--- Log opened Fri Aug 28 00:00:01 2020
10:30 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #utreexo
10:33 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 240 seconds]
11:50 -!- belcher_ [~belcher@unaffiliated/belcher] has quit [Quit: Leaving]
12:40 < adiabat> I don't think non-inclusion is that useful for bitcoin.  Unless it was an stxo accumulator, then in combination with an SPV proof it shows the txo can be spent
12:49 < adiabat> I need to read that paper more closely, but I think a difference is that in order to remove an element from that accumulator, you need to have the whole thing
12:50 < adiabat> whereas in utreexo, if all you have is the roots, and someone gives you an inclusion proof, that's enough to delete the element proven
12:50 < adiabat> in this paper it looks like if you want to delete something, you need to know more than just an inclusion proof for that thing
20:14 < adiabat> OK reading it more:
20:14 < adiabat> the deletion actually doesn't seem like much of a problem and could be adapted to something like the utreexo model
20:15 < adiabat> you'd just need to keep some data about which leaves are near you as part of your inclusion proof & deletion data
20:15 < adiabat> so maybe a little bigger but I think would still work
20:15 < adiabat> the bigger problem in this paper for our model is that *additions* need lots of leaf data
20:16 < adiabat> since the whole leaf set on the bottom is sorted, if you make a new utxo, in order to insert it you need to know what comes before and after it
20:16 < adiabat> so you'd need a bridge node & proof to create a leaf
20:17 < adiabat> that means you'd never be able to get rid of bridge nodes, even in theory, since you don't know where your new outputs will show up
20:18 < adiabat> though in practice utreexo will probably never be able to get rid of bridge nodes because it's bitcoin, if you started a system from scratch it could
20:18 < adiabat> also it just seems much nicer to not need any proofs to put new elements into the accumulator
20:19 < adiabat> I didn't even realize that was needed for a while when reading the paper because I'm used to utreexo and... there's nothing to "prove" when inserting :)
--- Log closed Sat Aug 29 00:00:02 2020