Topic: Important Lighting Network reading- for everyone! - page 3. (Read 1262 times)

This is the first thread I've stumbled across dinofelis and you have confirmed my initial observation. The only saving grace to his tl:dr posts is that at least he normally highlights in bold the wrongest parts of his argument, which saves a lot of time.

You see, ad hominem is all you can bring.  You don't enter into a well-argumented technical argument.  Nobody ever did.  As I said, no single logical counter argument.  Like usual.  This is what confirms my understanding of the system.  Nobody ever succeeded in giving a logically built counter proof to the evident thing I'm saying, because I'm just repeating a basic design aspect of bitcoin, as it was presented in the original paper.

There WERE some valid arguments that have some validity to have sufficient full nodes, and these are:
- some advantages for the user himself (but that doesn't have external influence of course)
- in case of a massive crack-down or other catastrophe, to have some guarantee that "the blockchain will survive the cataclysm" (higher probability that at least one copy will survive)
- some resilience of routing through P2P if ever there were a global attempt to isolate the mining industry (the multi-server) from the users (clients).

But the argument that "full nodes keep miners in check" is totally false, is never contradicted by any logical argument. It only attracts ad hominem, which is a proof of its solidity.

As to whether this is off topic in this thread, no it isn't.  If one doesn't even understand the fundamental data flow and structure of the "layer 1", it is somewhat ridiculous to go and talk about the network properties of layer 2.

The fundamental network structure of layer one is a client/multi-server structure.  Users vs a backbone of mining nodes.  It is because some people argued about the importance of full nodes, that the discussion came about.  It is hence essential to talk about this.  Because in as much as a true P2P system has scaling problems, a client/multi-server structure doesn't.  The client/multi-server structure of bitcoin scales perfectly. This is what Satoshi explains in his second e-mail on the nakamoto institute document server, and what happened in reality.  I'm only essentially explaining again what Satoshi said back then, and what is obviously observed in reality.

In fact, that should be good news.  The fact that this basic observation (that Satoshi saw this in 2008, and that the system indeed, evolved this way) is met with such vehement ad hominem resistance, is remarkable, and in need for an explanation of course. That explanation seems obvious to me: where Satoshi had in mind a "P2P network of mining nodes" that would count still hundreds or thousands of nodes, in reality, this reduced to something like 10 or the like.  In as much as you can still sell the religion of decentralization of a P2P network of hundreds or thousands of mining nodes, with a reduction to 10 or so, that becomes hard to sell.   So one was in need to save the religion with a narrative.  The point is however, that we also observe that this higher loss of decentralization doesn't have, after all, negative effects.  Even though the mining pools are only 10 or so, bitcoin continues to work correctly.  Satoshi's P2P network of decentralized mining nodes is much smaller than he thought, but it is still there.  So this client/multi-server system, even though it is much less decentralized than initially conceived, is working quite well and scales even more easily.  The only problem it faces, is that the belief of its value proposition was attached to "decentralization".  People are forgetting that decentralization was a TOOL to make the thing work.  It turns out to be much less decentralized than anticipated, but it turns out that visibly, this decentralization is not needed for it to function correctly.  However, now that one has sold bitcoin everywhere as being special because decentralization, and that decentralization is the fundamental belief tenet, it is quite annoying that it isn't there in reality.  Even though we see that it isn't needed to make the thing work.

So one has introduced a whole narrative about a false form of decentralization, which brings with it a lot of technical difficulties (and unnecessary solutions), just to keep up that narrative, because its belief system, and hence its "value" is thought to be linked to that belief.

And the kid that says that the Emperor has no clothes, gets an ad hominem reply.

It would be more constructive on your part to give arguments, not ad hominem, but I know you have no arguments.

What else does your "checking node" look like from the outside, else than a talking IP socket ?   How do you hope that that talking IP socket is going to influence the Byzantine agreement ?  How do you justify the claim that "a big majority of nodes imposes its will on miners/keeps miners in check" apart from counting them ?  And what exactly do you count, apart from IP numbers ?



