Could you just procedurally generate a massive reference file of every possible combination of 1s and 0s, then use that to "compress" data by referencing the reference file? Then all you could have to store is the algorithm to generate the reference file and the "compressed data" that references it.
such a reference file would be infinite in size since there is no end to the possible combinations of 1's and 0's.
>>62147633
The data referencing the file would be about as big as the data you are referencing. So you don't gain anything. It is basically another form of PiFS.
https://github.com/philipl/pifs
>>62147647
What if you limit it to n byte blocks? There aren't an infinite amount of those, even if you make n exceedingly large
>>62147685
>https://github.com/philipl/pifs
Why is it so shit?
The idea sounds great.
>>62147692
Then how would it be a reference to everything? It would be a reference to SOME things, and very limited things at that.
>>62147705
Simple:
>say you have a 1GB file
>that's a *specific* ~8 trillion digit binary number
>you need to find this exact sequence in pi
>the number describing it's starting location in the sequence is itself over 8 trillion digits
In the end, you're basically asking "how many random 8 trillion digit numbers must I generate before I get the exact one I want"?
>>62147725
True but I'm also not OP and was just curious.