@SpunkyEnigma@steveklabnik I have my own config library I use for Rust CLI tools. It supports multiple languages, but I basically only use Rhai. The README has an example of what a Rhai config file might look like: https://t.co/cUzViMRpgq
I would perhaps vibecode a simple Rhai reader to try things out.
@TheZvi As a thought exercise, just think how you would watermark random looking sequences of 0s and 1s without "loss of quality". You could just use a good RNG with a secret key to generate the bits. If you later want to prove that you are the author, simply reveal the key.
@ProgrammerDude@cafkafk Austrians/Swiss Germans are just as bad in that regard though, probably even worse. Maybe having an awkward Germanic language/dialect is the secret recipe for success.
@Jonathan_Blow@BlakeNastri2403@veorq Finding a collision for a proper cryptographically secure hash function (say BLAKE3) is on the order of 2^128 hash evaluations, which would take billions of years with current technology. That's not a gotcha, that's merely stating facts that you seem rather confused about.
@Jonathan_Blow@BlakeNastri2403@veorq If you truly believe this then you should disconnect from the internet forever because the cryptography securing everything we do rests on the assumption that, no, you will never ever get a collision, even if you try your hardest.
@Leonidas76167@Jonathan_Blow when we're dealing with probabilities to the tune of 10^-100 then that's irrational nonsense, but we'll never gonna agree on this
@Leonidas76167@Jonathan_Blow People use hashes as shorthand IDs for optimization. The collision likelihood for say BLAKE3 is so absurdly low it's quantifiably inane to not make use of it. You're more likely to have a cosmic ray flipping a bit WHILE you're getting hit by lightning and it's still not close.