If I present you two different block chains, which one are you going to accept ?  The one with most PoW, right ?  If that's not a vote, I don't know what it is.  If you have a given group of miners, divided in two groups, each making their own prong of the block chain, which one is going to win ?  That prong that accumulated the largest amount of hash rate (PoW), right ?  It is because most PoW voted for that prong, that it won.  That's the essential principle of Byzantine agreement with PoW: of all different proposals, the proposal where most PoW put its vote on, wins.

How, in case of consensus disagreement (two blocks propagated in parallel), consensus is reached again ?  By the PoW vote of the next block.  The following block is mined on top of one of both.  So the block that got voted for by the next miner, and puts its weight of PoW on top of it, wins.  The block that is not voted for, is orphaned, because its prong now has less PoW voting for it.  If ever some other miners vote for the second block, there's no tie break, until one of the two prongs clearly wins the PoW accumulated vote.  That chain that has most PoW voted for it, wins.  And the property of PoW is that you can only use it once, that's why it is a "voting token": an amount of waste that cannot be re-used.

I guess that if even that is not clear, there's no point to discuss this further.  

Who said anything about IP numbers?  I don’t even expose one; I’m onion only.

Who said anything about votes?  Bitcoin is not a democracy.  There is no “vote by PoW”; that’s nonsense, and shows a total lack of understanding of what POW is and how it is used to achieve BFT transaction ordering.  There is no vote at all.


Quoted for craziness.  You forgot to add that the CIA designed Bitcoin, and gives all these very knowledgeable people their mendacious talking points about nonexistent “decentralization”.  I admit, I work for the CIA, too.


The design of Bitcoin is a subject about which you demonstrate worse than zero understanding, insofar as misconceptions must be unlearned.  You really ought to go study up on how Bitcoin actually works before you spout off.  You don’t even grasp the basics.  You talk as if you learned all you know by reading /r/btc.

---

Posted whilst I typed up the foregoing:


Redoubling my point:  You have evidently never heard of consensus rules and validation.  Among other things.  These are not set by the Byzantine agreement which miners produce for transaction ordering, and only that.  You really know less than nothing about Bitcoin.

(Did you just go look up “Byzantine agreement” on Wikipedia between your posts?  Though I’m curious, I only ask rhetorically; don’t bother answering.)

Now, this is an offtopic thread hijack of a good and important Lightning Network thread.  I desire to avoid that.  I will not take this as an opportunity to explain Bitcoin design fundamentals, much less argue about them with somebody who shows belligerent ignorance and an unscholarly attitude.

---

Edit:  While I was writing the above addendum, dinofelis edited and completely changed the post to which the addendum replied.  The above quote is as I first saw it.  From its much longer replacement, I wish to make one point crystal clear for the benefit of other readers who come across this thread:


No, the design of Bitcoin is not amenable to a Sybil attack; and indeed, Core developers tend to have a dim view of systems which are.  Bitcoin’s general Sybil resistance rises from the fact that there almost is no voting whatsoever in Bitcoin.  The only exceptions have been when votes of sorts have happened (or been attempted), via various “signalling” bits.  I know of instances when “XT” advocates tried to Sybil the network; it sort of wound up being a sick joke.

You could spin up a million of your own Sybil nodes on a million different IP addresses, and the integrity of the Bitcoin network would be unaffected.  (I here ignore DoS, since that is not a Sybil issue.)

I hope you understand that "Byzantine agreement" is all there is to bitcoin, deep down.  Bitcoin is nothing else but the fact that there is an established, unique, unfalsifiable Byzantine agreement over what coins were rightfully created, and how they were transacted.  There's nothing else to it.

You are perfectly right that "all else" (that is to say, nothing of importance) is done by nodes.

Those deciding on that Byzantine agreement are hence bitcoin.

