We'll see if this doesn't happen to bitcoin too...

My idea is, the more, the merrier. 

No, without this, bitcoin would have its currency value, given by Fisher's formula.  We both agree that this would be ridiculously small as compared to the actual market cap.  If a *currency* is used (you know, to buy and sell goods), there is a demand for it, and an average hold time, between the moment where you obtain the money when you sell stuff of value, and the moment where you spend it to obtain stuff of value.  This demand gives rise to the normal price of a currency.  That should normally be bitcoin's price.

All the rest is greater-fool.  And, as we can see, this aspect of it is HUGE.

In niche applications where fiat cannot go.

Huh, that's about the price of a bitcon transaction now these days.

I agree with you, only: crypto currencies (at least, the block chain based crypto currencies we know of today) start from an axiom, which is decentralized and trustless.  Nice as this may seem, it brings a huge burden over centralized and trusted payment systems (fiat).  This is like driving a tank, instead of a small and lean city car.  The tank will resist enemy fire, and it will be able to run over non-paved roads, it will be able to get through barriers and other types of limits ; maybe even through a mine field. 

However, to go shopping, to go and take the kids to school, to go on a holiday, to go to work, .... using a tank is a HUGE BURDEN.  It is extremely expensive to use, to keep running, to repair, it is clumsy, ....  Yes, you are very safe in a tank.  Yes, you will protect your kids against a terrorist attack (or an outright war) when you go and get them at school.   But by far, the small city car is way, way, way more practical, because it doesn't waste resources, energy and all that on mitigating a problem that is not the principal difficulty.

So tanks are not competitive with city cars.  Unless we are in a situation where the tanks' use and safety is of paramount importance.

And that's the problem of crypto and "mainstream", and "adoption".  It is simply not competitive with fiat in all those cases where fiat can be used, because fiat doesn't have the burden of having to be decentralized and trustless.

But there are a few niche applications where crypto is useful.  Like there are situations where tanks are useful.  But don't expect most people to drive with tanks.  They will use city cars.  It is a better use of resources, more performing and all that. 

When you can do it with a city car, don't use a tank.  When you can do it with fiat, don't use crypto.

So what remains ?  Niche activities, and greater fool games.  Because that's what crypto, in its block chain form of decentralized and trustless at a high cost, can only be about. 

hard is not "nodes vote".  Hard is: miners DECIDE to fork (to make two chains). 
But it is only when there is a permanent double coin on exchanges, that the USERS can vote in the market.

If miners decide to hard fork and are united in that aspect, they are just as well sole deciders as with a soft fork.
The only difference is that with a soft fork, there doesn't need to be unity amongst miners to fork: a MAJORITY of miners can impose their choice on the minority of miners that do not agree.

With a hard fork, the minority of miners that do not agree, can continue to build the old chain, making two coins, on which the USERS will vote in the end.

But as long as miners only make one block chain, nodes nor users can vote.  Well, users can only vote by leaving that crypto currency.

(nodes can never vote: it is not vote by node, but vote by proof of work).

The only thing that this "node" stuff does, is to try to fathom what will be the market response ; or as a means of propaganda. 

Most probably, the safest method is to use an exchange that lists both coins, and has provisions to do the split for you, that is, to send the "unified old coins" to an address attached to you on the exchange.  The exchange will then give you IOU of both types of coins.  You can trade those (for instance, selling off your bitcoinx and buy more bitcoin, or selling off bitcoin to get more bitcoinx).  When you will withdraw the coins, they will be 1 type only.

But be sure you use an exchange that knows how to handle this.  Most big exchanges learned this during the ETH/ETC split.  Some made big mistakes.  Some stole the ETC of their customers, pretending they didn't "list" ETC.

With ethereum, there was a smart contract that could do the split for you, but with bitcoin, that's not the case, and the only way to do so is to use "newly mined dust", which is hard to come by for a normal user.

