quotingI discovered something quite remarkable today after seeing podcasts with Sean Bowe [1] describing the new tachyon system and then one today with Ying Tong [2] mentioning the fabled 'sandblasting attack' . It turns out that zcash put out a ZIP zips.z.cash/zip-0317 with what seems to me extraordinary content: it says not that there is some resource limit for blocks, but that individual transactions *should* be treated thusly: fee should be linear in number of ins/outs, but 2 outs should be treated like one (for a privacy reason), that certain types of transactions (their different shielded pools) should not be discriminated, and they disrecommend relay of transactions with other fees, and then give a long RECOMMENDED section to miners on how to construct blocks. This is totally nuts - the miner incentive is always to maximize fee revenue, and while it can be hard to work under that scenario sometimes, it's crazy to try to say things like randomize your candidate transactions and only take high paying txs in this ratio, or similar, as they do. Bear in mind that the sandblasting attack, which genuinely crippled the network afaik because they couldn't verify, in a normal node, as fast as the attacker could create transactions, happened because they had the insane idea of a flat fee for every individual transaction, no matter how big it was! (to be sure, they must have done that for better privacy, but it's an utterly broken concept).
nevent1q…4gl3
These are some of the very smartest cryptographers in the world, and I am not exaggerating for effect, there. How did they get such batshit insane ideas (or lack of ideas?) about how a permissionless p2p network works?
[1] its on the recent Zero Knowledge podcast, look it up, [2] the recent BTCKVR podcast 'BitVM optimizations', around 35 minutes
#cryptography #bitcoin #zcash
npub1va…knuu7 on Nostr: Oh, and in case you were wondering, isn't something like witness discount similar? ...
Oh, and in case you were wondering, isn't something like witness discount similar? The answer is a resounding no! Think about it - what the witness discount controls is *how much stuff can go into a block* and is therefore a *global* consensus rule. If miners break it they fork off. Here what the ZIP is discussing is having everyone sing kumbaya and agree what kinds of fees are fair, reasonable and keep good privacy and ddos resistance for everyone. It'll work fine, until it doesn't.