That said, there IS indeed an important point that will give true power to full mining nodes, and I neglected it to some extend: the need to keep up the belief in decentralization even when in reality it is gone.    The belief in decentralization (whether the system actually IS decentralized or not doesn't matter) is the driving force in people pumping money in this thing.  If you somehow could too obviously see that that belief is bogus (while it is, but as long as it makes for a story for the gullible, one can convince), you may kill the very religion that makes the stake holders rich.   So in a way, stake holders like miners need to PRETEND to be kept in check by nodes, so that the node owners continue to BELIEVE they have power and the belief in decentralization can be propagated.

This is a bit akin to wanting your kid to keep believing in Santa Claus because that's how you can make him behave the rest of the year.  You're obliged to play Santa Claus in order to maintain his belief, and you have to give him the presents he asked Santa, even though of course it seemed that you had the full power not to: if the belief in Santa by your kid is necessary, you have to act as if Santa were real.

In a way, then, it is true that nodes do keep miners in check, because miners have to act like they do so to keep the belief up, like you have to act like Santa exists to keep your kid believe in it.   This is a game-theoretical aspect I didn't realize.  

In as much as it is true, bitcoin is now entirely open to a Sybil attack by nodes.  Because miners have to pretend being kept in check by nodes, and nodes can easily be sybilled.  In as much as miners have to keep up the appearances, they will have to do what the sybil nodes tell them to do.  Like Dad has to do what Santa tells him - or blow the illusion apart and tell the kid that Santa doesn't exist.

(that's what I called earlier: social systems based upon lies and deceit - all social systems - eventually crumble when they hit the wall of inconsistency).

1. So Lightning Networks are like few dedicated nodes with fast network?
2. Node still has to rely on those Lightning Network participants ?
3. Isn't the idea behind node connecting to nearby peer is similar ?
4. Or its more about the contract protocol than the network making it faster ?

Sounds to me just marketing gimmicks. If really want to make it faster and better reduce the avg block mining time, reduce difficulty so that more gets leverage of mining blocks faster rather than giving it to handful giant mining setups and importantly help reduce electricity and save nature.  

Edit -
To add to this, with respect to communication better change will be Voting for best nodes by their uptime/network latency, higher voted nodes gets rewards and are discoverable by some means like torrent peers. This does not require bringing whole new terminology to entire bitcoin networking paradigm.

Nodes are "vote by IP number", which is what Satoshi wanted to nullify by vote by PoW.  Third page of his paper.  There's no link between the number of IP numbers you control and your market stake.  The last is not called "vote by full node" but proof of stake (which IS a sensible way to do things).  Read the bloody Satoshi paper !

So of course, given that this is so obviously evident that "vote by IP number" cannot be considered of any importance in a system that was from the start, designed to avoid it, there must be another reason.  People very knowledgeable of that system cannot ignore the basic design principles of that system, can they ?  So there must be a deceptive reason for telling this, given that it is objectively wrong.

What is amazing in this, however, is how elementary and fundamentally wrong it is.  It denies the very design of bitcoin !

You really want to know ?

This is my way to get my own ideas cleared up, confirmed, or contradicted.  I'm highly fascinated by crypto currencies, I want to understand it.  It is a mixture of deception, lies, beliefs, good intentions, underlying brilliant ideas, propaganda and some elements of truth - as such, it is a great experiment that explores the typical social structure of human (and all complex, living) societies.  Me being something like a mixture of an anarcho-Darwinist (if that word exists), a libertarian and also a scientist, I consider that all forms of official communication are tools of oppression and deceit, and that this leads to quite functional societies for a while, until they hit their wall of inconsistency.  I think that "good" and "evil" are inverted notions and the funny thing is that this is very very well experimented in the crypto world.  I consider hence that just "official" information, the thing you get answers about, concerning crypto is of course most probably totally deceptive as it should be.  That's why you cannot really learn from asking.  

Like in cryptography, and in science, you learn from finding out yourself.  But you cannot trust yourself either.  So the best way to see whether your ideas are right, is to explain them to others, while having no "official" position at all yourself (otherwise, your "priests" will "protect your holy word" and will destroy your ability to see you were wrong).  From the reactions, you can see by yourself how well your own ideas work out: are some criticisms justified, or is one essentially repeating some dogmatic mantra ?

