You are viewing a single comment's thread from:

RE: Reputational Enhanced Delegated Proof-of-Stake (REDPOS)

in #blockchain4 years ago

Aren't witnesses suppose to be like UNL trustees? So, what would qualify as a UNL? When you say Unique, does that mean unique IP Address, MAC Address, that is the physical device of the witness, the node, the server? Do you want to hard-wire into the blockchain a mechanism for verifying unique nodes or IP addresses for each UNL? But what would you do when you run into proxies?

REDPOS Over DPOS

REDPOS would be better than something like DPOS to the extent that REDPOS can integrate REP into the DPOS system. Yes, it would stop the Tron puppets which came out of nowhere in February of 2020, at least in theory. So, I like this REDPOS idea. The details to how REDPOS would work specifically could be debated and tinkered with.

Sort:  

If you could translate the UNL idea into STEEM, it's like a witnesses list that any witness would have. So for example, @blocktrades could have an "UNL" with 50 witnesses that he would like to consider for determining consensus.

But you can't translate that idea literally into DPOS withut loosing some trustless aspects of the protocol.

The idea is just to understand that the UNL list represents "reputation" for each server.

On REDPOS the equivalent of an UNL would be to just include the REPUTATION of each witness as part of the witness rank.

So for example: An account less than 2 years old, could be excluded (by protocol) from witness elegibility. Or if not excluded, acount age would be a factor affecting the "voting power".

image.png
source

But what if a new account was very popular?

Well it depends on how it would be implemented. The parameter "account age" could be used to exclude the posibility of ranking as witness unless a set mininum account age is achieved. In that case no other parameter of REP (followers, upvotes, etc) could help to avoid the exclusion.

If that is what people want, they should vote on that. I may be in favor of accounts that are at least 3 months old. Well, I might be ok with no minimum age required and base the requirements on REP, etc.

I'm thinking more in terms of human society. STEEM is just 4 years old, but would you put a 3 month old in charge of it? A baby ruling a kindergarten.

50% of STEEM age would be a better guarantee. Of course if STEEM was a 100 year old community, a 50 years old mininum account age would be too much, so a limit on the requeriment could also be established.

Something like: At least 50% of STEEM age or 10 years. A very stong barrier for newcomers.

Age is not relevant because age will not automatically make a sock puppet account become real. Tron could make a bunch of accounts and those accounts could eventually become old enough for your standards. And that means they can some day be old enough to ride, like tall enough to go on your roller coaster. What really matters is letting people vote for witnesses. But at the same time, perhaps Steem needs states and an electoral college.

a bunch of accounts and those accounts could eventually become old enough for your standards.

If you can delay an attack for 2 years, that's a feature. It's a lot of time to prepare a defence, don't you think?

PS: I'm not trying to present a bullet proof idea, just ideas for a better protocal than DPOS, making use of data provided by the use of the accounts over the years and somehow incorporate that data on the protocol, per comparison increasing the cost of an eventual attack.