No, because they don't need it.  Just any code that counters segwit is good enough.  They will never activate it.
That said, measuring people's capacity to modify a code by the number of Twitter followers is doubtful. 

What I mean is that if bitcoin hard forks, you make two *different* crypto coins.  If you think that you "double the coins" then you should take into account all the coins on other chains too, all altcoins.

And that's indeed fundamental.  The idea of "sound money" only makes sense if that money has the monopoly on money.  The idea of sound crypto money only makes sense if there is one crypto.  But if the market is shared with others, then "sound money on one" doesn't make sense.  Yes, if there is BTC, and BTC-U, there are 21 million BTC, and 21 million BTC-U.  But there are also 7 million DASH, 90 million ETH, 14 million XMR, and so on.  If you are worried about the doubling of bitcoin, you should also consider all other crypto. 
Bitcoin is not fundamentally different.  It is a crypto like any other.  It just was there first, and that gives the bitcoin brand name a huge advantage.

As I said, soft forks are not protocol changes, but protocol restrictions.  
There is BIP 66, BIP 65, BIP 34, BIP 30 and BIP 16

BIP 34 was about the version number of blocks and from 2012 (still the cheap years of bitcoin)

BIP 65 introduced a new instruction in the bitcoin instruction OPcodes

BIP 66 solved an ambiguity concerning OPENSSL

All these things are just technical aspects, soft forks, and do not affect in the slightest bit the usual way of how people use bitcoin economically.  I even don't think that these aspects are much used.  All of them have only been activated with 95% consensus which was easily established.  You can hardly say that they modify the original protocol in its way of working.

These are technical improvements that don't affect what people actually do with bitcoin, they don't affect the economics of fees, rewards, and so on.

In other words, they were not "significant changes".  Most people would never notice.

I think you're being very naive about LN, much more so that Satoshi's naivity about proof of work.  The LN only makes sense when there are VERY BIG hubs to which normal users connect.  The reason is that in order to be able to pay a random person over the LN you need to find a series of nodes that can propagate your payment to destiny before settling on that path.  Settling costs a fee.  In order for the LN to be competitive with direct on chain transactions, a LN channel needs on average to transact AT LEAST a number of transactions, equal to the number of hops between sender and receiver.  In other words, if on average, a transaction takes 20 hops between a sender and a receiver, that means that on average, each LN channel must transact 20 transactions before settling, or direct on chain transactions would be cheaper.

This means that you have to "lock in" on average 20 times the amount of a single transaction in each of your LN channels to your neighbours EVEN TO BREAK EVEN, if you charge the same fee to your customers, than an on chain fee.
The more you can keep your LN channels open before settling, the better your margin becomes (between the fees you can charge for a LN transaction, and the block chain fee).  But you are not alone in settling: your counterparty can settle too !

The economies of scale in this system are evident: the most profitable LN nodes, that can offer the most competitive LN fees, and can guarantee the best probability to reach destination, are BIG HUBS that connect amongst them, with very large amounts locked up in their channels.  Small players are CUSTOMERS, that have to open a channel to their BANK/hub and keep it open to process at least say, 20 transactions before closing and having to pay an on-chain fee.

The economies of scale in the LN are even much more pronounced than the economies of scale in proof of work.  So the centralization will go even faster.  Bitcoin is now in the hands of 5 people (the 5 mining pools that make up more than 51%), and for sure in the hands of 14 people (controlling more than 90% of the hash rate).

Bitcoin is hence a large majority 14 people oligarchy.  But it took several years for this centralisation to take place.   The LN will do this much faster.  Probably 10-15 "banks" of hubs will form rather quickly if the LN is to function.

I'm not proposing to bail out anyone   I'm just saying that whining about a crypto because it doesn't have the right features and "should evolve" is not the right attitude: changing to a better crypto is.  The only thing where this hurts, is in the monetary belief in eternity, but things are never eternal. 


I think my principle is right.  Not "morally right" but "factually right".   And yes, I don't think that bitcoin will change unless it is totally centralized and that central authority (a cartel of 5 miners for instance) decides upon whatever should happen to bitcoin.

My conviction is that as long as bitcoin is decentralized, the protocol is protected by the dynamics of immutability.  And all the animosity we see is nothing else but that dynamics at work: the fact of sufficiently diverse antagonists (decentralisation) not being able to agree over anything but status quo, because all change can be done in 10 different ways, and each antagonist has HIS way.  So I think that we are now stuck with the bitcoin protocol as it is, and 1 MB blocks.  For ever.  Unless a cartel of 5 miners (and maybe a Chinese government official) in a room somewhere decide otherwise.