This is exactly how science works.  You would think that scientists publish so that others learn about your work.  No.  You would think that scientists go to conferences to tell others about their work and learn about what others do.  No.  If you're at a conference as a speaker, usually you're NOT interested so much in what others say.  You're interested in how well you can answer their nasty questions after your talk.  You can see that: at a big conference, half of the audience is working on their laptop - usually on their own presentation.  They couldn't care much about what's being said.  But they play the game: "can I find something that makes the guy in front ridiculous ?".   And your stress when you give a talk, is: "is there a guy going to find a hole in my presentation, and make me look like a fool ?".  Scientific conferences are a horrible arena where scientists try to slaughter one another.   The same game is played in writing: it is called "peer review".   This is how you learn: from nasty questions at conferences, and from rejections and nasty comments when you submit your work to peer review.  Scientists are battle-hardened by exposing their ideas and see how people try to find LOGICAL holes in it.  That's how you find out elements of truth in a "trustless" way.  Science invented trustless consensus long before Satoshi did.  That "block chain" is running for some 400 years already.

So I explain my understanding here, and I see if people can find logical holes in it.


It is not because some well-worked out propaganda scheme worked, or corresponded to what it claimed to achieve, that there is a single ounce of truth in it of course.  UASF is akin to the joke of the guy in Alaska that claims that he has invented and used a spray against elephants that works when it freezes.  When people tell him that his thing is a joke, he claims that the fact that there are no elephants in Alaska proves how well his spray worked.

As a propaganda mechanism, of course, it can work, like a self-fulfilling prophecy.  If the crowd is gullible enough to believe that if something totally harmless could do something, that something totally harmless can be used to obtain things from the crowd.  That's like Dumbo needing its magical wand in order to be able to fly.  

I'm fascinated by the entire lack of logic in the mantra about non-mining full nodes.  Nobody has ever replied something LOGICAL against the quite obvious claim that it is a totally erroneous understanding of the system, which is why I'm now absolutely convinced that this view is correct.  Of course, I can perfectly understand the official need to maintain that mantra, but I'm still amazed at the gullible nature of people, not being able to see the obvious joke.  That said, given that 3/4 or so of world population subscribes to one or other religion, one shouldn't be too surprised at the lack of critical thinking ; but then, religions are instilled in most people when they were vulnerable, that is, when they were children.  Most people here were not instilled with this mantra when they were vulnerable, so it remains an amazing phenomenon.  Of course, at first I considered I could be wrong myself.  

But there has never been given a logical argument why Satoshi, me, Gavin and others never understood the decision mechanism in bitcoin.

The remarkable thing is that one doesn't want to reason: one wants to know "in what camp you are".  One doesn't want to reason about how Jezus could have walked on water: one simply wants to know of what church you are.


This kind of argument is like the guy in Alaska with his spray, and there are several aspects to this:

- miners are in a game-theoretical battle amongst themselves too.  They didn't trust one another.  They were forced in a Keynesian beauty contest themselves.
- miners are highly sensitive to the market.  After all, the market is what pays them.  If they took erroneously for granted that the number of nodes represented a "market poll" they may have, like Dumbo, thought that they were going against the market.
- miners may be gullible idiots themselves, falling for a game of fear.
- miners may be intelligent people that don't want to demonstrate their power (and hence the true centralization of bitcoin), because they may kill the illusion of the story of a decentralized system, which is the "value proposition" of bitcoin in a way.

So the observation that miners wanted to do something, and then backed off, is not a proof that this is because user nodes have power (unless, of course, by a self-fulfilling prophecy like Dumbo's wand, if miners are gullible, or miners are afraid to show their real power and kill the religion entirely).

And the PROOF of this is so very simple, that Satoshi explained it already in his paper. If it were true that node count could influence anything, the system is entirely open to sybil attacks.  That's so obviously evident, that anyone claiming anything on the basis of "node count" must have missed Satoshi's third page and must ignore why there is proof of work in the first place.  This is so ELEMENTARY that it is mind-boggling that knowledgeable people even dare to repeat this.  

