Increasing Filesize & Compression Ratio

Bernard Robertson-Dunn
Tue, 15 Oct 1996 19:48:48 +1000 (EST)

Colin Richardson wrote:

>Beyond compression?  Stewart Fist once wrote of "stuffing a wombat down a 
>garden hose", so I guess file compression is analogous with transmitting 
>a perfect miniature of the animal but receiving it full size.  Why not 
>send the wombat's tiny genome instead?  Perhaps DNA is a good analogy for 
>today's FIF (for graphics), MIDI (for music) and TVML (for virtual 
>reality) file formats.

There are two sorts of file compression "no loss" and "some loss" (I refuse
to use the single word versions)

Some types of files such as .txt .doc and .exe can suffer no loss and there
are limits to the amount of compression.

Other files or data such as voice and image can withstand various amounts of

There are also different compression schemes. One removes redundancy,
another uses different encoding styles. The wombat example uses the DNA
encoding to represent the original. Unfortunately just sending the DNA of a
wombat is not sufficient to reconstitute the original wombat. Learning and
experience within the aformentioned wombat would be lost.

The reason for this brief burst of theory is that the compression ratio of
176:1 mentioned in Colin Richardson's posting would only be acceptable under
certain conditions. I would not like to see people recommending that all
files be compressed to this ratio.

Reducing traffic congestion on the Internet is like many other Internet
issues, not simple. That does not mean that we should not try to educate
people to use the rresource responsibly.

Any comments on the suggestion that users should be licenced, like gun
owners, car drivers and dogs?

Ducks and runs away quickly.

|Bernard Robertson-Dunn                   |
|Canberra                                 |
|Australia                                |
|                       |