Just reviewing the cloak coin anonymity claims. This is interesting quote from the Whitepaper: "Again in the Phase 2 pass the originator and recipient are not recorded, instead the transaction occurs between the Phase 1 and Phase 2 nodes."
My problem with this is:
1) the transaction are broadcast and, thus, propagated to all nodes. A malicious node could simply record all these transactions and easily publish the records linking the transactions.
2) the elected node could itself be a spy and publish records of these transactions for all transactions it pretends to anonymise.
Further, there are a few other serious problems:
3) there are only two levels of mixing, which is hardly enough to provide any real guarantees
4) no mechanism is described that would prevent an anonymising node from stealing outputs
Hopefully my concerns and questions can help put the discussion back on to a productive track.
Thanks.
Obfuscation. You have to think about PoSA like tor.
1. We can black list malicious nodes and rely on the fact that there will be so many nodes on the network (unlike centralized masternodes) that the majority will not be spy nodes and to find any relevant data will be next to impossible. Nothing is 100% approaching limits is the best way to handle things.
2. Same question as 1
3. 2 levels out of a random selection. powers of two magnify probabilities of not being interfered with exponentially.
4. Again the same questions
Thank you for the technical post. Hope this answers some.