I didn't think of it as centralized because anyone can be an exchange and whoever wants the branch can choose the exchange.  Also, the choice on whether to peg the conversion or let it float is an option of the exchange and it's store.  But I understand that this forces openness on a not-necessarily-open process and standardizes for a particular use.  And I can certainly see the benefits of distributed administration of the backing (though would still point out that such an implementation of an exchange is not prohibited either).

I think, though, it's important for me to better understand the proposals here before appearing to suggest that existing practices solve the use cases.  I was just trying to get a feel for what they were targeted at solving, and your answer is a good start!

Thanks.  I thought there was some greater guarantee being implied (with later mentions of moon math). I see this doesn't actually prevent all ways of taking coins from others, and others have expressed the same possibility.  This is an area I'm trying to get a better understanding of, as I've been analyzing other algorithms, and I keep seeing these discussions pop up.

I still don't see what this solves that altcoin exchanges don't.  You have an entity that takes value into it's store and gives out value in an alternate blockchain that may or may not follow similar rules (some altcoin, maybe a Bitcoin clone or maybe something quite different).  The exchange can hold onto that value until the altcoin work is done and exchange back the value.  Validation is done in whatever currency is chosen for the transfer, which can obviously have any of a variety of zero-knowledge transaction and block validation schemes.

I've always expected altcoins are the natural transaction scaling mechanism, so I'm probably biasing myself, though.

Do sidechains solve any real problem not already solved by altcoin exchanges?

Also, I am confused by:


I know that zk-SNARKs can resolve transaction verification anonymity issues, providing true security.  I've read the zerocash proposals, and am familiar with the mathematical basis.  But the information I have seen is that this actually makes 51% attacks easier (because it slows down propagation through the network, wasting node cycles and effectively decreasing the computational power of the network).  The link is to one of the original program execution verification papers and is unrelated to cryptocurrency except in the general way that execution verification is the point.  So eating newly minted coins is actually not stopped, and I imagine that is the security risk mentioned in passing here.

Is there a new algorithm being mentioned that goes beyond the zerocash proposals here?  Does this refer to something else?