Oddbean new post about login | logout Btw another thing about ecash that I think people get wrong a lot is, the fact that it has a central point of failure is not just about trusting with money but it's just a really bad feature of an architecture; consider how much worse the privacy of it is because you have to send messages to a central server every time you make a transaction.
> you have to send messages to a central server every time you make a transaction. This is not necessary. As long as the ecash wallet has enough coins of different denominations, the sender does not need to send a message to the mint server. For example, Keychat’s ecash wallet will exchange a large number of 1-sat coins with the mint server at once, and these coins are then used as stamps. The sender doesn’t need to continuously send messages to the mint server.
Those 1 sat tx can become inefficient over time, mints are expected to incurr fees at some point (a minimum fee of 1 sat is realistic). There are msat keysets where this becomes cheaper though.
Sure, sorry to both you and @calle if I was unclear, I obviously do understand that only the receiving party (say, merchant) has to be online/contact the mint. I still think that it is an extremely bad architectural feature that each transaction requires such an online message *to a central party* to be sent. Btw it's important to note that many of the early papers (say 1990 period) on ecash were discussions around achieving offline payment, i.e., avoiding this problem. There was one paper that Chaum co-authored with someone else and there was def one by Brands, probably tons of others. They all focused on an idea of "proved fraud" in case of double spend, thus they used punishment mechanisms. I discussed this quite a lot in the early part of my talk in London, because I thought it was a pretty fascinating issue.
Only one of the parties usually has to contact the server at time of transaction.