Why is there so much interest in changing mining and

block times and everything related to that all of a sudden?

Стикер

DeFi (on chains AMMs) gets a massive UX boost from very fast block times

Also, Intermediate Blocks, which I am working on, are a thing for like 5+ years. It's not anything new. It's just that it was never needed before. But well, we did not have DeFi before.

well how about we get a few of those on chain AMMs going before worrying too much about optimising the protocol

Стикер

defi is not yet big enough to suffer from confirmations variance, but we regularly see people complaining, like the most recent one: https://old.reddit.com/r/Bitcoincash/comments/1ejzjo6/what_happened_to_the_10_minute_block_time/

Your tailstorm protocol is very good solution changing 10 min block or do anything that is not backward compatible or heavy change sha256 protocol is catastrophic

we are not yet sure on full implications of tailstorm and final design, but it does look like a promising way to address confirmation time variance

No need to rush since zero confirmation is here already but at least it should be aid by escrow right?

yeah but escrow is not universal solution, it needs opt-in by the user and opt-in by the merchant

Which is why it is amazing as temporary solution as it is more secure than zero confirmation alone

Cauldron made a bad design choice back in the day, and are now caught in the sunk cost fallacy. Ignore confirmations is the way to go for defi.

The trouble is that two people can make competing transactions (two different people spending the same anyone-can-spend output at the same time) and you don’t necessarily know one will revert, much less which one.

Стикер

it's not really a bad design choice, just a tradeoff

Intermediate Blocks Whitepaper 23% ready BTW

it doesnt solve 0 conf?

Is having a confirmation "solving" 0-conf? When you get a confirmation, it's not 0 conf anymore :) imagine this, you broadcast a TX, you have 0conf, mempools pick it up, and then: - after some 15s, you see 1 subblock confirmation, you now have assurance that your TX is picked up by at least 1 miner - after another 15s, another 1 subblock is mined, but in parallel, and it has a double-spending TX, oh no! But this is also a super DSP, because it works for any TX type, and can't be dodged by bad miners because incentives are such that they have to announce their block now, can't mine in secret else they risk reorg later - after another 15s, your original subblock gets extended by another subblock, so it's 2:1 in favor of your original TX ... the race continues until enough subblocks become known, and some miner makes a full block from them

to be fair, he has a point. Cauldron seems conflicted in trying to be a central match making service, but not actually taking the advantages that this brings. Like for instance being the one in control of the unconfirmed chain. That's the choice. Because if cauldron, which in effect is a central place already, owns that being central and in control, then there will not be anyone creating those conflicting chains. And then we don't need faster block times.

ya but no one understands that it has to be part of the UX

Стикер

maybe someone can explain first how a Cauldron utxo chain is possible to spend by a user not using Cauldron. I don't think wallets do that usecase. How is it possible for me to create a conflicting spend with a complex script that won't get accepted unless it's exactly according to spec. I'd guess that this software doesn't actually exist outside of Cauldron

Exactly if strong final settlement guarantees is not there the system wouldnt be trustless

Either we give secure solutions at the expense of time or give high time at the expense of security and decentralization. Carefull solution need to be thought before reckless implementation

Стикер

the rest of the comment you didn't quote is relevant, lest someone reply with links to code.

Стикер

It does because you will make transactions secure cryptographic fast then wait 10 min for sha256 to give the first conf