Apart from the very early years, I would doubt that.  I don't think that there has ever been a hard fork in bitcoin.  Soft forks are protocol restrictions, so in as much as 51% of PoW decides, imposing soft forks is not really a protocol modification ; going back however, is.  Can you give many examples of protocol upgrades ?  The block format is the same, the block header is the same, the signature scheme is the same, the bitcoin instruction set is the same as far as I know, .... I really wonder what has changed in the bitcoin protocol, that is, in the way to write valid block chains ?  (apart the very first beginnings when bitcoin was worth zilch and no mechanisms were locked in by profit and antagonism yet)

(the network protocol is much less important as it is volatile: what remains is the block chain).

You know of many hard forks of bitcoin in the past ?


It is my way to build my own understanding.

This is a typical application where banks could use bitcoin as unregulated reserve currency for their unregulated agencies in such countries.

Not a bug.  A bug is a thing about software.  I'm talking about a protocol error.  A mathematics or cryptography error, an error in the economical dynamics.  Software can change, software only IMPLEMENTS (correctly or not) the protocol.  But the protocol IS the coin. 

For instance, it is part of the protocol of bitcoin that just ANY UTXO from any block can ALWAYS be used as an input to a transaction, using your private key in a way that was written down long ago.  If you leave your private key to your kids, and they leave it to their grand children, .... and 500 years from now, that private key should still be usable if that coin still exists.  If that's not the case, then that's not bitcoin. 

Now, if the cryptography breaks down and that private key can be deduced by cracking 2 hash functions and elliptic key signatures, then bitcoin is simply done.  There's no point in trying to fix that, if by fixing that, you break the original protocol.

Now, imagine that there was software that had bugs in the crypto software and accepted erroneous signatures.  That is simply a bug.  It means that maybe 1000 blocks have been generated that were INVALID according to the protocol, even though the buggy software accepted it.

The right thing to do is to orphan these 1000 blocks, because they are simply not valid according to the protocol, and were ERRONEOUSLY considered valid blocks by faulty software, and to start mining again from the point where the last valid block (according to the written protocol) was made.

If not, then whatever thing is being used, is not bitcoin.  Because bitcoin was an immutable protocol. (not unmodifiable software, that *implements* a protocol).  A protocol is a white paper with all details of it.  Not a piece of software.

I read your post.  The point is, that there is no reason to "repair" a coin.  A coin is meant to be what it is, according to its protocol (not its *software*).  If that protocol fails in one fundamental way or another, one should move to another coin (that is, another protocol).

Not yet, but soon.  China is not going to fight bitcoin, it is going to use bitcoin dominance for its own purposes.

You got it.  So a coin should remain itself, and when it is outdated, people should switch to other coins.  Simple as that.  The coin remaining itself can maybe still maintain a niche application.  And most probably what will happen in the long run.

But that is entirely wrong !
The "bad guy" can most of the time not stake directly (it is not his turn to stake).  So this "third block only on the second chain by the bad buy" can come maybe only 500 blocks later, when both chains have been accumulating identical blocks for all this time.

I don't see why you think it is impossible.  Each staker will empty the (identical) mem pool in his block he is staking.  And especially when there is multistaking, all transactions will end up in both chains, and hence add identical amounts of coin-lost-days in both prongs.  Yes, sometimes one will be in advance to the other, but within an error margin that multistakers are not going to waste, because somewhere else on the network, a stronger chain may be propagating that they didn't receive yet.

This is not the definition of a "decentralized system", in fact.  You are confusing a decentralized system and a democratic (centralized) system.

A decentralized system is a system where there is no deciding power or hierarchy.  "the majority wins" means that the minority must undergo the will of the majority.   In a decentralized system, there are no votes, but there are emergent properties nobody actually decided on, and certainly not a "majority".  For instance, price setting in a competitive market is a decentralized system.  Biological evolution is a decentralized system.  Science is a decentralized system.  The internet is partly a decentralized system.

Of course, de facto, there is some "weight of the majority", but that doesn't come about with voting, but rather as an emergent property.  If the majority DOES X, then the consequences of X will lead to properties of the decentralized system.

From the moment that there is voting, governance and so on, the system is centralized.  The dynamics is not emerging any more, but decided upon, and imposed.

As to your question regarding bitcoin: bitcoin is losing its decentralized aspect, but it isn't yet fully centralized.  As long as there are more than 3 non-colliding miners of which none has more than 50%, centralization is not complete.  For the moment, 5 miners are deciding on the fate of bitcoin.  As long as they don't go and sit together in a room, bitcoin is still decentralized over a few entities.  The day these 5 people talk to one another, bitcoin reaches the end of its decentralized part of history.

The principal need for decentralisation in bitcoin is immutability (of history and of protocol) and permissionlessness.

Which doesn't mean that bitcoin will not continue to function.  It will be like facebook or google.  It is almost there, but not yet.  Immutability of history will remain.  Protocol however, might change.  And the most problematic aspect is most certainly permissionlessness.  You will need your "licence to use bitcoin" eventually.

Big, big big Amen.

You still have some time    Bitcoin will fail in much more mundane ways long, long long before that will happen. 

