Decentralization of Power

[Troglodactyl]

...
The reality is that bytemaster controls enough stake to vote anyone in/out at will - getting elected as a delegate right now is pretty much impossible without his stake voting for you.
...

I don't think this is true anymore.

The biggest problem I see with Bitshares right now is Bytemaster doesn't seem to acknowledge his opponents as having valid critcism.  While it might sound harsh, Anonymint for instance says, Bytemaster is a Steve Jobs type control freak figure so he's not going to make anything remotely decentralized.  This criticism has some merit because I honestly believe Bytemaster is designing a system with the assumption that he + this website will always exist.  In reality, the system needs to be designed in a manner where it functions at maximum efficiency assuming neither this website or him exist at all.  The system also needs to be designed in a way that assumes mass voter apathy and ineffective or non-existent delegate campaigning to find suitable candidates for security that fit game theory motives.
...

I think this is legitimate criticism.  I think a lot of people around here (perhaps bytemaster included) are too comfortable with reliance on bytemaster continuing post October.

If witnesses are voted for based on collateral, then withness pay would have to compensate them by effectively paying interest on that collateral.  Is that expense really worth it?

I also suspect that we'll have less voter apathy (especially of large holders) once we have a better UI.  Better gateways so less BTS lives on centralized exchanges will also help.

[bytemaster]

  This makes me think he believes he's always going to be there annointing the delegates himself.  Afterwards, talking about trying to lower delegate count to 17 on top of that so you only need an amount of people you can fit in one car to collude is just getting outrageous.

The reality is that bytemaster controls enough stake to vote anyone in/out at will - getting elected as a delegate right now is pretty much impossible without his stake voting for you.

I actually like your collateral proposal; I think it would be better to remove the requirement to burn 50K BTS to register as a delegate, and instead have that posted as collateral. Having a least one game theoretical aspect to becoming a delegate is better than none.

Well I would definetely prefer a bidding war over burned funds than over collateral. Once it's burnt it's forever, collateral is just temporary. Would make people who burnt more funds to commit more seriously. An extra motivation for not letting their funds go to waste.

From an economic perspective, burned funds is a sunk cost and should be ignored by everyone.   

[monsterer]

Once it's burnt it's forever, collateral is just temporary.

The idea is that collateral acts as an incentive not to be evil - if you're evil, then BTS loses value and therefore so does your collateral. If you burn it, you don't care about the burnt BTS losing value, since it's gone anyway.

[Akado]

  This makes me think he believes he's always going to be there annointing the delegates himself.  Afterwards, talking about trying to lower delegate count to 17 on top of that so you only need an amount of people you can fit in one car to collude is just getting outrageous.

The reality is that bytemaster controls enough stake to vote anyone in/out at will - getting elected as a delegate right now is pretty much impossible without his stake voting for you.

I actually like your collateral proposal; I think it would be better to remove the requirement to burn 50K BTS to register as a delegate, and instead have that posted as collateral. Having a least one game theoretical aspect to becoming a delegate is better than none.

Well I would definetely prefer a bidding war over burned funds than over collateral. Once it's burnt it's forever, collateral is just temporary. Would make people who burnt more funds to commit more seriously. An extra motivation for not letting their funds go to waste.

[r0ach]

I actually like your collateral proposal

Requiring a non-trivial proof of burn to run for position of delegate which you might not be elected to disincentivizes anyone from running at all.  I'm not sure who thought that was a good idea.  A collateral bid system that just locks the funds but doesn't destroy them is far better.  Bytemaster is even the guy that at one point said something along the lines of proof of burn being useless and a stupid waste of capital.

[monsterer]

  This makes me think he believes he's always going to be there annointing the delegates himself.  Afterwards, talking about trying to lower delegate count to 17 on top of that so you only need an amount of people you can fit in one car to collude is just getting outrageous.

The reality is that bytemaster controls enough stake to vote anyone in/out at will - getting elected as a delegate right now is pretty much impossible without his stake voting for you.

I actually like your collateral proposal; I think it would be better to remove the requirement to burn 50K BTS to register as a delegate, and instead have that posted as collateral. Having a least one game theoretical aspect to becoming a delegate is better than none.

[Akado]

I thing the average Joe is happy if he gets security wise a 2FA option and he gives a shit if it's more or less centralized...
So whatever is your last decision about the number of witnesses make sure we have the 2FA future ready ASAP on all clients!
It will be crucial for mass adoption. Don't delay it to long! Is it possible to have it on launch ready?

https://github.com/cryptonomex/graphene/wiki/Wallet%202-Factor%20Authentication%20Protocol

  Well said!

[r0ach]

The biggest problem I see with Bitshares right now is Bytemaster doesn't seem to acknowledge his opponents as having valid critcism.  While it might sound harsh, Anonymint for instance says, Bytemaster is a Steve Jobs type control freak figure so he's not going to make anything remotely decentralized.  This criticism has some merit because I honestly believe Bytemaster is designing a system with the assumption that he + this website will always exist.  In reality, the system needs to be designed in a manner where it functions at maximum efficiency assuming neither this website or him exist at all.  The system also needs to be designed in a way that assumes mass voter apathy and ineffective or non-existent delegate campaigning to find suitable candidates for security that fit game theory motives.