Miners have all advantage in keeping the number of transactions a scarce resource, as every product with inelastic offer for which there is large demand can rise very high in price.    So, miners don't really like scaling solutions.  The (idiotic) fact that block sizes have been limited to very small sizes (1MB), hard coded in the protocol, is an unexpected but lucrative benefit to them, especially when block rewards fall.
On the other hand, miners don't want the market value of the coin they are mining, to be affected.  They are probably potentially ready to scale the amount of transactions if otherwise the coin becomes worthless.  But I think that this is actually not a driving force.  For the moment, block rewards are still the principal revenue, but fees are making up about 20% of the miner revenue.  As such, a scaling solution that would bring down the tension on the fee market would need to rise the bitcoin price by more than 20% for this to be attractive to miners. 

Chances are that fees will only increase over the coming months.  Suppose that fees represent similar amounts of income than block rewards.  Miners would only be favourable to scaling, if they were convinced that bitcoin's market value would DOUBLE because of the scaling.  This is not the case, because bitcoin's price is NOT mainly determined by small on-chain transactions, but rather by off-chain (exchange) trading.  So I can't see miners ever be favourable to any scaling solution.

I think miners "support" BU mainly to counter segwit.  Their idea is most rationally NOT to provide a scaling solution, unless they are convinced that bitcoin's market price suffers *significantly* from that non-scaling, which, as markets show us, is absolutely not the case.

Ah, yes, sure.   Then we agree.  But then you can forget about "decentralized" in any case.  And then, in any case, your worries will be of a different type than caring for a crypto currency. 

What happens if the global electricity infrastructure deteriorates so much that mining by PoW will not be possible any more ?

I would say that if global internet infrastructure deteriorates so much, that we probably have MUCH MUCH bigger problems to handle than caring about a crypto currency on the internet.  Maybe, finding a gun and defend ourselves, and trying to find food or something.  Mad Max style things.

Crypto currencies are a side product of high network connectivity, cheap computing and sufficient electricity.

The problem with block chain based crypto currencies is that they start from some most probably erroneous principles.  Full trustlessness is very, very hard to implement, and the price to pay is way, way too high for a real-world situation where partial trust is often a good investment.   Block chain based crypto currencies, writing all transactions on a global ledger, are extremely wasteful on computing, storing, and network resources ; but as these resources grow over time, and as adoption of these crypto currencies is NOT THAT IMPORTANT, there is a natural throttling of the adoption by the technological cost.  What is strange, is to limit this resource artificially with block limits, instead of having the resource price be a natural limit.

Block chain based crypto currencies will never equal fiat style payment systems, because the investment in trust in these fiat payment systems allows for HUGE advantages in terms of capacity.  The price that block chain based crypto currencies pay, is that they try to adhere to full trustlessness, which is too high a price to pay to compete with trusted systems. 

This is like driving to your work in a fully blinded tank, with all reserve food to be able to sustain all kinds of attacks: for sure, your system is more "secure", but the resources wasted on that are so big, that you are not competitive as compared to the guy taking his bike.  Your costs in terms of maintenance, consumption, capital cost etc for your tank are so big, that you waste all the money you win in going to work in the first place.   The guy on the bike takes the risk of being shot down, true, but at least, he's way, way more competitive in earning money than you are. 

Block chain based crypto currencies are over-protected by trustlessness, and hence, are not competitive with fiat systems.

Of course, you can transform block chain based crypto into less trustless "banking systems" on top of them.  That is more or less what the LN tries to do.  You will get a network of LN banks (big hubs where you have "an account", that is to say, a small-user payment channel).

Yes, you can build faster and lighter structures on top of bitcoin.  By giving up trustlessness.  And building a fiat system.

My idea is that the only use of crypto currencies are those niches where fiat doesn't work.  Those niches are not that important, and advances in technology will push the limits of what crypto currencies can handle to a level where most of those niches are addressed.  In other words, the relative scarcity of adoption that crypto currencies allow will be taken by the most useful niches that need it.

And all the rest will be done by fiat.  I don't see the use of building another fiat system on top of bitcoin: we already have that.  Let crypto currencies exist for those niches where you can't do it with fiat.

And then, one day we will invent better systems than block chains.  That day may not be far away. 

