Topic: Ethereum Casper (Read 1238 times)

That's an entirely different attack vector. Having a permissioned set of known block producers makes the system more vulnerable than one in which anyone who qualifies can produce a block, since in the latter, the network recovers automatically.

Technically ALL proof of stake systems are permissioned ledgers, since if I choose to buy up the entire order book of some random PoS coin on Poloniex, I'm now the boss and overlord of that coin.

The kinda interesting thing about the consensus algorithm is how long they spent trying to eliminate the possibility of pools (there were several hypothetical iterations of PoW that were discarded).  But within weeks pools represented most of the hash power on the network.

That's a fragile approach, though. If some force targets all 101 block producers, the network is unrecoverable. A permissionless system is less fragile. Quite frankly I'm amazed they're even considering changing their consensus algorithm.

Not sure what you said but i think i agree.
Bottom line is a better solution for distribution is needed i think.
Especially at the start.

IPO's are scammy bs.
But mining is heavily exploited unfairly too.

THIS is the area where the majority of experimentation and research should be conducted.

Tacking a market or anon feature onto a poorly designed and launched coin is not time well spent.

I for example had proposed some ideas on how to use Ethernet or Wifi hardware as a Proof of Work system.
And that would obviously be meant for the initial launch also.

EDIT:
And do i get this correctly ?
These guys IPO'd the coin then are tacking on POS later ?
+ other misc drama / negatives.
And how is this the hottest coin right now ? How is it ruling Coin market cap ?
What am i missing here ?

ahh sorry i forgot..... g r e e d 

PoS can't redistribute from HODLers to users. Users drive the economy. HODLers strangle it.

However, PoW in its current incarnations can't redistribute either, due to economies-of-scale in mining such as ASICs and cheap electricity.

It's supposedly going similar to the Darkcoin route, except with an automated delete your collateral for misbehaving mechanism to try and negate "nothing at stake".  Ethereum Casper and Tendermint seem identical so most pros and cons will apply to both.

With most traditional PoS systems, the big holders just get more mega wealthy over time.  In the DPoS system, block validators don't collect an infinite number of fees as the network increases transactions or anything.  Delegates set a burn rate % to decide how much of the transaction fees to destroy.  In other words, delegates set their own pay rate and stakeholders can either vote for expensive delegates or vote for cheap ones and receive equity gain/stock buy back mechanism via the burn.

Elements of the DPoS system are both ingenious and stupidly overly complex all at the same time.  It's designed for maximum vertical approach, while I think Vitalik wants to attempt to turn this Casper thing into a partitioned network eventually.  In a collateral based system with a fixed rate, the working set of validators is inherently designed to decrease over time unless collateral is set trivially low, which it cannot be for security/functionality reasons.  Combined with the DAPP overhead, Casper would probably have more centralization than Bitcoin.  Bitcoin usually just appears on the surface to be overly centralized, when in reality, it's not because of all the transient miners, which Ethereum would not have.

Unless it brings you extreme problems in state recovery, I feel like a fixed number of block validators using a variable collateral bid system is superior to the fixed rate collateral system.

I haven't studied this. Is Ethereum going the Dash masternode or Bitshares DPoS route of the rich get richer and usership disappears?

Won't this need to have minimum collateral set artificially high in order to prevent people from intentionally attacking the system by bonding with sub par hardware?  Meaning this system will converge into something like 20 whales running the whole thing with enterprise hardware?  It feels like a better solution to have a fixed number of block validators like 101 (or other number), such as DPoS, then whales have to collateral bid for those positions.  In other words, a fixed number of validators + variable collateral bids instead of fixed amount of collateral and variable number of validators.

The vulnerability would still be there, but seems like it would at least be a slight improvement.  Using an artificially high fixed collateral amount would probably just send your working set towards 1 or 0 over time.