search (a cycle in item 2) until we find value seed, giving a level of compression satisfying us To = N/N'
7. The archive file represents structure from three elements: , s,  '.
______________________________
PS. As a variant, as flag  it is possible to use generally any bit sequence, it is desirable rarely meeting in  as "natural" entrances  in  will need to be marked means, for example, doubling: -> . (And then it will be necessary to track, that additional  did not arise in the course of compression.)

2

Re: About one method data compressions

... By the way, (1, s1), it is possible further to try to compress a file compressed with application of one pair of parameters with what that another 2, selecting any other value s2. And so yet does not bother.

3

Re: About one method data compressions

Ivan FXS;
If  qualitative, i.e. allocation of bits equiprobablly, estimate an amount of cycles for achievement of the necessary level of compression. The problem does not look very much difficult and, I think, the answer is not pleasant to you.

4

Re: About one method data compressions

The task can be reduced to  replacing  with sequence of integer numbers.

5

Re: About one method data compressions

It is normal algorithm LZW, only with errors and accidentally-generuemym dictionary:fail:

6

Re: About one method data compressions

scf;
In secret, this topic is carried out in an independent subject from a subject http://www.sql.ru/forum/1281135/dispersiya-dliny-arhiva - where strong was just considered compression with an expenditure big (in a limit - not restricted) amounts of resources (is desirable).

7

Re: About one method data compressions

Siemargl;
Surprise: the dictionary does not need to be pushed in an archive file!

8

Re: About one method data compressions

Just the same the first  as at Siemargl - type , but only the dictionary is replaced by algorithm. I.e. as though, being repelled from ideas , we build something of a class "compression by means of algorithm". In this connection (and considering roots of this subject) there is a judgement that 1 of 2: or it is necessary to be able to recover a source code, or reproducibility of coding of the given D.Vo  I is necessary so I understand the term pow. In case  moreover standard I doubt reproducibility. In my judgement, at same  the joint ventures will be different (and laziness to check).

9

Re: About one method data compressions

exp98;
You, probably, suppose that if repeatedly to cause  with the same argument (seed th) it returns other result...

10

Re: About one method data compressions

Ivan FXS, a code sample will be? It is desirable with statistics.

11

Re: About one method data compressions

Dima T, in sense "code sample"?

12

Re: About one method data compressions

Well ... Friday proceeds. Let's argue.

FXS wrote:

There is data  - bit sequence of length N, - which are required to be compressed, and standard .

The thesis about that that is certain "sequence in vacuum" me does not arrange. It - crafty.
I had questions. On what this bit sequence is similar?
1. The English text. Or the arbitrary text.
2. A flow of casual bits like/dev/urandom
3. An aforesaid compound.
4. Still any variant
[quote =] 1. We find bit sequence of the minimum length M, which never meets in D.Nazovyom its "flag" (.

I suggested to name it "escape" (' \')
[quote =] 2. In a cycle we sort out values seed (s), used for initialization :  (s).

It is necessary for us though will be formal to prove that length s in bits less than length random ().
For example for built in OS of linear congruent , the argument and function will be
On 32 bits. And function - an essence bijection.
If we enter that into task  of higher or floating digit capacity it is necessary
To justify that it is . Or it NOT-GPSCH. Or not bijection. Not the bijection - means
It is not covered . Etc.
[quote =] 3. By means of  (s) it is generated casual sequence (joint venture) of the same length N.
4. We compare   and the joint venture. All coinciding - in  and the joint venture - bit sequences, length of more M is replaced (in !) on F.Pri's flag it we track, that "on joints" an interposed flag and adjacent initial bits  did not arise "stray flags".
5. As a result of changeovers bit sequence  ' N ' (less N turns out lengths!).

The last point - a feeble place in reasonings.
It is necessary to substantiate reasonings somehow.  the next coder Grandmother's turns out.