Topic: A Peer-To-Peer Currency Exchange and Transport System Using Confidence Chains (Read 1285 times)

a thorough reply is at https://bitcointalk.org/index.php?topic=220155.new;topicseen#new

what use cases would that be?

I disagree with that as the main assmption and use case.  Most share trading is done via buying and selling, not via redemption (share buy-backs).  I made some more elaborated comments on this thread you pointed to now about the additional value PoW provides for shares:

https://bitcointalk.org/index.php?topic=220155.new;topicseen#new


I encourage re-examining requirements, its a habit of mine also, and that is often how improvements are made.  However I think you maybe missed some use cases for PoW/block chain security in share trading.

Adam

The system does appear to have some usability similarities to Ripple, but most importantly there is no XRP like commodity currency(although users can build in tx fees if they want).  Also it has many structural similarities to Bitcoin and doesn't require major rethinking to get underneath the UI level, so presumably it has more appeal to the Bitcoin crowd.  It's also fairly simple to implement.  Ripple was another one of those kind of projects that started off one way, and ended another due to the influence of VC.

The Confidence is a more generalized definition than PoW.  It can be used in a number of different ways.  Initially, the Confidence settings will be STATIC and BOUNDED(a predetermined set of notary nodes).  What this gives you is a multi-owner exchange that cannot be disabled without disabling every node.  Also the trades must abide by the consensus rules, you cannot get this by sticking orders in the BTC block chain.  Far superior to Mt. Gox or similar services.  I don't attempt to solve the 'unbounded consensus' problem.  There are other ways to derive confidence values, for instance membership in a social network.  I have not really explored these aspects much.

a good background paper by Ben Laurie: http://www.links.org/files/decentralised-currencies.pdf

There is a pretty good thread explaining some of these points here: https://bitcointalk.org/index.php?topic=220155.new;topicseen#new , if you have the time to get into it.


there is no unbounded consensus.  There are no party crashers, no generalized participation thus it is not subject to sybil attacks.

the fact that the Issuers MUST HONOR THE LEDGER TO REIMBURSE, gives us an additional factor to rely on here.  Bitcoin was designed to be a currency without backing and without ownership.  We have shifted the goal to 'asset index' or 'bearer shares' (I think thats the term you used).  This allows us to safely remove PoW because ultimately we must use a ledger which is honored by the issuer.  This is discussed here: https://bitcointalk.org/index.php?topic=220155.new;topicseen#new

It's a subtle difference in requirements that results in a major change in architecture.

-bm

Doesnt this have the same problem as ripple?  How do you assign confidence.  I believe pragmatically ripple is assigning confidence based on external business relationships with gateways.

Plus some manual trust links by individuals.

In either case this is either vulnerable to systematic sybil attacks, or legal attacks to the gateways, where because consensus has no mining security, they can be ordered to undo history.

When that happens it damages fungibility, and introduces dispute cost and defensive reputation management cost into the transaction layer, and fraud cost when those best-effort mechanisms fail, and then you are back to the status quo.  ie there is a reason businesses go to equifax, and that credit cards are expensive, beyond banks having regulation costs and high margins due to barriers to entry: paying for the above reputation evaluation counter-measures and fraud costs.

Adam

https://docs.google.com/document/d/1cKlN55wX7n0SLvxidLoFVrJnNMJO-Iefr8bVyeHBseg/edit?usp=sharing