42.zip: A single 42,374-byte zip file that uncompresses to one million 4.5GB files (4.5PB)

I guess the only way this file could have been created is recursively unless I am missing something.

Link to gzip quine:

http://groups.google.com/group/comp.compression/browse_threa...

and it contains at most 42,374 bytes of information, no matter how verbosely it is spelled out once unpacked. compression doesn't hide information, it just displays it in the most concise way possible.

but call me up when you can archive information into thin air, such that the data can be smaller than the smallest self contained equivalent. there some ways to do this. for example, "the first million digits of pi" is a very concise way of referring to a number with a million digits. and pi doesn't have to be stored anywhere, it can always be calculated. so the information is compressed to pi[0..10^6] plus the overhead of the algorithm for calculating digits of pi. since the algorithm would be of constant size, there would be a threshold (likely less than 10^6) after which this approach would be more spacewise efficient than a self contained zip.

too bad there isn't an easily recognizable off-the-shelf constant for every million digit number one would want to compress. pi is transcendental, so EVERYTHING is theoretically in it, but it's tough to find anything in particular and, if you do, it's going to be REALLY deep. and referencing something at a random location deep in pi would probably take no less space than the original data we're compressing.

i wish there was something like that, though. it wouldn't be practical, but... for theory's sake.

This is a Trojan Horse.

YC admin should remove this post immediately.