I would like to add something, that illustrates even further the "futility" of the distributedness of the non-mining node network.  You can compare the block chain to a kind of very, very heavy digital signature, not of a single transaction, but of a whole crypto currency.  Now, if you can verify for yourself the validity of a "digital signature", the only thing that matters to you is that you got it.  From whom you got it doesn't matter because the security is in the cryptography, and not in the nature of the one providing you with the signature.  The cryptography proves that the signature is valid.  The validity of a digital signature doesn't increase by "other independent entities telling you that they too, agree with the signature's validity, or not".  If it depended on peers validating, it would mean that the cryptography of the signature is broken.   Normally, the verification of a signature doesn't depend on others: you can verify that all by yourself.

So, if you get to a block chain, which is a kind of big digital signature of the whole of what the crypto currency represents, and you verify it for yourself, and you agree with it, that is all that matters.  Even if you got it from a "centralized server".  You do not need peers to acknowledge that it was valid: if they acknowledge it, you STILL want to obtain a copy to verify it for yourself, and if they don't acknowledge it, they won't even transmit it to you.  So, your peers validating it or not, doesn't contribute anything to YOU finding it valid or not.

The only point of possible contention with a "centralized block chain server" is that that server might hold back the "best block chain", to serve you a "valid, but less long" chain.  However, in as much as you see that chain growing (for yourself), you know that who-ever is doing that, is spending a lot of hash rate on a chain of which HE knows that it is not the longest and strongest, because the cryptographic proof of work can be verified by YOU.  No miner in his right mind is going to WASTE all that proof of work on a chain of which he knows HIMSELF that it is not the longest.  So there is actually no danger of being "tricked into a miner making shorter chains to cheat on you".
The miner's "web site" has all the reasons to provide you with the most up to date, longest chain he's REALLY working on.  There is no danger to be tricked there.

In fact, this has been, in my opinion, the biggest mistake in most crypto currencies: that the consensus protocol enforcement is rewarded, be it proof of work or proof of stake or something else.  This induces centralization by the nature of economies of scale and eagerness for rewards.

Well, that only depends on what the owner of that node wants.  

1) it is "just a node": doesn't matter, but it will eventually find other nodes on the network that have valid (according to his protocol) block chains.... or not, in which case it will simply come to a halt.

2) it is a user, that is to say, someone who has coins and want to send transactions against value, or the counter party: who wants to receive coins, and provide value for it (goods, services, IOU, dollars.... whatever).  Of course that user will have to CHOOSE what coin (what block chain) he wants to use (in agreement with his counter party of course).  And so, that user has to pick one of the miners making the "right" block chain.  Miners have all interest advertising themselves to the users that want to value the chain they are building.  So normally, both having all interest to get in contact, I don't see the problem them finding themselves.  If it is a big user, say, an exchange, it is the small miner (mining pool) that will find the big user advertising himself ; if it is a small user, and a big mining pool, the mining pool will advertise himself.

3) miners have also all interest in connecting to users, but especially to like-minded miners.   So miners join automatically the strong network of like-minded miner nodes.  They have to, if they want to use their hash rate usefully and not always be late and make orphaned blocks.

Users have no difficulty finding facebook's servers, don't they ?  Users have no difficulty finding coinbase's servers, don't they ?  So I don't see how they could miss one another, if they are both eager on doing business together.

Miners find one another.  Users find one another.  Because they will advertise themselves.  This is not really "centralisation", in the sense that what is decentralized in a crypto currency, are those entities providing for the consensus.  In a proof-of-work system, those are the miners.  And nobody else.  But they are sensitive to the users as they provide market valuation, and pay the miners through their market valuation.

Pools are miners and users.
Exchanges are (big) users: they transact a lot of coins.
Users, well, are users.
And users that set up nodes, well, are users.

All these entities are USERS, that is to say, are interested in transacting tokens against value.

You might think that exchanges are something special.  They aren't really, they are very big users.  They receive coins, and they give you web-site IOU with which you can play on their web site.  They also sell web-site IOU for (fiat) money.  That's their ultimate goal.  And they send you coins to destroy web-site IOU.  In other words, they transact a lot with their customers: essentially they transact coins ultimately for fiat money.

