The prize for most annoying word in the cryptoverse lexicon must surely go to: Decentralisation.
The amount of energy wasted on the concept is possibly on a par with that used to sustain every blockchain. Decentralisation is not even a thing; it is an adjectival noun that has become unmoored from any code to float in a haze of politics and social fluff.
A decentralised system in systems theory is a system in which lower level components operate on local information to accomplish global goals. The global pattern of behaviour is an emergent property of dynamical mechanisms that act upon local components, such as indirect communication, rather than the result of a central ordering influence of a centralised system.
Indeed, Wikipedia is a great example of how a decentralised protocol becomes a control-tool through the constant influence on the "local information" by coercive global forces. Thus, local influence leads to emergent behaviour under the big assumption of global ignorance. If there is nothing to stop "global knowledge", then there is nothing to stop "emergent centralisation". This is precisely what we see on many blockchains.
Indeed, I'm starting to get the feeling that this big-D is merely a propagandist tool like another big-D: Democracy. The process of democratisation is where a country is weakened by enforcing a system of government based on social ignorance backed by private coercion and violence. One need only look at the primary promoter of this concept.
The internet is itself already decentralised in the sense that it has a resilient redundancy built into the network. However, it is in no way under a decentralised governance system; it has layers of control, from private servers to government licenced ISPs. The "purity" of the internet as a concept is becoming fractured into digital islands.
I would replace the whole debate around decentralisation with a better word: dependencies.
Dependencies are much easier to define and analyse; they are also easier to resolve or classify as critical. What this also does is to clarify at which level of the network the dependencies exist. Whenever decentralisation is discussed, one always ends up with a porridge of concepts that confuse the important distinctions between different network levels. Hardware, infrastructure, software, financial and social issues take place at different levels and may have different rules and dynamics. Bowing at the altar of some big-D in the sky seems just a way to bamboozle believers while ignoring any solutions by avoiding to properly define the issues. This has happened; we have seen it happen - it continues to be unresolved and undefined on some platforms.
I don't wish to write a lengthy epistle, so let me close this with a concept I mentioned above: emergent centralisation. This is often so obvious that I'm surprised when it gets dismissed; it makes me suspect there are reasons.
Let's take a quick look at the idea of DPOS witnesses. To be a witness, you need some hardware and software - some functional dependencies so you can actually operate. But to make the activity net positive to the individual, that witness needs votes - votes from vested stakeholders. With 30 possible votes and top 20 places, it doesn't take a mathematical genius to figure out that witnesses are beholden to large stakeholders. That is how DPOS was designed. So, witnesses are sponsored by stakeholders - they are dependent on them. This would not be so bad if witnesses were merely guardians of the blockchain, but they also have the power to change that blockchain by the process known as forking, a collective change in the consensus algorithm. So, the governance of a chain is dependent on its witnesses who are then dependent on stakeholders. What could possibly go wrong?
Don't get me wrong: this isn't necessarily a bad system, but just stop calling it "decentralised" - it isn't. From what I've seen, none of them are; they are all prone to cartels, either explicitly or implicitly. But analysing the network of dependencies makes this much clearer than the quick-sand of ersatz decentralisation. In the example above, witnesses can be seen as decentralised at the level of block production but not at the level of governance because, as we have seen, centralisation can happen.
As the above definition states, true decentralisation works on local information and global ignorance. However, blockchain economies are really quite small; they are digital constructs that thrive on information processing. This means that every user has both local and global information at their fingertips. This also means that emergent behaviour can happen very quickly; economic consequences that can take years in real life can evolve in mere days in a cryptoeconomy. Thus any tendency towards re-centralisation also happens quickly.
The urge by some individuals to control those things that can be controlled wins out. What may have started as a decentralised system quickly evolves into a centralised one. In my opinion, that just means the original naive implementation was not decentralised in the first place - it was just badly constructed and over-sold.
Let me stress this point: if a decentralised system manifests centralisation as an emergent property then it isn't really decentralised at all. A truly decentralised system needs also to be decentralising, meaning that it self-corrects away from centralising tendencies.
It is an important question as to whether one even wants all levels of a network to be decentralised, but if one does then one must list all dependencies and then find ways to mitigate them. This goes all the way down to bare metal and all the way up to social interactions. One will be left with some critical dependencies, but at least one will be aware of them and put in place plans in case they become real threats.
Dependencies will control the system, so control those dependencies.
As a coda, this is the attitude of DARPA towards their blockchain research. For them, decentralisation means minimising dependencies. Not easy.
Please vote for my witness account @busbecq.
Can be done directly here.
Thanks!
So I understand that decentralization is at the level of the process within the chain, such as the production of blocks, but that this is subject to the group within the structure that makes decisions, then we could be in front of a consensus organization where the majority of the Group of 20, decide everything about the blockchain, is that right?
So, there is an important aspect that should be considered in making decisions that should not be neglected, such as human relationships or soft skills that each of the members of the witness team must possess, to manage under consensus. a better decision-making process.
An interesting topic, greetings Rycharde.
Good vibes.
Yes, good point - the off-chain relationships are often the main influence on on-chain actions (such as votes) that otherwise, looked from outside, have no logic purely within the ecosystem. But that ecosystem is much larger than the blockchain cryptoeconomy... hence... that system should respond in some ways to such actions. This is, indeed, where off-chain relationships can manifest in very similar ways to vote-rings ;-) even if that is not their intention.
Back to decentralisation, yes, we saw it in the JSun Steem-experience that, not only did one person control all top witnesses, but they were also all on the same physical machine, thereby making the chain dependent on one piece of hardware! I think it's issues like that that mean taking the time to develop and design a robust governance model - both attack vectors from outside and from within, some of that is through code and protocols but some is mathematics and game theory.
Thank you. It has bothered me over the years reading others make their proclamations of decentralized (and democracy as well).
Reminds me so much of the house divided falls saying.
I know you are super busy, but I do wish you would write more. You speak of truths most would leave to the shadows and ignore. Your insights continue to reveal a strong will that doesn't cower in the frightening power that surrounds us all. Such a will allows one to see a thing for what it is, not what one wishes it to be as so many do.
Many thanks - deeply appreciated.
Those few who can truly see what is happening need to focus on their personal power to have the clarity to make good decisions. That means having experiences of what that actually means - becoming suprahuman before being forced into being transubhuman. The Luddites were right - but they lost.
"Cardano is now the most decentralized blockchain".
But doesn't that also lead to cartels?
A large stake-holder just needs to fragment their holdings into more and more pools?
How do they overcome that?
They can fragment their holdings, but the different admins of the different stake pools may vote for different projects ... and they don't really depend on the stakes of certain stake holders because anyway their pools normally won't surpass a certain size ...
I don't say one couldn't circumvent these hurdles at all, but at least it's an interesting try, isn't it?
I'll have to dig up sources, but I do recall members complaining that Cardano had been over-run by cartels very quickly. Not sure what the feeling is now.
Using the language of Blurt, and looking towards Blurt2, if witness voters are paid a proportion of witness earnings, then the formula for that should be such that per-VP income should come down as the VP-votes increase, thereby making it more lucrative to also vote for those lower down the witness ladder.
However, for that to work it will need a different algo that selects witnesses for each block-cycle.
As always, feedback loops are necessary to avoid convergence towards centralisation.
Thanks for your input - will read the article link.
Ah, sorry to say, but that link is not really so detailed ...