Re: Repeatable compression is possible and easy to do, here's how...

it seems you're really into personally attacking people on these
boards without any actual research of your own.. so, i'll be patient
until you implode.

How much compression do you think you would be achieving for these
463 bit files? Something significant like 10%, or something
less significant like 1% (4 or 5 bits, say)?

Actually, you hint at an answer to that in the following:


i do actually.. funny that huh, an example of what i'm talking about
being referenced inside the actual message.

if you can imagine this you're set: 5000bits long already. looking for
30% compression. per test each file is either compressed 30% or
expanded 30%, the distribution is 50%:50%. 5001 bits nope. 5002 bits
nope. 5003 bits nope. 5004 bits yep. success.. 5004 bits to 3503 bits.
recomposition microcode parameters 300 bits long. 300 bits + 3503 =
3803 bits. saving 201 bits total. Add another block of information
from the stream to get > 5000 bits to get ratio back.

My god, you're a complete freaking loon.

oh there it is.. so lets see here, i'm not talking about 'me' at all,
i'm talking about streaming compression algorithms. unless, of course,
you think i'm a streaming compression algorithm and you're not a
streaming compression algorithm, in which case i'd say to simply shut
up and observe since i'm going to know what i'm talking about and you
know nothing about it. or else you're seeing me as a human being and
are attacking because you like it, in which case i'd say 'Upgradz UR
SKILLZ OR STOP phr3@kin 'out''

one of those will stop you, great. back to the story in progress.


Your technique is likely to turn a 5000-bit file into a file so large
that it wouldn't fit onto all the hard disks ever manufactured.

Phil

actually dude, i've said that only 50% of all 5000 bit files will be
turned into resultant files that have 30% less bits in them. to
equalize on that, 50% of the 5000 bit files will have 30% more bits in
them.

this, is logic. obviously it'll work like that to conserve entropy.

the trick, oh skill-less human or non-streaming compression algorithm
thing, is to look at it as a streaming sequence and apply maximal taps
at all entropic points, all places where the pattern shifts.

because, you see, if you have something that has the same 50%:50%
ratio for 30%:30% expansion and compression, and you can do it for
5000 bits and 5001 bits and 5002 bits and 5003 bits in length, and
*each* of those different tests gives the maximally random
distribution on what exactly compresses and what exactly expands.. at
4 samples there is a 200% chance of finding a 30% compression matrix.

easy. you just have to be there to get the pun.

chris.

.