Exchanges are amongst the most important users, because (unfortunately) not much ELSE of value is exchanged for crypto coins.  A little bit of services and goods, but not much.  Far most (real) crypto currency transactions is to and from exchanges.  Coinmarketcap also includes the transactions between exchange-web-site-IOU which are promised to be backed by the actual crypto currency, but these transactions are not the ones on block chains we are talking about: these transactions could take place totally independent of even an existing block chain.  Hundreds of alt coins that essentially only exist on Poloniex could continue to be traded as Poloniex IOU even if their block chains stopped entirely.  It would be like Play Station game assets, centralized on Poloniex ; not a crypto currency as such.

So, all the entities you talk about are users.  Users vote with their money, because they establish the value of the crypto token.


This "infrastructure" is not really needed, and has strictly ZERO decision power, that's what I'm trying to tell you here, and this is why the network of non-mining nodes is absolutely UNIMPORTANT, and doesn't mean zilch as "the controlling community" ; "the guardians of the consensus" or whatever things are told about it.  This infrastructure is just a proxy structure that helps (only helps) the miner network communicate with the users: miners make a block chain for the user (and the MINERS decide upon the protocol, by making the blocks, and by choosing on which other block to build), and by accepting transactions submitted by users.

The only thing that a node network can do, is make this propagation happen in a P2P way.  But as long as there is a direct connection possible between the miner network and the user, that P2P network is not even needed.  That's what I'm getting at.  And if the network is not needed, it has zero decision power. 

The USER has of course the possibility of choosing between SEVERAL different versions of a block chain *if there are miners making them*.  If MINERS decide upon a hard fork, that is, their network splits into two networks, building two different chains, then the users (including big exchanges) can decide to only accept one ; or both.  And they will then determine the respective market caps.  Users can also decide to quit the network, and to plummet the value of the coin, even if only one chain is out there. 

This is why miners are sensitive to the user value estimation.


Miners have power, because they make the block chain or the block chains, so they (and they alone) decide upon the protocol or protocols used to make the block chain (s).

Users have power because they give value to the crypto tokens on the chain(s). 

So, miners have power over users because the miners determine the protocol ; users have power over miners because they determine the value of the tokens the miners obtain (in other words, the users pay the miners).

And there is nothing else in a crypto currency.

As I said, suppose that 3 big exchanges, and 7 mining pools (having a significant part of hash rate) agree upon a protocol.  That's all that is needed.  The exchanges list the coin (and their customers play with it), ; the mining pools make the block chain and include the transactions.

Whether 4000 nodes agree or not, they don't give a shit.

You see, the problem is that you are introducing ill defined concepts about what "consensus" should be, like a political or philosophical goal or something ; "protecting the coins".

But there are only two kinds of agents in a crypto currency.  There are the entities that use it.  Whether they are exchanges, merchants, speculators, ...., they are people that want to use a monetary function of bitcoin, and *transact* (buy them, sell them, exchange them.... whatever) against value.

And then there are the miners: people building block chains.  Miners are also users, because they do the mining to obtain coins, and sell them, or hold them and sell them later against something.  But most users are not miners (contrary to what Satoshi had in mind).

These two sets of entities are what makes a crypto currency.  The users transact against value, the miners make the block chain.  The users need the miners to make the block chain, and put their transaction on it.  The miners need the users to spend their fees and block rewards on.

And, essentially, that's it.

The entire crypto currency dynamics is a matter of the interaction of these two sets of entities: amongst themselves, and between one group and another.

Users need the block chain in order to "prove" their transactions to their counter party.  Without your transaction on the block chain, you didn't prove your transaction and the counter party will not provide the market value.  Users also need the block chain to prove their right to spend (if they are on the receiving side).  And that's all they need.  Users are entities, people, that want to transact tokens against value.  That's why they are there.  And they need a block chain to be able to prove their right to spend, and to prove their transactions.  This is the only thing that gives "market value" to a token.