The only external effect of "Joe running a node in his basement" is the behaviour of open sockets on an IP number.  If I can control a million IP numbers, I can, for all practical purposes, be 1 million full nodes.  I don't even have to copy the block chain and RUN for real, a full node behind all these IP.  I can do this with a single centrally controlled master, and sockets opened on these IP numbers.  I am not going to check a million times the very same block of course.  My 1 million IP numbers with open sockets will SWAMP all the honest Joes with their home node.  And this for the cost that is a tiny weeny bit (namely obtaining IP numbers and a very small server behind it) as compared to mining.  That's why Satoshi reverted to mining !  That's the FUNDAMENTAL IDEA behind bitcoin.  You want me to have "true home IP addresses", so IP numbers that are only really associated by home numbers (as if that were established in the node count).  Great, how much does a botnet cost as compared to installed mining hardware ?  As I said, the botnet doesn't need to do a lot of work: the IP sockets are very light, and simply MIRROR node communication.  They don't really do all the running of a true full node, but that cannot be seen from the outside.

Thinking that those providing real proof of work are at the mercy of "vote by IP number" is not understanding even the very fundamental working principle of bitcoin.  And "run your full node to keep the miners in check" is exactly that.


Yes, yes, vote by IP number wins from vote by proof of work.  That's the "basic design principle of bitcoin", sure.

pebwindkraft, at some point, you may be wasting your time.

dinofelis is wrong, has repeatedly been wrong, is persistently and incorrigibly wrong.  In every thread he fills with long, tortuous argumentation, everything he says is shot full of holes by people who actually know what they’re talking about.  He still fills threads with long, tortuous argumentation.

Downplaying the importance of nodes is a hallmark of people pushing a certain agenda, for which dinofelis gives talking points.  (“The only reason why they talk Joes into running nodes in their basement, is because bitcoin needs a story, and decentralization sounds like a good selling argument.”)  Whereas nodes represent the economic majority on the network, and collectively have more power than miners or exchanges.  We have seen that repeatedly demonstrated within the past 6–8 months.  Indeed, if “nobody will give a shit” about nodes, then NYA/2X would have taken over the network; instead, NYA/2X was cancelled in a face-saving maneuver.  Originally, closed-doors power-player NYA had been backed and indeed, promulgated by some of the most powerful exchanges and miners!

This is in addition to, and an extension of, what you said about UASF and whence came NYA in the first instance.

dinofelis speaks of a “PoW oligarchy”, which demonstrates how little he understands about how Bitcoin works.  (Hint:  Miners have one, only one, exactly one very important job—Byzantine agreement for transaction ordering—whereas all else is done by nodes.)

Moreover, on the face of things, anybody who uses the phrase “nobody will give a shit” as a technical argument is neither competent nor serious.  Sloppy and vulgar speech on a technical topic belies sloppy and vulgar technical thinking.

Perhaps your time could be better spent.

Even if we are leaving the track of the OP here - I do not understand why you use such black and white words ("miners don't give a shit"), as if you are unsatisfied with the system. I also read your posts about the banking assumptions. Can't make up my mind yet, but looks like you are highly desperate of the system, still you contribute a lot of text to the discussions - puzzled 

I want to mention quickly the UASF discussion from last year, and why dramatic things happened. Segwit got introduced, and Bitcoin went up like a rocket... If you read the segwit2x story from a miners point of view, then one could think, they met in NY, found an agreement, and implemented it. If you take a look at the reality, you find, that suddenly many, many nodes where on the net, not relaying transactions, if the miner doesn't signal segwit support. So I think it is correct, if you say, a single user can switch on or off a node, and the miners don't care. But when many users do the same, miners HAVE TO CARE. A bit like in democratic systems, and basically the design of Satoshi design/genious/network. A single malicious code can't do harm, and in general it is better to play the incentivized game, than trying to fight it. But when many users work together, miners "give a shit" :-) I think it is very well balanced...

