Topic: Delegated Proof of Stake (DPOS) White Paper by Daniel Larimer - page 2. (Read 11482 times)

Finally the real reason for all the sudden interest comes out. You and your group awarded all to yourselves and are now promoting it. I don't know whether you really believe what you wrote in that post is it part of the act.

I don't think you developers should work for free, but you should be honest. If you want to award even 50%, state that clearly and others will still support. Trying to keep it hidden, awarding yourselves all and claiming it was fair later on won't fool us here. We have seen enough to spot when it happens. I suppose I shouldn't be surprised as Bitshares are closely linked to the Ethereum group who even with all their advertising efforts couldn't fool the community with their IPO.

Good luck with your project. Its very interesting and anything which can emerge as an alternative to the cartel has my support. I will just wait for a Counterparty equivalent of your project.

Agree! (Perceived) Fairness correlates with how many people know about the possibilities, that is why the principle is to be as transparent as possible and make threads also here on BTT like this one. Everyone is welcome to join on our forums. Also Daniel and the whole team goes to conferences regularly to make Bitshares known as much as possible. We also head a lot of mainstream media coverage.
The rule is to give at least 10% to PTS/AGS. It is up to the developer of a DAC that uses Bitshares codebase to decide which allocation gets him the most support for his project. Mostly it is more than 10% for PTS/AGS... With Bitshares X it was even 50/50. The snapshot for Bitshares X took place on the 28. February. There where delays due to problems with Bitshares X out of which DPOS came out as a solution.
We discussed the Maidsafe fiasco here https://bitsharestalk.org/index.php?topic=4090.0

I knew Protoshares was the IPO coin, but didn't know about the other coin. The prices are much lower than what they used to be so its quite a good bargain being informed late

You are awarding 10% at the start as IPO. How are you distributing the remaining 90%? If you have another IPO for that I would be interested. Whatever you do, make it a fair one, even a proof of burn style process as Counterparty is the only new project which has any credibility. Look at what happens when you invest in a unfair distribution with greedy developers like that of Mastercoin, you get screwed as they try and make their money through deals with third party like Maidsafe.

The facts are: The Bitshares donations period goes a long 200 days and everyone can participate!! That is more than half a year. Whenever a DAC is ready there is a snapshot. And the distribution will be a whole lot more equal than with Bitcoin and most POW altcoins. Also the code is completely open at ALL stages. Check it out on https://github.com/BitShares/bitshares_toolkit
The last question you raised is indeed interesting

Hi bytemaster,

I think this is some good progress. The great weakness in PoS is that money gets distributed on the genesis block. This is potentially much more corrupt than the initial bitcoin distribution. Although coins can always be purchased on an open market, this process leads to mis-trust from the beginning. BCNext chose a rather fair distribution via the forums, but this was before the release of the code. It will be interesting to see how PoS clones will do on the market, because there is no hurdle of creating clones at all. Also what will be interesting is the relationship between corporations (Invictus LLC, ethereum GmbH, ...) and currencies. I think creating legal entities creates unnecessary attack surfaces.

Exactly, how Nxt will work.

Invictus recommends to all developers using the Bitshares codebase to honor BTS PTS (aka Protoshares) holders with at least 10% as well as BTS AGS holders with at least 10%.
Bitshares XT is tested at the moment. You can look at it here and test it yourself https://bitsharestalk.org/index.php?topic=4480.0;topicseen

Sounds quite good. Do you have anything running or is this all just hypothetical at the moment? You speak as if you have the algorithm working and have survived through attacks.

They cannot produce dubious transactions... if they double sign then that is incontrovertible proof the delegate should be fired and is easily detected by all nodes on the network.  In less than 50 minutes every delegate should have confirmed your transaction and a 51% attack / double spend becomes impossible.  So large financial institutions have nothing to fear if they wait just 1 hour... not even bitcoin has this level of security... it takes 6 blocks and even then a fork could result in a doubles spend. 

All clients automatically recognize when a block has been missed and the 'confirmation window' is automatically lengthened and wallets are warned.   No one would lose money... everyone would immediately see that 49% of the blocks are being dropped despite being produced on time.    The client would view this as a network-split at least and automatically inform everyone to stop trading until the fork is resolved.  Under normal conditions the network would reconnect and all delegates would start producing blocks again and the fork will be considered resolved.   Under an attack the 'fork condition' will remain until the attacker can get enough votes to fire the other 49%.... this could take a while and may be impossible once people realize what was happening.    The minority (good) delegates and majority (good) clients would manually flag the attackers blocks to be ignored.  In a relatively short period of time the attacker's delegates would be voted out of the minority fork and it will quickly regain its title of being the longest chain.   

