Email exchange between MIT Media Lab and the IOTA Foundation [pdf]

Post by mbgaxyz » Sun, Feb 25, 2018, 4:45 PM UTC

strictnein wrote: post #83

KKKKkkkk1 wrote: post #81

As a participant in the SHA hash function contest who broke one of the 51 Round-1 SHA-3 proposals and who worked on security proofs for another SHA-3 proposal I can say with some authority that using the sponge construction and showing statistical properties of the transformation function is not sufficient to ensure security. Of the 51 Round-1 SHA-3 proposals all of them passed statistical tests and at least one roun…

You may dislike the listing of authority, but he's right, and you're kind of misconstruing what he's saying. You can "invent" all the crypto systems you want, you're just foolish to do so without proper review, because there will be flaws.

Post by sneurlax » Sun, Feb 25, 2018, 5:11 PM UTC

QML wrote: post #5

merlincorey wrote: post #2

Second Preimage Resistance is important with most cryptocurrencies based off bitcoin because they are often calculated as a hash of a hash, which is what second preimage resistance is all about preventing predicting. In this case, the `curl` function is not being used as a hash function, but a different type of mapping. Unfortunately, that mapping is supposed to be psuedo-random and now it is known that it is not.

How would using the hash of a hash strengthen the property of second preimage resistance? The space of all possible hashes is still 2^256. The only impact I see is that the cost of computing a hash is doubled.

Post by mikeash » Sun, Feb 25, 2018, 6:03 PM UTC

Ar-Curunir wrote: post #120

mikeash wrote: post #82

Earlier quoted context omitted.

That’s because other fields of computer science can either prove or demonstrate that their solution works. Cryptography almost never has solid proofs, and demonstrations prove nothing. When you’re working on something where “works great” and “completely broken” are almost indistinguishable, the only way to even have a hope of avoiding the second one is by having a lot of smart people bang on it for a long time.

What? Modern cryptography is based on proofs and definitions. Sure, we can't prove that SHA2 is cryptographically-strong (that imply P≠NP), but we can show that it resists certain kinds of attacks. Furthermore, assuming certain properties are satisfied by SHA2, we can order that different constructions based on it (eg a Merkle tree) are secure. Cryptography is highly mathematical.

Post by Blackthorn » Sun, Feb 25, 2018, 6:12 PM UTC

tmz wrote: post #95

amenghra wrote: post #84

The first time I heard about IOTA was when I read https://shitcoin.com/iota-cannot-be-used-for-iot-loss-of-fun... It’s a fun read and you’ll probably end up unimpressed to say the least.

Maybe you should also read https://twitter.com/ShitcoinDotCom/status/965910847645212672 It is from the same author

Post by charleslmunger » Sun, Feb 25, 2018, 6:41 PM UTC

rkagerer wrote: post #92

That was a painful read. Considering the amount of money invested into this coin, someone should have coughed up travel expenses to get these guys into a room together for a few days so they could more efficiently clarify themselves to each other, maybe establish a bit of trust and rapport from working alongside each other on a problem, and reach consensus as to what facts and opinions they agree on vs. what is in di…