I mentioned there's no automated fallback mechanism for voter Apathy, yet can easily be fixed by having clients automatically vote for delegates that put up the highest collateral bid if they don't vote manually, but Bytemaster doesn't even seem to believe voter apathy is a problem.  This makes me think he believes he's always going to be there annointing the delegates himself.  Afterwards, talking about trying to lower delegate count to 17 on top of that so you only need an amount of people you can fit in one car to collude is just getting outrageous.

Since those 9 people colluding require no substantial collateral to be delegates in the first place, they can just sell all BTS they own, accept a bribe from "Come-from-beyond" at NXT, all place shorts at the exchange then black swan attack the network all day to destroy it or at the very least all of it's credibility.  Bytemaster is unable to even envision this attack because he seems to think he's going to be there manually controlling everything forever.  Even if your goal was to straight up run a centrally controlled company, everyone on the outside believes that since you're operating under the Bitcoin banner, that attack vectors like this cannot be allowed or it's not a valid product.

If your goal is to operate a centrally controlled company, things like this still can't be allowed because in a centrally controlled company, the board of directors all have tons of stock in the company and million dollar salaries.  No substantial amount of collateral is even required to be a delegate in BTS, and nobody is making anything near a million dollars.  The only way this system can work with a low number of delegates (anything 100 or less is low) is if they're all required to lock up large amounts of shares to function as one to prevent themselves from attacking the network  (i.e. my collateral bid solution).  The minimum bid would need to be somewhat high, then you would need to bid against others on top of that.

[jakub]

Did I miss something? I thought in BTS 2.0 the number of wittnesses anyway depends on the amount of witnesses shareholders want AND can vote on (one can not choose more wittneses than once has selected to vote for).

I wondered the same.
My impression is that this discussion is only to help us make an informed decision when we vote in 2.0.
But I think that however this discussion ends, there is not going to be any hard-coded number of witnesses except the lower bound which I believe is going to be 10.

[santaclause102]

Did I miss something? I thought in BTS 2.0 the number of wittnesses anyway depends on the amount of witnesses shareholders want AND can vote on (one can not choose more wittneses than once has selected to vote for).

[Empirical1.2]

What anybody here thinks is technically the 'best' option is, is largely irrelevant.

The only relevant question imo is whether the 'best-selling' option, significantly adds to costs or security concerns.

Every business and product in the world is trying to find the 'best-selling' option.  They do extensive market research and modify their product and packaging so that it's presented in a way their market will like the most. This forum is almost unanimous, (even the people that agree with the 17 number here), that the perception among the current alt-coin market will be negative.  So we should shoot for a higher number of witnesses and also perhaps place limits on proxies imo.

 
Excerpt from Rich Dad, Poor Dad. One of best-selling book series of all time...

A few years ago, I granted an interview with a newspaper in Singapore.  The young female reporter was on time, and the interview  got under way immediately. 

"My work does not seem to go anywhere,” she said quietly. “Everyone says that my novels are excellent, but nothing happens." 

“Someday, I would like to be a best-selling author like you, Do you have any suggestions?”

“Yes, I do,” I said brightly. “A friend of mine here in Singapore runs a school that trains people to sell for many of the top corporations here in Singapore, and I think attending one of his courses would greatly enhance your career.”

She stiffened. “ “I have a master’s degree in English Literature. Why would I go to school to learn to be a salesperson?"

“Do you see this?” I said pointing to her notes.  On her pad she had written: “Robert Kiyosaki, best-selling author.”

“It says best-selling author, not best-writing author,” I said quietly. 

“I am a terrible writer,” I said. “You are a great writer. I went to sales school. You have a master’s degree."

The world is filled with smart, talented, educated, and gifted people. We meet them every day. They are all around us. I am constantly shocked at how little talented people earn. 

“They are one skill away from great wealth.”

[bytemaster]

Producing blocks is only one part of security.  Providing seed nodes is another.  Attacking the P2P protocol is a third.   Of the three of these, attacking the block producers is probably the most difficult because no one knows their IP address.   Attacking the seed nodes on the other hand could completely disable new connections.   More importantly, attacking the P2P protocol could temporarily completely disrupt all communication among witnesses. 

How would an attack on the P2P protocol and an attack on seed nodes work?

Attack on the seed nodes is simply a DDOS.
An attack on the P2P protocol would be similar to DDOS, create many connections to fill up the maximum allowed connections on all publicly available nodes.
To prevent these attacks from disrupting the CONSENSUS process the witnesses should form a PRIVATE (Invite Only) P2P network and then bridge to the public through several bridge nodes.    A witness could still operate on the PUBLIC P2P network but attacks on the public infrastructure could end up causing them to miss blocks.

[bytemaster]

Is cold storage an option somewhere in the new system?

Yes, simply update your account authority to a public key that has never touched an internet connected machine.

[santaclause102]

Producing blocks is only one part of security.  Providing seed nodes is another.  Attacking the P2P protocol is a third.   Of the three of these, attacking the block producers is probably the most difficult because no one knows their IP address.   Attacking the seed nodes on the other hand could completely disable new connections.   More importantly, attacking the P2P protocol could temporarily completely disrupt all communication among witnesses. 

How would an attack on the P2P protocol and an attack on seed nodes work?

[mf-tzo]

Is cold storage an option somewhere in the new system?