Would it be possible to use an algorithm like the Library of Babel's to, instead of generating pages of text, generate strings of 1s and 0s to create any file type? Data on a hard drive could be stored as it's coordinates in the library instead of the files themselves, saving storage space.
Idk prolly
>>62089430
That's how computers already work dumbass.
yeah it would be
Since there are as many valid files as there are combinations of 1 and 0, each file's index would be no smaller than its contents on average. Introducing a seeded randomizer/sequence-of-pi factor just adds an extremely heavy CPU load on every save and is as likely to come up with assignments like:
Hello.<CR><LF> = lol 4chan spamfiltered the 1000-digit random number I had here
as it is
vicky_12.avi = 238
or
Master_of_Puppets.mp3 = 9082
>>62089430
This sounds awfully like block-level deduplication.