>BTW, do you have any links for "inconditional data compression?" I ran a Google search just now and didn't find anything. You've got me curious :-)
A search for
"data compression" faq gets to the following, very exhaustive, page:
http://www.faqs.org/faqs/compression-faq/See part 1, section [9], "compression of random data".
The "counting argument" is simply that with 3 bits (for instance) you have 8 different possibilities. If you want to compress this to 2 bits, you have only 4 different possibilites. There is no way to uniquely assign any one of the 8 original variations to the 4 in the "compressed" version.
The fact that compression is possible at all is that quite often, data is
not random; some combinations are more probable than others.
Difference in opinions hath cost many millions of lives: for instance, whether flesh be bread, or bread be flesh; whether whistling be a vice or a virtue; whether it be better to kiss a post, or throw it into the fire... (from Gulliver's Travels)