Miners make block chains for the users, and obtain tokens as block rewards and fees, that they also sell to users, to obtain the market value.

As such, miners need to communicate with users to get their transactions, and users need to communicate with miners, to get the block chain.  This is all communication that is needed.

The phenomenon of consensus makes that users as well as miners agree upon a protocol of using the coins (of what are valid transactions) and of making the valid block chain.  the consensus mechanism involves proof of work and a whole set of rules that tell people 1) when A block chain is valid and 2) when there are several valid block chains, which is the preferred one.  The proof of work mechanism is such, that it costs a lot to make an alternative block chain, so if a miner makes a block chain, he has put a lot of investment in it.  He would lose a lot of money if he were not trying to build blocks on the "best block chain".  So you can accept that an active miner is always building on the best block chain around according to him.

So, as a miner, what do you need ?  You need to have the best block chain, and quickly, to build blocks upon.  You cannot afford learning late about the best block chain: you waste all your hash rate.  A miner can only be successful if:
1) he learns quickly about the latest blocks on the best block chain
2) he divulges very quickly his OWN found blocks to other miners
3) he gets a good filled mem pool of user transactions

This is why a miner is fully motivated to have good network infrastructure, announce his own best block chain to everybody (and especially to other miners) and takes in user transactions.

A miner is the best source of block chain, and the ultimate destiny of transactions.  There is no better source of block chain around.  It is the *principal server* of the best block chain.

Of course, a miner also has all reasons to VERIFY the integrity of the blocks of his peer miners.  If he mines upon INVALID blocks by peer miners, he is just wasting his hash rate.  So miners have in general a strong incentive to adhere to the protocol: there is a strong mechanism of consensus amongst miners.

Now, what can users do ?  If there is full consensus amongst miners, NOTHING.   Users can just accept the single block chain that miners in full consensus, make.  Users can download the block chain from miners, and like it, or not like it, but as there is no other block chain, they can't do bloody nothing, except not use that crypto any more.  

So what do non-mining nodes mean in this respect ?  Simply: nothing.  In as much as the full non mining nodes on the network are in agreement with whatever protocol the miners use amongst themselves, they are just proxy servers of the main servers which are the nodes of the miners (the pool nodes in most cases) ; and in as much as they are not in agreement, they simply come to a halt.   Because the miners, amongst themselves, make only one single block chain.

==> even if tomorrow, all miners would collude on a different protocol, say, they all agree upon not halving the block reward next time, all full nodes will come to a halt.  The only thing that users can do, is to use a node that is in agreement with the new protocol of the miner consensus.   And the only reason why miners might NOT do this, is that amongst themselves, there is a consensus mechanism at work, that would orphan blocks AMONGST MINERS.  The first miner making a block with a non-regular block reward would get his block orphaned by *other miners*, unless they collude.  And IF they collude, and implement this hard fork, and if this is the only block chain that is made, whether there are 10, 100 or 4000 non-mining nodes "protesting", there's nothing that they can do about it.

==> as long as miners are in consensus, non-mining nodes are at most proxy servers.

It becomes more interesting if there is dissidence amongst miners: if different miners use different protocols, and build TWO (or more) DIFFERENT block chains.   Then, non-mining nodes can chose between the chains.   But in the end, users will now have the choice between two chains.  The miner community will have split in two communities, building two chains.   The user can now chose WHICH block chain he is going to use.  But nothing changed in the relationship between users, and THEIR preferred miner set, and their preferred block chain.  Users will STILL communicate with the miners that correspond to their choice of block chain.  

The set of full non-mining nodes will split into two different types: those that download and transmit chain A, and those that download and transmit chain B.  But they are STILL just proxy servers of one or other block chain.  

==> even in the case of miner dissidence, and the building of multiple block chains, full non mining nodes are still proxy servers of one of these block chains.

However, now, users can appreciate the different tokens on the different chains differently concerning their market value.  Miners are sensitive to this, because miners earn tokens on the chain they are mining.  So miners will take into account the user market value appreciation when they mine different block chains (different coins).   This is the ultimate interaction by which users impose their will onto miners: by market value.

