101

Re: Dispersion of length of archive?

There are variants  without electric power burning: the community selected responsible, they of the  signed everyone, reward for it drank between itself. It is ready. Any useless searches.
There are here points at issue in a choice , but, think, as a result this circuit gets accustomed. Because the electricity expenditure on  already passed all reasonable limits. Recently in news wrote that in China  want to restrict an electrical supply since the infrastructure does not cope with such loadings.
Otherwise will be as above wrote: "rates all grow" since on an infrastructure money it will be necessary.

102

Re: Dispersion of length of archive?

Basil A. Sidorov wrote:

Personally for me both occupations

are senseless
It is repetition it (was already told), therefore and me it is necessary to repeat:  it is senseless in itself , but in inside  (at all , namely !) it fulfills quite defined (so comprehended) function of autoregulation of frequency of origin of new units.
Precisely also, for example, are senseless ( in itself ) efforts on expansion on a factor of large integer numbers. But in the project of search of prime numbers they are absolutely necessary (I do not consider, this project) is how much comprehended.

Basil A. Sidorov wrote:

for   there is an accurate algorithm and who has more than computing powers at equal architecture will win. Including the architecture specially developed under a specific target.

- process probability: "wins" - on each step - the one to whom carries. But chances it is more at at whom more powerful equipment. Actually, in the "modeling" task of search of prime numbers too so: "wins" (finds each new number) not mandatory at whom the most powerful computer.
(Yes, by the way, it is possible and to say explicitly that I exist projects (and even, apparently, implementations) , in which  not , namely (new) prime numbers!)

Basil A. Sidorov wrote:

That  compression levels each time is required or to develop algorithm of compression under specific sequence that looks a little utopian or to select a certain mythical parameter.

- "or to select... Parameter". At last we started to distinguish goal seek and the algorithm invention.
And if in process " compression levels" are invented new, more effective algorithms of compression, - so the same simply feast of spirit, direct favor for grateful Mankind!

103

Re: Dispersion of length of archive?

FXS wrote:

- "or to select... Parameter". At last we started to distinguish goal seek and the algorithm invention.

you, a pancake,  , parameters of normal algorithms saw?
These parameters - countable number. Even it is less, than it is possible to count a human body (720).
Really you think what to sort out these parameters - heavy work? Yes archivers already use parallel compression with different variants for a choice "the best on the given unit".
Or, maybe, you think, what manage to provide parameter of capacity of the order 2^64 and with stochastic dispersion of a level of compression on this set?
Remove, at last, pink glasses and go down from clouds on the earth.

104

Re: Dispersion of length of archive?

Basil A. Sidorov;
Well I invented (directly on a knee) an example (not working, for certain, but an example) algorithm of compression which uses search seed ().
And you continue about " normal algorithms"! In my setting of the task it is necessary not " normal ", and special  algorithm.
(What favor from   -  it  - happens out of the task " to a compression level" is a separate talk.)

105

Re: Dispersion of length of archive?

FXS wrote:

well I invented (directly on a knee) an example ( not working , for certain, but an example)

I underlined a keyword. Idle ideas at me - though a pond .

106

Re: Dispersion of length of archive?

Basil A. Sidorov;
So, I here play at giveaway, and you give in on this game in .
But generally, why generated (on the basis of ) "dictionary" (which, thanking him , it is not necessary to push entirely in an archive file) does not offer advantage before storable (in an archive file) dictionaries, - anybody yet did not explain to me.

107

Re: Dispersion of length of archive?

FXS wrote:

But generally why generated (on the basis of ) "dictionary" (which, thanking him , it is not necessary to push entirely in an archive file) does not offer advantage before storable (in an archive file) dictionaries, - anybody yet did not explain to me.

Lemovsky "Fairy tales of robots" esteem - there practically you of a case is considered. Generation of the information from white noise.

108

Re: Dispersion of length of archive?

Basil A. Sidorov;
Yes-yes, you, probably, and enciphering of messages by means of  will deny. And only: we generate on the basis of next (written down on a piece of paper both at Stierlitz, and at Centre) seed casual bit sequence and it is summable it () with the sent message.

109

Re: Dispersion of length of archive?

FXS wrote:

enciphering of messages by means of  will deny. And only: we generate on the basis of next (written down on a piece of paper both at Stierlitz, and at Centre) seed casual bit sequence and it is summable it () with the sent message.

the Gamma does not generate the information, and hides it in white noise. To understand it it is not required even higher mathematics.
Here when learn to cipher one gamma, and to decrypt another - then will make mischief.

110

Re: Dispersion of length of archive?

Basil A. Sidorov wrote:

the Gamma does not generate the information

and the dictionary generates the information?

111

Re: Dispersion of length of archive?

... It, actually, too ciphers the information. The question exactly in, whether will be possible to make the ciphered message more shortly (bit count), than initial.

