undefined | Better HN

0 pointskeepamovin2y ago0 comments

That’s not really true, the priests are people.

This confuses (or may just respond to only one of) two things I’m saying: field is kept artificially weakened by the priests; and new designs by outsiders are deliberately kept out. These are connected.

What you’re talking about it seems is just genuine competence and performance. I’m fine with that. I’m okay to compete. But it’s not just competition we’re dealing with, it’s artificial bias, designed to keep the field weak.

With crypto hashes it’s more than just entropy: there’s many hashes that sit at the top of good entropy performance regarding smhasher results, but that’s not enough for crypto: you need analysis. And for analysis you need experts.

But if i was worried about just competing on performance, which I’m not (many of my hashes have excellent entropy results), then your piece world be encouraging, it’s a good pep piece! Well done!

0 comments

3 comments · 1 top-level

msla2y ago· 2 in thread

You haven't proven that there are better designs out there being kept down by the priesthood, just waved your hands in the direction of Grand Conspiracy and anointed Bruce Schneier a member in that supposed priesthood. The thing is, cryptography isn't kept obscure. It's a mathematical discipline welded to the practical discipline of actually writing software, but so are other topics in that part of CS, such as compression and formal verification.

But here's the nub of my point: It is hard, though, and it's harder than it looks, and this matters more for cryptography than compression because of the things people do with cryptography. If you design a bum compression algorithm, you make files you can't really decompress and you get discouraged. If you make a bum encryption algorithm, and you get all excited and actually use it for something important, and it's easy to break, well, you might get a lot worse than discouraged. Major companies have been bitten by this. Is your algorithm better than HDCP? Probably not. You know what happened to HDCP? Yeah. Intel wishes it had taken the kind of advice any new cryptographer gets about not taking their own inventions too seriously.

keepamovinOP2y ago

Sorry about my late reply! I've been very busy.

So, basically, nobody would have any interest in preventing an explosion of new strong cryptography is that right?

msla2y ago

> So, basically, nobody would have any interest in preventing an explosion of new strong cryptography is that right?

Not really.

j / k navigate · click thread line to collapse