Initial Witness Pay & Number of Witnesses

[bytemaster]

1.5 BTS per block.

[bobmaloney]

Was there any kind of consensus here?

What are we going to be starting with?

[mint chocolate chip]

The OP left out the job of a Witness also includes publishing price feeds and the costs associated with doing so?

[r0ach]

+5% Let the code do it, dynamically.
https://bitsharestalk.org/index.php/topic,18549.msg238114.html#msg238114

Your solution doesn't have any Sybil resistant parameters.  Why not my collateral based solution? 

https://bitsharestalk.org/index.php/topic,18584.0.html

[kenCode]

I don't think you guys are in touch with reality if you don't think 17 is a public relations disaster.  It's not good security wise either.  Easy to collude.  It's a bad solution from any angle.

+5%

+5% Let the code do it, dynamically.
https://bitsharestalk.org/index.php/topic,18549.msg238114.html#msg238114

[betax]

even if we want less witnesses they must be dynamic auto changing...

even if we want less witnesses they must be dynamic auto changing...
nobody will successful accuse us for the number of witnesses when they know the decision was made from the blockchain-code and not decision's made from one individual, in our case Daniel Larimer (doesn't really matter the name)...

I thought the same, we could have a set of 100 (or more) grouped in 20s and dynamically change them.

Can we not make a list of who wants to be a witness? We can use named forum users.

[liondani]

even if we want less witnesses they must be dynamic auto changing...
nobody will successful accuse us for the number of witnesses when they know the decision was made from the blockchain-code and not decision's made from one individual, in our case Daniel Larimer (doesn't really matter the name)...

[liondani]

I don't think you guys are in touch with reality if you don't think 17 is a public relations disaster.  It's not good security wise either.  Easy to collude.  It's a bad solution from any angle.

+5%

[r0ach]

I don't think you guys are in touch with reality if you don't think 17 is a public relations disaster.  It's not good security wise either.  Easy to collude.  It's a bad solution from any angle.

[xeroc]

This may be a quick but a little dirty solution for those that care about the pay, the amount of individuals in control and the perceived decentralization:
Since "perception is everyting" .. why not have 17 individuals that each run 3 witnesses (3 different accounts)
That way we have 51 "witnesses" ... and only need to pay 17 machines :)

and you can clearly communicate it that way:
if you are not ok with that guy running 3 out of 51, vote him out, or reduce the amount of witnesses with your stake

Is there only around 17 trusted individuals capable of running a witness?
It could be seen from the outside as a closed loop.
Too much control can be counter productive.

My proposal is pretty much exactly the same thing as BMs proposal in the OP .. it just removes the perceived issued with 17 being "so small" .. :D

[r0ach]

This solves all problems with DPoS imo:

https://bitsharestalk.org/index.php/topic,18584.0.html

[iHashFury]

This may be a quick but a little dirty solution for those that care about the pay, the amount of individuals in control and the perceived decentralization:
Since "perception is everyting" .. why not have 17 individuals that each run 3 witnesses (3 different accounts)
That way we have 51 "witnesses" ... and only need to pay 17 machines :)

and you can clearly communicate it that way:
if you are not ok with that guy running 3 out of 51, vote him out, or reduce the amount of witnesses with your stake

Is there only around 17 trusted individuals capable of running a witness?
It could be seen from the outside as a closed loop.
Too much control can be counter productive.

[liondani]

Pay X% from the total marketcap annually for security.  (that means you pay with bts, total share supply / 100 * X)

and pay the witnesses relative to their total votes.
More votes for a witness means a bigger share from the X% "pie"



PS    X= a percentage value that change dynamically depended on what the shareholders vote

[r0ach]

I'd like to see 51 at launch with the power to vote to expand that number as the network grows.  The way it works is, Bitcoin uses delegated representation with mining pools, but if the pools do something bad, it has quick fault recovery.  DPoS can be more decentralized in that regard, having a much larger number of"pools", but it has worse fault recovery.  Due to this dynamic, the DPoS system obviously can't utilize a similar number of block validators as Bitcoin, it needs to be greater since fault recovery is worse.  Anything below 51 is also bad for collusion.

I was going to wait till after launch to run for delegate but can probably get up to speed on running one beforehand.

[xeroc]

This may be a quick but a little dirty solution for those that care about the pay, the amount of individuals in control and the perceived decentralization:
Since "perception is everyting" .. why not have 17 individuals that each run 3 witnesses (3 different accounts)
That way we have 51 "witnesses" ... and only need to pay 17 machines :)

and you can clearly communicate it that way:
if you are not ok with that guy running 3 out of 51, vote him out, or reduce the amount of witnesses with your stake