So as you can see there is a simple algorithm in place that allows the good majority of shareholders to quickly recover control from an attack.  No funds can be stolen in this process unless people choose to ignore the automatic chain fork detection warnings. 

Is it just stop including transactions or putting in some dubious ones too? I can see how a 51% attack can benefit someone trying to manipulate big money. You will be running banks, lotto etc so the incentive is there.

wouldn't last for 24 hours doesn't look too good if bitshares is aspiring to replace the big financial institutions we all hate now.

First of all someone with 51% of the delegates only has one power:  to exclude transactions or to stop producing blocks. 

If they gain 51% of the delegates and fail to include transactions that would vote them out then it will be very obvious to everyone on the network, the blockchain will be hard-forked by some developer and the network will continue to function.  You see what binds a network is not the software rules, but the social contract among the community.   It is almost inconceivable that in the event 51% of the delegates stop including transactions that the community would sit back and take it for long.   

Competition among chains and no barriers to entry or forking keeps people honest because there is NOTHING to gain by a 51% attack. 

Note that these 51 delegates would have to pretend not to see blocks from the other 49 delegates and thus the attack would be so obvious and disruptive that it wouldn't last more than 24 hours before community members took action. 

What happens if somebody sets up 51 delegates, provides incentives to get the votes, and then takes over? Its much easier than trying to get 51% of the mining power.

What are you even talking about? CPU coin means botnets mining at a rate which is unprofitable for normal users. Which means it will be controlled by thieves.

If you are going down that route I would rather stick to the cartel.

AnonyMint, you are welcome to trim your arguments and debate Bytemaster on mumble on Saturday 11 AM EST https://bitsharestalk.org/index.php?topic=4203.30

Anyway anyone is invited to ask questions and critique concepts.

The main problem is concentration of power. That is the whole reason for the decentralization movement.

So the most important thing we can aim for is decentralization of knowledge. Centralization of knowledge is one of the main causes of centralization in a Proof of Stake network.

Knowledge such as which blockchain to invest in. People who don't understand the difference between DPOS, POS and POW will be at a knowledge disadvantage. As a result they will lose out in the race for the limited token supply. Even if Proof of Stake is decentralized it will still favor those who have the most knowledge so there is always going to be a centralizing force. The question is whether or not it's the kind of knowledge that anyone can acquire and if it's the kind of knowledge anyone can acquire then at least you can say there is no barrier to entry.

For example people who think Dogecoin is going to be a great Store of Value for their life savings should have done a bit more research. People who think Proof of Work is going to be as good as a Store of Value as Proof of Stake need to do more research. There will be a lot of centralization around the brightest minds and most knowledgeable persons in a Proof of Stake community while in a Proof of Work community the centralization is around whoever has the most expertise and best equipment.

Money is central to both because usually if you want to invest it's going to cost money and if you want to mine it's going to cost money. Usually the more money you have the more you can make up for lack of knowledge or expertise. People who mine the right coin at the right time do so because they either have the knowledge/expertise or they just have a lot of money to buy whatever equipment they need to keep up with the race for tokens.

Perfect equality is impossible. All brains do not have equal knowledge. All people aren't on the same place on the starting line. Some people have had years to study programming, to study economics, to study cryptography, and have been waiting for something like Bitcoin to come along so they could play with it. Those people are the majority of the early adopters who were paying attention to these subjects prior to Bitcoin even existing.

Then you have people who started paying attention when Bitcoin went over $1, and these people are generally technologists, programmers, geeks, nerds, or whatever you like to call them. But the demographics who understand the technology the best are positioned to benefit from the technology the most and it's always that way even if you try to make it fair.

DPOS in my opinion should be measured compared to what is currently out there and not some hypothetical or mythical Proof of Work which doesn't currently exist. CPU Proof of Work is not going to be decentralized unless everyone can 3d print the CPUs. Proof of Work is not even designed to be decentralized because if you wanted true decentralization you would go direct democracy by saying every registered human being gets to vote on every block rather than every CPU gets to vote.

