I have still waited for response from NYZO team about the block duration (7 seconds).
HOpe that the team will reply soon.
7-second block is too fast and risky for the network's health.
I haven't seen any emails come in with questions about the block duration; even checked the spam folder. If you would like to converse over email, could you try to send it again? The email is
[email protected]. I personally enjoy the theoretical questions immensely.
Also, if we'd like to carry on the conversation here, I'll start.
With a proof-of-work (or proof-of-just-about-anything blockchain), a 7-second block time would be a bad choice, in my opinion. I'd enjoy the engineering challenge of trying to make it work, but I think that you would be dedicating so little resources to any particular block, due to the amazingly short duration, that it would result in an unnecessarily weak system. Thoughtful engineering can only do so much to overcome a fundamentally poor design.
Proof of diversity is very different, though. In most blockchains (speaking very generally here), we use cryptographic signatures to secure transactions, but then we use proof of some sort of resource (computing power, hard drive space, with some modulations like proof-of-stake) to secure the blockchain. In proof of diversity, we also use cryptographic signatures to secure the blockchain. We have rules that govern how quickly we allow new verifiers into the blockchain, and the only valuable signatures are the ones produced by verifiers that have managed to get into the current cycle.
Not only does proof of diversity provide a clear definition of which keys may sign blocks, it also provides a clear order in which those keys must sign blocks. We have a scoring system that defines the optimum block for any height, and we have an automatic consensus system that allows the verifiers to agree which block to "freeze" at a height. When a block is frozen in Nyzo, that block is final and cannot be removed from the blockchain. In order to freeze a block, we must have 75% of in-cycle verifiers vote for that block (to avoid a less-than-50% sabotage of the blockchain, there is a manual override for this down to 50%, but that would be an exceptional circumstance). Nyzo verifiers don't start working on the next block in the chain until the block before it is frozen, so if Nyzo is extending a version of the blockchain, that version of the blockchain has 75% or more of the weight of the cycle behind it. Whether this is done in 7 seconds or 10 minutes doesn't have any effect on the strength of the blockchain.
There are some very real engineering issues associated with such a short block time. I'll save discussion of those for another post, though.