On Tue, Jan 23, 2024 at 8:59 PM Thiago Macieira wrote:
 
3) there's a non-zero chance of false positive. It might be only 1 in 2^128
for random data (and read-only data is not random), but it's there.


I am not sure if you're being exhaustive and verbose just for the sake of being exhaustive and verbose, but the point you're making here is nonsensical.

I'm going to generate a 128-Bit random number right now, give me one second . . . ok here's 16 random bytes in hex:

    3c1fed6f74bb6c814173c5409cb7d0aa

That 128-Bit number has never been seen before. You won't find it in a database anywhere, nor in an encoding of a genome of an axolotl crossed with a salamander, nor from measuring the dimensions of fractal methane crystals on Europa, nor will you get it from another random number generator anywhere ever. Perhaps in the distant future if we can travel at many multiples of the speed of light, and encounter billions of billions of galaxies with trillions of trillions of homeworlds, we could ask the aliens there what 128-Bit numbers they've generated, and compare their list to the one's generated on Earth, but still we might not get a duplicate in this circumstance.

We don't live in a black and white world -- we live in a world of uncertainties and probabilities, and we go with that. Some people have a psychological reliance on certainty in life. If we have a reliable random number generator and it gives us a 128-Bit number, we can use that as a magic number to store in the read-only section of an ELF file.