There will be way more CPUs than human beings. This favors CPU makers because CPU makers will always have the advantage from the start. It also favors the electric company because most people favoring Proof of Work don't have solar panels to generate their own electricity. You have to pay these third parties just to play the "serious game" we call Proof of Work mining.

Proof of Stake is a "serious game" as well but it's simple. Anyone can play this game by simply buying tokens. These tokens act like lottery tickets and by owning them there is a chance that you'll mint new tokens. This could be done in proportion to how many tokens you own (proportional) or it can be randomly distributed such that someone wins a jackpot. In either case this is more decentralized because you remove the need to have expertise in chip making.

You also flatten the hierarchy greatly by removing the need to pay an electric company a huge amount. This way the electric company doesn't decide who can and cannot mine profitably anymore (this limits participation).
It's possible to decentralize POW more than it is. I had some discussions on the topic and the best idea I could come up with was to juggle the hashing algorithm and try to make it pseudo-random or even random.

That is really the best you can do. You can make it so people will not know what to invest. Eventually people will start buying all the chips and investing in everything and once again it would centralize around whoever has the most money, but it's definitely better than what Bitcoin is doing.

Bitcoin is not decentralized at all and isn't trying to be. This is why I don't understand why they cannot raise the Max_Block_Size because it's already centralized so what difference does it make?

Bitcoin has major problems and it's not just ASICs. The Bitcoin community suffers from a concentration and centralization of knowledge. The core developers have so much power because there are so few of them and there are so few of them because Bitcoin code is esoteric for the average programmer to understand.

As a result we have a lot of altcoins which exist to allow programmers to learn the code base but also to experiment. That is a good thing because it decentralizes the knowledge and expertise, but for some core devs it could be seen as a bad thing.

Nice discussion going on here so let me chime in a bit:

Before you can call something decentralized or centralized you must first define what these terms mean in a concrete and objective manner.  I will define them as follows:

1) You are fully decentralized when you are fully in control of your own assets.  The free market and voluntary trade of physical goods is full decentralization. 

2) When fractional ownership of a common good is in play, then you are decentralized when your control/influence is proportional to your own assets.


Mining can never be decentralized because the miners are in control over the shareholders assets and control/ownership are completely decoupled.  Sure you can sign a transaction, but it must be 'approved' by the miners.   


I like Nxt and their transparent forging approach is the closest thing out there to being truly decentralized.   Unfortunately, because only a subset of the users forge and there is no restriction in how concentrated forging power can be.   

I think that this entire debate of mining vs POS is founded upon a lack of understanding on what Bitcoin really is.  It has not solved the Byzantine Generals Problem... it just defined a rule set for consensus that allows anyone with money to control the network and once controlled there are no alternatives.

Likewise, CPU only POW merely means the government will certainly control it given the percentage of the population employed by the government and government-friendly corporations and that each of these employees has a computer.   Mere economic analysis is all that you need to understand that POW is the ultimate means to assure centralization because it erects barriers to entry that are insurmountable by everyone wanting fork as a result of the miners taking the chain in the wrong direction.

Because of a false belief about the nature of consensus many seem to think that 'Zero-Trust' models are the only ones that are valid.   I contend you are always trusting something or someone.  Random selection of anonymous individuals to produce blocks is both slower and allows 'random' bad behavior on a temporary basis.  It is less reliable.   

DPOS is all about creating a system that gives people financial incentive to do the right thing and maintains the ability to 'fire them' quickly if they don't with 100% certainty of getting caught.   The delegates have no power *BUT* to do the right thing.   

Can they collude?  Assuming they did collude the only thing they could do is block transactions and it would be instantly detected by all peers.   A fork would be launched immediately with these delegates stripped and the two chains would compete against each other in the market.  Assuming the delegates were just random people that attempted to screw with the system then consensus would be maintained.   






 

What about botnets of zombies? The number of computers that can be infested is practically limitless.

Long ago I realized that botnets are a finite resource, especially with a cpu-only currency there is competition for the same resource, thus if you increase the demand for them, the price will rise eventually to the point where it costs the same as buying the hardware.

Problem mitigated or solved with sufficient scale.

ASICs and proof-of-stake are worse because there is no hope of keeping them decentralized.

I am not only a programmer, I am (perhaps a polymath) and above demonstrating I am an astute economist.

Another form of centralization.