Here we go again.  And, if they do so reject them, they stop.  And nobody cares.  Because the 10 miner pools are connected by their own backbone network, and don't need your or anybody else's node to receive a block from one of their 9 competitors.  A full node is just a proxy, nothing more.   If your full node is switched off, nothing special happens to bitcoin.  The only nodes that need to remain up and running, are the light wallet servers (which can be run by mining pools who have their own distributed, although centralized) net of nodes of course, and the nodes of exchanges.   Whether there are still 10 000 other Joe's that run nodes or not, doesn't make the slightest difference.  If they switch off their nodes, nothing dramatic will happen, and the miner pools, nor the exchanges, will notice.

If miners decide, from block N onward, to have a change in the protocol your node doesn't like, then your node will reject block N+1.  It will wait for a "good" block N+1.  But no miner will make it.  Miners are OK with their block N+1.  So they make a block N+2 (which you also reject of course: it is not OK, and it is not on top of the last good one, N).  And they'll make N+3.  And N+4.  And N+7000.  And your node is still waiting for a good block "N+1" that is never broadcast.  One year later, while miners are working on N+52000, you're still waiting for the good N+1.  As do all your peers that ran the same software.  Nobody made an N+1-bis block for them.  So they wait.  And wait.  And wait.  Newcomers will not connect to you: your node is not up to date.   They will connect to a node that has N+52000.  And they will find it on the network of the miners.

It is a different story if miners disagree of course.  Miners can split.  Some miners may want to make N+1-bis.  And then there's a fork. This is how miners keep other miners in check.  But if miners keep agreeing amongst themselves, your node will never see a good block again.  So it doesn't keep anyone in check.  It simply stops, and waits and waits...

However, the mining pools are REALLY kept in check by the market: they have made huge investments in hardware that can only serve one purpose, they are rewarded in coins, and they don't want these coins to be worth nothing.  This is why they keep following the rules.  To please the market. Not because poor Joe has a node running, that will switch off from the moment that it doesn't like the sole block chain out there.

Exchanges have a lot of power too, because these are the gateways for miners to cash in on their coins.  Exchanges are the keepers of the market.  If exchanges decide to de-list a coin, their miners are in deep trouble.  So yes, miners will respect the nodes of exchanges.  

However, if ever exchanges and mining pools agree upon a protocol change, nobody will give a shit that 10 000 Joes find their nodes switching off because they don't find the "right" block chain any more, and come to a grinding halt.

This thing is not decentralized, but it works, because of profit motives.   The only reason why they talk Joes into running nodes in their basement, is because bitcoin needs a story, and decentralization sounds like a good selling argument.  Moreover, bitcoin being a kind of religion, running a node in your basement must be like turning those Tibetan pray mills.  But bitcoin is not decentralized at all.   The PoW oligarchy could technically impose any protocol change with a collusion of 3 or 4.  The only reason they don't, is the market (with exchanges being a big other oligarchy to take into account).


As I said, that's technically absolutely not true.  And if you're a small fish, not an exchange, nobody cares about your "checking miners".  Your node simply stops, and gets removed from the network.

TL;DR: Satoshi expressly designed bitcoin NOT to be sensitive to "vote by IP number" and hence NOT to take into account even a massive Sybil attack from non-mining nodes ; it is why he introduced PoW in the first place.  So non-mining nodes have no influence on the functioning of bitcoin.  Not as attackers, and not as "guardians".  

There's a common assumption that miners and big Bitcoin businesses want the same thing as the users, because it's supposedly irrational to want to cause any damage to Bitcoin.

This is very short sighted.

Miners and big Bitcoin businesses can:

• Represent other industries that don't like Bitcoin competing against them. They would be strongly incentivised to use miners or businesses as a proxy to damage Bitcoin
• Falsely believe that wresting more control over Bitcoin for themselves is beneficial
• Falsely believe that any damaging actions are not damaging

Users can only control the Bitcoin they want to use by enforcing it's rules on the Bitcoin network. That means some users must be running fully validating nodes, even if some users (i.e. SPV users) are depending on those running full-nodes to look after their interests.