112

Re: Dispersion of length of archive?

Ivan FXS, now came across, you tried it? (I - was not present)

113

Re: Dispersion of length of archive?

exp98, no.

114

Re: Dispersion of length of archive?

FXS wrote:

But generally why generated (on the basis of ) "dictionary" (which, thanking him , it is not necessary to push entirely in an archive file) does not offer advantage before storable (in an archive file) dictionaries, - anybody yet did not explain to me.

When the next inventor comes to academy of Sciences with the next idea of a perpetuum mobile, anybody explains nothing to it, simply show the door.

115

Re: Dispersion of length of archive?

S.G., where in it academies of Sciences a door?

116

Re: Dispersion of length of archive?

FXS wrote:

S.G., where in it academies of Sciences a door?

it is that through which enter.
It through which quit smile

117

Re: Dispersion of length of archive?

S.G. wrote:

in academy of Sciences with the next idea of a perpetuum mobile

And nobody remembers "final verdicts" French Ak. ?
Stones cannot fall from the sky, for it not  the earth.
Devices more heavily air cannot fly.

118

Re: Dispersion of length of archive?

exp98, at all a problem: someone unfamiliar (we will not point a finger... On S.G.) insulted. Well, insulted and insulted.

119

Re: Dispersion of length of archive?

exp98 wrote:

it is passed...
And nobody remembers "final verdicts" French Ak. ?
Stones cannot fall from the sky, for it not  the earth.
Devices more heavily air cannot fly.

Well naive they and improvident. It is necessary so: "the perpetuum mobile Project will be considered after its operating prototype within two hundred years smoothly supplies with an electricity the Academy building".

120

Re: Dispersion of length of archive?

I know as how to transfer any oblate data of enough big length even more shortly, than their length. Basically it not a joke, it from area .
What horse-radish tsp-packages empty run? We encode a part of the packets given by numbers. Over variants it is possible to think. For example one inverse of numbers (i+1, i) == 1, 0 - to absence of inverse.
For  reliabilities it is possible to use noise-resistant coding. For example among the codes . 1 == odd swap of several numbers, 0 - even. The encodable piece of the data on other end is joined to "legal". Certainly such method is powerless against "order by desc" the provider.
However, if an initial part to segment, transferred  it is possible to encode for example bits sessions. Or still as-thread.
The main thing, to the provider not to speak.

121

Re: Dispersion of length of archive?

We speculate with term . But if it () to consider in a composition of Markov model of the automatic machine
Which generates syllables, letters, words and word combinations that Ivan's idea quite to itself the worker.
But the main complexity we transfer to other function which also it is necessary is compressed to transfer on a communication channel and it   than .

122

Re: Dispersion of length of archive?

mayton;
Generally speaking, if we on a Centaur Alpha send the file compressed by the archiver, and we want that local  it unpacked, we should together with a file of archive and (de-) the archiver to send, and... An operating system, and it is a lot of that else, even I will not undertake to enumerate.

123

Re: Dispersion of length of archive?

mayton wrote:

But if it () to consider in a composition of Markov model of the automatic machine
Which generates syllables, letters, words and word combinations that Ivan's idea quite to itself the worker.

The unoperated idea since as a matter of fact it is the same dictionary as at normal compression, but is the dictionary with huge redundancy, the length of the identifier of each record of the dictionary is as consequence superfluous.
Compression after all follows the account of that that the most frequent records have the shortest identifier.

124

Re: Dispersion of length of archive?

Dima T wrote:

Compression after all follows the account of that that the most frequent records have the shortest identifier.

-- It is true, but attributing various to "frequent records" identifiers, shorter, than these records could promote rather far, if not necessity to create for each such changeover "dictionary entry". Present that in our input flow in the end the end found out two entrances of a line in length of 1000 bits, and possibility to create short identifiers is settled up to the identifier in length of 998 bits... Anything, we designate (and we replace) this 1000-bit line the 999-bit identifier - we compress, apparently, an input line on 2 bits! However is not present, "the expenditure of bits" on instructions of correspondence of a 1000-bit line to the 999-bit identifier will be in an any way more than this "saving" of 2 bits.
That is dictionary creation stops when "" expenditures () on dictionary addition become equal in bits of saving (bits) from application of new "dictionary entries".
And here this last condition on another will "be absolutely calculated" at usage of "the generated dictionary" (than at usage of the dictionary inscribed in archive). Thus that in itself "the generated dictionary", of course, noticeably will be less effective (on saving of bits at application of "dictionary entries"), than the "free" dictionary inscribed in archive.

125

Re: Dispersion of length of archive?

FXS wrote:

it is passed...
-- It is true

more precisely, is not true: compression follows the account of that records they have shorter identifier, than. And here " good " (" the best ") compression becomes "because the most frequent records have the shortest identifier" as you wrote.