Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.compression > #2103

Re: New Data storage method with severe compression capabilities.

Show all headers | View raw

Thomas Richter <thor@math.tu-berlin.de> wrote:

> what I wrote. *Data compression* is the science of detecting and 
> removing redundancy (and sometimes irrelevancy, for lossy compression) 
> from data. How that data is represented is irrelevant. If you use bits 
> or trits, or any other representation does not matter, the theory stays 
> intact, just the unit changes.
 
> Saying that you can "compress data by changing the representation" is as 
> good as saying "you can shrink distances by measuring them in miles 
> rather than kilometers". Yes, the numbers get smaller. No, you still 
> need to walk the same distance.

I think that isn't quite right, but it is so obvious that many ignore
it. If the data has a poor representation, then many compression
algorithms will find the redundancy in the representation.

Consider a file of random and uniformly distributed ASCII letters.
The redundancy is in the representation (ASCII) and not in the data
(the letters). 

LZW, for one, will find that in the same way it finds repeated strings,
and compress them down, not so far from the ideal. LZW does it without
knowing that it is compressing a poor representation of random data.

If you knew in advance that the data was uniform ASCII letters, you
might just encode them directly, but you don't always know that.

We often choose poor representation for convenience, though. Human
readable ASCII is nice, though often inefficient.

-- glen

Back to comp.compression | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-04 16:02 -0800
  Re: New Data storage method with severe compression capabilities. Noob <root@127.0.0.1> - 2013-11-05 10:03 +0100
    Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-05 04:44 -0800
  Re: New Data storage method with severe compression capabilities. Sebastian Garth <sebastiangarth@gmail.com> - 2013-11-07 23:28 -0800
    Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 01:00 -0800
  Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 04:56 -0800
  Re: New Data storage method with severe compression capabilities. Sebastian Garth <sebastiangarth@gmail.com> - 2013-11-08 09:50 -0800
    Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 10:41 -0800
      Re: New Data storage method with severe compression capabilities. Sebastian Garth <sebastiangarth@gmail.com> - 2013-11-08 10:57 -0800
        Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 13:25 -0800
          Re: New Data storage method with severe compression capabilities. pfraser <pete_fraser@comcast.net> - 2013-11-08 18:05 -0800
            Re: New Data storage method with severe compression capabilities. Sebastian Garth <sebastiangarth@gmail.com> - 2013-11-08 18:59 -0800
              Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 19:06 -0800
      Re: New Data storage method with severe compression capabilities. Thomas Richter <thor@math.tu-berlin.de> - 2013-11-08 22:28 +0100
        Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-08 16:40 -0800
          Re: New Data storage method with severe compression capabilities. Thomas Richter <thor@math.tu-berlin.de> - 2013-11-09 17:31 +0100
            Re: New Data storage method with severe compression capabilities. glen herrmannsfeldt <gah@ugcs.caltech.edu> - 2013-11-09 19:02 +0000
  Re: New Data storage method with severe compression capabilities. Fibonacci Code <anglikai@gmail.com> - 2013-11-11 05:48 -0800
  Re: New Data storage method with severe compression capabilities. Fibonacci Code <anglikai@gmail.com> - 2013-11-11 06:03 -0800
    Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-11 06:13 -0800
      Re: New Data storage method with severe compression capabilities. Fibonacci Code <anglikai@gmail.com> - 2013-11-12 06:57 -0800
        Re: New Data storage method with severe compression capabilities. Michael Harrington <michaelharrington4rep@gmail.com> - 2013-11-12 09:40 -0800

csiph-web