Satoshi's original model turned out to be wrong; fraud proofs for SPV nodes still do not exist, and so real world incentives have adjusted such that a middle-tier of users running full nodes is needed to keep the interests of the system's users in balance.

It's a client-server structure only for SPV clients, while full nodes are peers of the network - even if they can't produce new blocks, they can verify all blocks and reject those that are invalid. This is a vital part of Bitcoin protocol - users can keep miners in check, making it impossible for them to cheat, except for 2 very expensive attacks (double spend and DOS). The big block model doesn't have this feature - for now fraud proofs don't even exist, so miners would be able to cheat as they please, but even if there were fraud proofs, miners and datacenter-nodes would still have so much power, that their relationship with users would be the same as that of banks - they would be able to get away with a lot of violations, because there's no mechanisms to stop them. We've already seen how easily big nodes (exchanges, services) and 90% of mining can act as one entity when they have tried to attack the network with their fork in the last year - this is a great example that we can't trust anyone, and instead we should verify everything.

I realize that you are right, and I was wrong here.  There is indeed a solution I didn't realize.  If A, B, C, D and E are linked in a ring, of course, by the direct links, A can pay B only until the AB channel is exhausted.  But I forgot that A can also pay B by sending through the AE, the ED, the DC, and finally the CB channel.  I stand corrected. 
So, half of the time, A pays B though the direct channel AB, and when that reaches its limits, A pays B "the other way around". And then again in the direct sense.
Thank you.

You provide no reasoning at all for why you, the bakery, the supplier and the farmer cannot simply pay each other with direct links. Well, except when you say that they need "flow in the same direction", which is precisely what having direct channels between one another achieves. This is a reason, but it disproves your point completely. There are actually people reading your posts sometimes, y'know!

Routing can be achieved by hubs, as you point out, but it can also be achieved other ways (which you suggest isn't desirable or possible). The protocol design doesn't care whether a node uses nodes with high numbers of channels (i.e. hubs), or uses a lot of multi-node hops, or opens a new channel. It only cares about minimising fees (which is in the user's interest).

Your description implies channels can never be closed, they can. Just one on-chain transaction can be used to close 1 channel and open a new channel at exactly the same time, and the new channel can be sent in the direction where no route existed before (i.e to where the money is needed). It's quite simple. And if it's cheaper, or more desirable for privacy reasons, than using a pre-existing route ("hub" or otherwise), that's exactly what users will do.

---

You're also repeating the use of the word "banks" to describe nodes with the highest number of channels (you've been doing this in other threads too). This is inaccurate. How many channels do I need to open to become a Lightning "bank"? There is no way of distinguishing between Lightning "banks" and regular nodes based on how many channels they have, everyone will be opening and closing channels as they please (including their "bank accounts" with nodes that have alot of channels, unlike when dealing with actual banks where one must ask permission to close one's account).

There is no entity in the Lightning network's model analogous to a bank, in any way, shape or form.

You could do, but then you would be completely missing the point.

Well, usually there's a money flow and an opposite goods/services flow, and these flows tend to be conservative, that is, flow in circles ("div B = 0" if you are mathematically inclined).  

Usually, money flows from my employer to me, from me to the bakery, from the bakery to his supplier, from the supplier to the farmer, from the farmer to the store, from the store to my employer, to make it simple.  I rarely pay my employer, my bakery rarely pays me, the supplier rarely pays the bakery, the farmer rarely pays the supplier and the store rarely pays the store.  Look at your bank account: you usually receive money from certain entities, and you usually spend money to other entities.  There's rarely a back-and-forth motion.  With all these entities, there's no point of opening channels.  They will quickly get exhausted.  It is more interesting for all of these entities to connect to a "single settlement node".

