Re: I know this won't work, I just want to know -why-.




dalek_no2@xxxxxxxxxxx wrote:
> Another crackpot thought pattern from me. I occasionally poke my head
> into this group and rattle off some inane, but unusual thought
> regarding compression methods. I understand the basic concepts of
> compression, and I get the whole "you can't get something for nothing"
> line of reasoning, but I still like thinking towards the more exotic
> forms of data compression.
>
> Anyway, recently, I found myself reasoning thusly:
>
> It is possible, nay trivial, to create redundancy data for any file
> which will detect and correct any error up to the size of the
> redundancy data. Simple, yes?
>
> Why, therefore, would it be impossible to take a file, create just over
> 50% redundancy for it, delete the original file, create a -completely
> random- file of exactly the same size, and apply the recovery data to
> the random file? Statistically, only about half of the bits should be
> wrong. Also, you're not compressing N-bytes into fewer than N-bytes,
> since the random file is N-bytes, and the recovery data is 1/2 * N
> bytes.
>
> I suspect that my misunderstanding here lies in my understanding of
> error detection and recovery, but I just wanted to make sure.

Instead of speculating over if it will work or not, why not just code
it and see if it works :-)

Its not that hard to code this thing, and you will then know for sure.

Sachin Garg [India]
http://www.sachingarg.com

.



Relevant Pages

  • I know this wont work, I just want to know -why-.
    ... compression, and I get the whole "you can't get something for nothing" ... It is possible, nay trivial, to create redundancy data for any file ... Also, you're not compressing N-bytes into fewer than N-bytes, ... since the random file is N-bytes, and the recovery data is 1/2 * N ...
    (comp.compression)
  • Re: [opensuse] Backup Process
    ... but I don't know of a tool that can record removable media compressed with recovery data designed to bypass the common types of media errors. ... I'm not talking of state of the art maximum compression: ... No on the fly compression with error recovery data, and on the fly recovery, tailored for the backup media. ...
    (SuSE)
  • Re: [opensuse] Backup Process
    ... Did you ever use the old PCBackup for Dos, from PC Tools? ... The technology is out there somewhere, but I don't know of a tool that can record removable media compressed with recovery data designed to bypass the common types of media errors. ... I'm not talking of state of the art maximum compression: ...
    (SuSE)
  • Re: I know this wont work, I just want to know -why-.
    ... Send redundancy data. ... Create random file of proper size. ... Apply recovery data to random file. ... If error detection and correction can work the way I currently ...
    (comp.compression)
  • Re: [opensuse] Backup Process
    ... rsync could ... Compression is good until you get an erroneous bit. ... They are compressed, yes, but they also contain recovery data to ... The other items mentioned are all either command-line or scripted. ...
    (SuSE)