Compression - Re: Please, PLEASE, hold your questions/comments/elsewhat til the end. Thank you. :)

Author

Re: Please, PLEASE, hold your questions/comments/elsewhat til the end. Thank you. :)

Threshold

2006-03-31, 9:55 pm

Not really. I have more than a clue, as far as compression,
compression theory, information theory, and whatever else you'd like to
throw into that general category goes. I actually have a CLUE about
programming-I understand it, but I'm not up to writing yet. And you
can't tell me you weren't there once, as well-everyone had to learn
once. And it's not gibberish, either. The simple fact of the matter
is that you are all under the impression that the counting principle
does not allow for recursive compression. The simple fact of the
matter is that IF 2^128-1 > 1, and IF 2^N + 2^N-1 > 2^N, which they
are, then the counting principle DOES allow for recursive compression.
The likelihood of finding an algorithm that produces intermediate
compressed files having a different MD5 hash than the original,
and-this is the important part...EITHER one bit smaller than the
original OR the same size, FOR EVERY FILE, is highly unlikely-I readily
admit that. But it's not impossible-impossible would be if all of the
intermediate files had to have a unique MD5 hash. They don't. And
highly unlikely and impossible are NOT the same thing.

I'm also fully aware that some of my original post content was...easy
to tear apart, shall we say. While my sleep-deprived state for one or
more of the previous posts is not a sufficient excuse, I would hope
that you all could give my current posts a fair treatment regardless.
Again, I've only CLAIMED to do one thing here-convert BPE to a 256
"base". And I never stated how well that went, either. All of the
rest of this is ideas, which was the original intent of MY posting
here-to discuss ideas.

If someone can PROVE, beyond the SHADOW of a doubt, that it is
ABSOLUTELY IMPOSSIBLE to come up with an algorithm such that all
intermediate files have an MD5 hash independent from the original (but
NOT unique) and such that all intermediate files are EITHER 2^N-1 OR
2^N size, WITHOUT being a copy of the original (as that would mean they
had the same MD5, would it not?), for all files (of course-and if you
fall back on the counting argument, it doesn't count-2^N is always less
than 2^N plus something) then I'll be happy to drop this. As I said
previously-prove me right, and I'm happy. Prove me wrong, and I'm
happy. It's just ideas.

Personally, I don't really believe it's provable OR disprovable. It's
sort of like the existence of (a) God...the only way to really prove it
is to present and, it it's not too easy to disprove it, either,
although many have certainly tried. But I'm not trying to get into a
discussion on religion here, just attempting to use an analogy.

As far as I can see, the math supports my position. If you feel you
can open my eyes, without merely telling me "You're wrong, and
everyone's already told you that"-as you may have noticed, my position
has become a bit more refined and the objections that have been
presented previously are no longer valid in this context-then please,
do so. But I'd really like a WHY.

Thanks. This has actually been enjoyable, so far.

Sponsored Links