If we want to do this with a LN network, the only solution is to have a central node (a bank).  Otherwise, channels will always get exhausted, because they always need to flow in the same direction.  Yes, you can replace that central node by a "web of central nodes".  But my employer, me, the bakery, the supplier, the farmer and the store have no reason to have links between them, only to that "central web".  And inside that central web, if similar flows have to flow in systematic circuits, they all have also a reason not to connect along the periphery but only to a more central part.  And so on.

The ideal LN structure is hence a single central hub, with all others connected to it.  It is the only structure that can support indefinitely all possible "circles of money flow".  That's not even a criticism.

This analogy to real money can be discussed as an interpretation, of what money is. Friedman, Nash, Schumpeter, Mises, or the gold standard...? The usual bills and coins (Pound, Dollar, Euros?), or the virtual money (FIAT), which only exists as figures, where you don't know if you can get it, if you really need it... you name it!

I don't see how money goes in circles, my view is it is a one-to-one connection. I open a 1:1 channel, when I pay with cash, I put a 10 Euro bill, and I get maybe some return. There is no circle at all involved. Same I can do with a Lightning Channel...
Circle might come into the game, when my 10 Euro bill, which is in the hands of some else now, continues his travel, cause this someone else pays again for something. So the re-usability of this bill is maybe the circle analogy. The exactly same thing can be done with Lightning. But this money has a major, major disadvantage: you open/close the transaction channel within several seconds/minutes, you cannot keep it open. Every time you pay something, you need to re-open a process like pulling out your wallet, find the right bills, and hand over the money. Lightning gives me the extreme flexibility of "lending money" to a trusted channel, from which a specific amount can be used over and over again. The assumption is, this makes it very cheap for especially smaller amount transactions.

The best thing with lightning is on top: in real live I cannot tell my grocery shop to keep some of my money, because I have to buy some bread around the corner, and they shall send it to the bakery. With Lightning I can do this. I'd call this hopping over service providers. Impossible in this money world (limiting the circle analogy).

With Lightning everyone can do this, and as such a network gets created. If the perfect game for lightning is, that 10.000 traders open channels with the exchanges, then I don't see how this can be called centralized. And another 10.000 with their cable service. And another 10.000 with Amazon (or what so ever). I don't have to use these existing channels, I can still open a channel with my spouse independently.

People here are discussing a lot centralization, predicting the future based on very limited assumptions. I am not seeing the centralization idea, as it is especially unclear, what centralization means. If there are 20 main hubs around the world, can this be called centralized? Or is centralized, when 1000 major hubs collect 10.000 users around them? Or is something centralized, when one company owns all these 20 worldwide hubs? Or is centralized, when you have to use a specific hub? There is a broad spectrum of personal understanding behind centralization, that people use to predict the future...  I always come to think, that a lot of personal fear is involved, and overall I am missing real tests from those people. It is easy nowadays to create several thousand or ten thousand nodes in lightning, and let them work. And then have real statistical values as an underlying basis for such statements. But I also see, that not everyone can create such a setup.

Summary: centralization of lightning cannot be predicted, nor proofed, due to:
 - incapability of large scale measurements
 - limited assumptions
 - limited personal understanding of the network
 - different scope of meaning (of centralization)
 - and a bit of personal apprehension

All this make predictions look like hobby research, throwing shells, maybe conspiracy theory, but not at all (bullet proof) research.

One can also think that this kind of image is simply rhetoric.  Association of "the product to sell" with other things that have a positive image is the basis of publicity.

All analogies break down if you try and push them further than they were intended. The point of an analogy is to use a similarity in something people are already familiar with to explain something new. Pointing out differences doesn't make an analogy invalid.

In this case the analogy is of blockchain being early Ethernet segments, a bus architecture with CSMA/CD being suitable for small LANs and then Lightning Network being the addition of routers and the TCP/IP stack at the next level of the OSI 7 layer model. In both cases they remove a restriction on bandwidth and allow you build an infinite network.

The analogy works in getting across the point it is intended to. LN is an additional layer on top of Blockchain.

Of course Bitcoin isn't directly comparable to a network, and of course, LN isn't directly comparable to a routing protocol. It's an analogy, not a comparison.