Introduction to Decentralisation | Hacker News Reader

Introduction to Decentralisation

ledgerpath.com

77 pointstheocs17 comments

I love how most blockchain related articles I read pass from “baby-talk description of decentralization and distribution” to “PoW, permissionless, PoS, miners and mining” in a single swipe.

If I could understand those terms why did you bother to explain such basic concepts to me in the first place?

If I need to be guided throught those ideas, do you really think I can comprehend what are you talking about with Proof of Work? (Which is never spelled in its full words in the article)

By the way, I always maliciously infer that this is the actual depth of the understanding of the author. You just grasped some very high level concepts and have a very pratical knowledge of lower level details — but just chaos in the middle.

I may be wrong, obviously.

I have seen this phenomenon so often in explanations/tutorials of different things. People seem to not have a concrete target audience in mind. They simply write of cuff.

I'll go one step further: for many slightly-obscure technical topics, the intermediate writing between beginner and advanced is rare or simply doesn't exist. I struggled with this when I was trying to learn functional programming--writing on the subject is mostly linked-lists/lambda/map/reduce/filter, or όλα είναι ελληνικά για μένα. This goes even for better writers who do have an audience in mind: their audience is either complete beginners, or people with advanced academic degrees in the field. It took me years to find writings that bridge that gap.

What material did you find? I too have noticed this gap and been searching.

I found a lot of the stuff on Matt Might's blog helpful. He writes a bit about compiling Scheme to C, which allows me to see what each piece of Scheme is doing in a language I know.[1] Following this approach, I found LISP In Small Pieces[2], Compiling With Continuations[3], etc. There's more but I have to dig it up.

[1] http://matt.might.net/articles/programming-with-continuation...

[2] https://pages.lip6.fr/Christian.Queinnec/WWW/LiSP.html

[3] https://www.cambridge.org/core/books/compiling-with-continua...

This X 1000. I tried learning programming several times in my younger life, maybe 3 or 4 efforts at getting some 'basics'. Never thought I would be able to program until I picked up Unity on a whim and started learning C#. Lo and behold, documentation is now plentiful and tutorials are a vast landscape.

Nearly every 'beginner' lesson or tutorial or book on programming that I have seen state in the beginning that 'some programming experience recommended.'

A decent editor (a person, not a "text editor") would quickly solve this problem.

>> - Architectural decentralisation: how many physical computers the system is made up of? What’s the topology of the network and its geographical presence?

>> - Political decentralisation: how many people or entities are in control of the network?

>> - Logical decentralisation: does the whole system behave like a single entity?

>> The take away from Vitalik article, is that all blockchains combines these three dimensions into one system.

Actually, Vitalik states in his article that individual blockchains are logically centralized: https://medium.com/@VitalikButerin/the-meaning-of-decentrali.... It depends on whether we're talking about a specific blockchain or the set of all blockchains. That's why I don't agree with the position of Bitcoin maximalists; their vision is not fully decentralized. When you factor in the centralization of mining power, BTC is only architecturally decentralized. Bitcoin is not so different from a centralized system which has replicas in multiple data centers.

It is the architecturally decentralized nature of bitcoin that you mentions, that can make it become more and more decentralized over time.

You say it is not so different from a centralized system that has replicas in multiple data centers.

- if we speak about bitcoin nodes: don't forget that centralized systems can be shut off very quickly by governements

- if we speak about miners: new mining pool software is just being developed (and some of if finalized), that takes away power from pools, and gives it to the many small miners:

https://medium.com/hackernoon/betterhash-decentralizing-bitc...

and

https://www.reddit.com/r/Bitcoin/comments/ci2daj/making_mini...

I think the point cryptica is making is that there is only one, central blockchain (logical centralization). This has a few centralizing effects on political centralization which is what most users probably care about:

1. Governments need only to track one blockchain to track every BTC transaction. Attempts to deal with this mainly entail coin tumbling, which involves giving your coins to a centralized trusted entity. Larger tumblers are more effective, creating an incentive for further centralization.

2. The central blockchain is a performance bottleneck to event throughput. Attempts to deal with this mostly have to do with federation, i.e. forcing people to transact through another entity (usually a company you pay fees to) if they want faster transactions. Federation is much closer to centralized on the centralized/decentralized spectrum. Usually companies respect regulation from centralized government (meaning a government could force them to not process your transactions) increasing centralization further. And historically, federation happens unevenly, i.e. a large number of email users federating under GMail, which creates further centralization.

3. Mining competition has a very high entry cost because the minimum investment for entry is very high. The solution to this is pools under a centralized pool manager who typically collects fees and may or may not collect data on pool members.

Blockchains have a single centralized view of data and this view is distributed worldwide with everyone keeping each other in check. Blockchains are centralized but distributed in this way.

Federation protocols have different, decentralized views of data and these views are selectively distributed with peers and distributed in ad hoc ways, and sometimes peers keep each other in check. Federated protocols are decentralized and distributed in this way.

>“hypothetically”

Second word of the article and i almost stopped reading. Is this a sarcastic hypothetical, where what follows is obviously true?

>This is an impossible structure for an organisation to function

So... Why is this "hypothetical" instead of hypothetical?

I suppose im being unreasonable, but i find this extremely offputting and intrepet as a very unprofessional writing.

It's weird that decentralized architectures are becoming what the racoons and quants are now flocking to via means of block chain via means of cryptocurrency. Seems like they are getting pulled down the buzz word rabbit hole and we are the only ones the wiser

I call writing “code for the human compiler”. That code should address the knowledge level of reader. No general purpose article can get that right

Oh, it was blockchain again. I tought it would be about the political practice.

... or about a myriad other technologies.

It's sad that cryptocurrency speculation and hype sucked all the air out of the room vs other decentralized tech.

In the end its leading to a condition where decentralized equals Ponzi finance scheme in the minds of many people.

Reply on news.ycombinator.com