Last month an altercation began between Charles Hoskinson of Cardano and Vitalik Buterin of Ethereum, which remains an endless topic to date. This heated discussion began when Buterin questioned the competence of Cardano's Ouroboros protocol.
Buterin had initially published on subreddit of Ethereum, some points concerning the differences between Ouroboros di Cardano and Casper of Ethereum.
On August 9, Charles Hoskinson made a refutation on Twitter in response to Vitalik's comment.
After the first blog post of Confutazioni di Cardano, Buterin replied with a post on Reddit. Part of his answer was the reply to Hoskinson's allegations that they were unaware of publications or evidence that sufficiently described Casper's way of doing business. He wrote:
"WTF? First of all, there is the Casper FFG card.Therefore, there is the complete FFG mini-specification.There both contain definitions of key guarantees (security, plausible liveliness , fairness …), along with evidence that they are satisfied.There, also there is evidence verified by the computer of the properties of Casper FFG. "
He continued to further accuse that the protocol had no proof of correctness:
"The Ouroboros protocol is analyzed in a model that is fully described: it defines unequivocally all the participants' programs, their performances and interactions, their communications – including network properties … Without such a model (especially in the Casper FFG white paper or other available sources related to Casper), it is impossible to prove the correctness of any statement on the protocol. "
On August 13, Hoskinson took to Reddit to respond to Buterin's pre accusations ceased to publish.
Later, last night [13th August] Buterin published on Reddit, to respond to the latest Hoskinson's refutation:
"Answering this is simple and also shows the importance of having concrete statements on protocols and proofs of their properties. "
Quotation points in reference to Ouroboros theorem 1 Genesis paper and Ouroboros 9 theorem Praut Buterin wrote:
- For example, consider a tenfold increase of network delay. Link Delta and 10Delta in the alpha inequality and you will see that in the second case an additional factor of about [1-f 9 Delta] will appear, changing the lower limit for honest and active parties.
- For reasons of concrete numbers, suppose that f is originally set as f * Delta ~ <0.03; in turn, this means that the protocol will provide security in the 10Delta setting assuming that the honest parts are at least 69% above, compared to over 50%
- So 1 / f is the "average blocking time" "(approximately), Delta is the network latency, so when Delta> = 1 / f * ln (2) (that is, the network latency exceeds the block time multiplied by> = 0.693), (1- f) Delta <= 1/2, so it is not possible to prove any security level.
In relation to the previous points, Buterin believes that security has been taken. He said:
"For me, it seems to me that you are assuming a security assumption that Delta <1 / f * ln (2) in order to get a guaranteed security / liveness, and therefore it would be more accurate to call the 39, synchronous algorithm, or is there something I still miss here? "