But in all of this, non mining nodes are just server proxies to miner nodes.

Let's do a small calculation.  Take a medium-sized country in Europe, say, France.  About 60 million inhabitants.  If these 60 million people paid their taxes with bitcoin, then that would mean: 60 million transactions.  At a rate of 3 transactions per second, that would mean, 20 million seconds.  231 days of full blocks.  By the time your taxes are paid, using the FULL BLOCK CHAIN, you can almost start over again.

France only.

Go figure if the US citizens and the Russians started paying their taxes in bitcoin too.

I would rather argue for infinity minus one

Nope.  Nodes don't mean anything. The interaction between cryptographic proof of work on one hand, and users spending money in the market, is the boss.   Nodes are means of communication between miners and users.  They could do it by e-mail if really needed.  A network of other-minded nodes doesn't matter.  As a user (that is, as someone wanting to spend value on bitcoin, or transfer bitcoin for value), you only need to have the block chain that your counter party is also agreeing upon using.  You don't need a network of other nodes for that.  You only need your wallet, the counterparty needs his wallet, and both of you needs access to the common block chain that you've both agreed upon using, and which is produced by the miners sticking to the protocol you want.

Whether there are 4000 other nodes, preferring another block chain, or whether there are 100 000 such nodes, doesn't matter for you to have access to your preferred block chain, and for your counter party to have access to its block chain with a wallet.  Oh, yes, and you have to be able to send your transaction to a miner to include it in the block chain, because that's what your counter party is waiting for.

In this essential scheme of a crypto currency, you don't need "most of the other nodes".  So their preferences don't matter.

What matters is the miners making block chains, and the users spending value on coins when transacting, and a means of communication for the miners to give their block chain to the user, and a way for the user to give their transaction to the miners.  By e-mail if necessary

Proof.  Suppose you adhere to bitcoin as it is now and so do I.  Suppose that on one hand, there are miners still working on bitcoin as it is now, and that you and I adhere to it, too.  Now, suppose that 95% of the nodes are "against" it, and implement, say, segwit, but you, I, and those miners don't like that, and continue the old chain.

Suppose that I cannot get a connection to the miners.  But suppose that I make my transaction with my wallet on my off-line computer, and I send that transaction to a miner by e-mail.   Suppose that 10 hours later, I download by FTP, 60 blocks that follow up on the last block I have on my off-line PC, and that my transaction had 59 confirmations in that piece of chain.  I tell you to download that too.  You now have my transaction ; you are pretty sure that the miner has been mining "for real".  You check it with your node software that all blocks are valid.

You have accepted my payment. 

We didn't need the node network.  Whatever most nodes do, whatever they censor, whatever they do, you and I don't need it.  We only needed the miners, and the communication with the miners.

I would go for 10 000 coins or more.

But the notion of "coin" is maybe different in my mind: it means something like "block chain", that is: suite of recorded transactions.  I would count each payment channel on a lightning network as a coin in that respect.  You start a coin by opening a channel and giving it value with a bitcoin collateral.  You end up destroying the coin, and releasing the collateral.

There's a huge difference between DASH instant pay and LN.   In the LN, transactions are NOT written to the block chain.  The block chain only serves to hold the "collateral" of the LN nodes.  Bitcoin would be a reserve currency, and the LN nodes would be banks that transact your payment over the network.  LN is a banking layer over bitcoin, that ressembles very, very much, traditional banking, except that they have a smart contract regulation with bitcoin as reserve currency, rather than having a legal regulation. 
What goes on the block chain, are only settlements between banks after they have processed many transactions (like now with real banks).

On the other hand, DASH instant pay is only a confirmation that your transaction is waiting to be put on the block chain.  That's it.

You could do this with bitcoin too, except that with bitcoin, nothing stops a user to send a double spend with higher fee that will be preferred by miners.  With DASH, you trust master nodes that they will not send a second confirmation for the higher-fee double spend.  Probably you can trust master nodes somewhat more than miners in that respect.