DPOS versus POS - should humans vote?

2년 전
in eos

The debate around POS versus DPOS seems to represent quite a philosophical divide. I have read and listened to various views, some quite strident, polemical. From a technical standpoint I am satisfied that both systems are resilient to attacks, that neither is inherently more vulnerable, that both approaches are viable. So why choose one over the other? While both approaches are viable I think that they would nonetheless produce very different chains.

In POS to become a block producer I need to stake a bond and have suitable hardware to run the software. That's it. The implementation then takes over without any more of my involvement. The hardware hurdle is significant. How good does it have to be to run the software? Perhaps systems will be designed that will be able to harness distributed/disparate computing power making for a very decentralized platform. If anyone with even a small stake and a home pc could participate that would be great. But it's a challenging problem. I hope it can be done in a scalable way with low latency and sufficient transaction throughput. I hope that it can be done in a way that can succeed.

In DPOS block producers are voted in and campaign for the privilege, perhaps offering something more, working harder. More concentrated and powerful hardware that allows the platform to scale is deployed in the service of the community. Steemit has 21 block producers with many more at the ready waiting for your vote, campaigning for your attention. The voting makes the system agile. DPOS can respond to change more easily since change is not always a code change. Change, for some things, is only as far away as a vote. DPOS puts the community centre stage. It asks us to participate, requires us to pay attention. When B1 talks about aligning interests it is human interests that they are talking about, it is human interests that are being aligned. Not just the consensus algorithm's interests. You express your interests with your vote.

I'd like to see POS ala ethereum, rchain, cardano etc. go forward. I can't wait to see what happens, to see how they evolve. They might be well suited to some types of applications where human voting is actually not desired. Where humans get in the way. Maybe human voting is fundamentally flawed.

Or maybe if we can't participate, if we can't co-operate, if we can't be bothered to understand the issues and cast our votes then we are stuffed anyway.

POS versus DPOS. An interesting debate! There is fantastic research going on around POS that will undoubtedly lead somewhere. I am certain that important benefits will emerge. But so too with DPOS. I guess that what I am really arguing is that DPOS has certain properties that make it central to the decentralization of politics and governance. It is the only consensus algorithm that is socially disruptive in a good way! The AI apocalypse is less likely under DPOS! BTS and Steemit prove that strong communities can form around DPOS. Of course there is more to be done, but that's up to us. This article by @samupaha talks about how DPOS could evolve to become even more human centric, how it could be even more in the service of and run by us humans. A great post.

At the end of the day we are building these systems for us humans aren't we? Which consensus algorithm do you vote for?

While it will be interesting to see how it all plays out in the years ahead I believe that we need a scalable platform now. ASAP. And notwithstanding some unexpected developments elsewhere I believe that platform is EOS!

Am I missing something? Do I have it all wrong?

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  trending

Thanks for the links! I have read them and I agree - DPOS could be just flat out better. But the POS advocates are so dismissive. I keep feeling that I must be missing something - they are intelligent people. A comment I read recently, maybe one of yours, wished academics were doing papers on DPOS, getting into formal research on it. But I am interested in the automated consensus versus human mediated consensus angle. That might also make DPOS more desirable.


in summary, we don't know what's best. see bitfury pos vs pow paper where they admit dpos solves a ton of problems, but have a lot of uncertainty as we do to this day.

you are a good writer..



Thanks to provide Great New and Useful Information

I am a pioneer in the market of cryptography. I follow your messages. Thank you for sharing!