It is much simpler to use other crypto whenever that comes, than to try to repair a broken car.  But for the moment (and many many years to come), it isn't broken.  Forget the idea that a crypto currency is "for ever".  It has a life cycle, like any product.  And really, bitcoin will reach the end of its life cycle long, long before that cryptographic system breaks down.   There's nothing wrong with systems reaching the end of their life cycle.  You simply switch to better systems in that case.
It means you stop buying bitcoin and selling bitcoin, and you start buying Xcoin and selling Xcoin.  Nothing dramatical.  But don't dream that bitcoin is forever.  Nothing is.  Especially nothing that is in this technology.  Even today, bitcoin tech is oldfashioned and clunky, where it was brilliant and shiny when it came out.  It will still last a while.  But not as long as elliptic curve crypto.

Indeed, that would be the simplest thing to do, except for a few reasons:

1) those capable of forking (miners) don't really want to fork.  BU is bluff to keep segwit away.

2) the nasty clunky old-fashioned 2000 block difficulty adaptation in bitcoin.  Unless the fork changes that too, you cannot afford to have a serious fall in hash rate.  If you are the minority, with, say, 20% of miners on your side, each block will take 50 minutes instead of 10 minutes (5 times less hash rate for same difficulty that gave 10 minutes before), and you will have to do this for 2 - 3 months before difficulty is adapted: by that time, your chain is dead, customers voting for it or not in the market: you have no room on the chain, and the whole point was to have bigger blocks !  If you have only 10% of the miners, you will have a block every 2 hours, and it will take half a year to change that.

3) the bitcoin brand name is all bitcoin still has.  "being an alt coin" is a horrible idea for bitcoiners.


That was already the case when the first altcoin was made.  It is a stupid idea that sold very well though.  If one would have kept the block chain rewards constant, with constant additive, and diminishing relative inflation, all of this fee war wouldn't have happened.

I honestly don't understand what this means.  There are different PoS mechanisms, but I was thinking of just any mechanism where "randomly" someone can stake, and nobody else, and that the probability of it to be you, is proportional to your stake ; I don't want to go into details of how this is done (there are different ways).  The idea was simply that at this point in chain and time, it was Joe's turn to stake.  And then Jack's turn to stake.  And then....


But you don't care about that "power to decide".  You only care about reaping in the reward.  We were thinking about otherwise "honest" stakers, that do not want to double spend or anything.  Just reap in staking rewards (minting).  You mint wherever you can, because that's why you are staking in the first place.  Not to impose any decision you don't care about.


This very very small "proof of work" doesn't compensate the possible loss of a minting reward if ever your chain doesn't win because of propagation delays in the network, or because of an orphaning 5000 blocks from now.


So your PoS is in fact PoW ?



If nobody is multistaking then 1) you cannot know and 2) then the consensus is just imposed by randomness (which is good), but you are not protected against multistaking.  That's the point.  As long as people only run official "random selection" staking algorithms, indeed, PoS will find a consensus.  But it is vulnerable to multistaking and having no possibility to come to consensus.  It's most probably because PoS is only used on rather small chains with not much at stake, that nobody has been implementing multi stake algorithms.


Of course, that is exactly the divergence of consensus that is the "nothing at stake" problem.  If, however, people are so gentle as to adhere all to a random selection algorithm and not multistake, then by definition, this divergence cannot happen.  That is similar to saying that if nobody is attempting a 51% attack on a PoW system, then the chain is secure with PoW.


Huh ?  But what if you have 2 identical blocks, where only one single transaction is different:
Block A: Mary has spend 10 coins that are 517 blocks old to Joe
Block B: Mary has spend the same 10 coins that are 571 blocks old to, Janice.

Block A and block B are identical except for this double spend.  (this is the ONLY case where consensus is necessary).

Now, Alice receives block A at 12:00:02 and mints block C on top of it.  However, it only gets to Joe at 12:01:00, but gets it at Jake at 12:00:30.
Alice receives block B at 12:00:08.

Bob receives block A at 12:00:09, but received block B at 12:00:05, and mints an IDENTICAL BLOCK D = C on top of it.  Bob gets it at Jake at 12:00:45 and at Joe at 12:00:45.

So now Jake receives Alice's chain before Bob's and Joe receives Bob's chain before Alice's.

So Jake will build block E identical to block F on top of Alice's chain
Joe will build block F identical on top of Bob's chain.

Everybody is honest here.  But at the end of the day, after I don't know how many identical blocks containing identical transactions, there will be a difference, orphaning one of both.  As long as all stakers include all the transactions in the common mem pool, their blocks are identical according to your criterium, and the double spend by Mary is not resolved by consensus on both prongs.


If the transactions are identical, yes.  And you stake for sake of the minting, not because you want to decide something.

Still no reasoned arguments, right


Religious zealot blah blah.  Believe !  Believe !  See the Light !  But no hard logical arguments, no gedanken experiments (and if you do, they talked against your stance each time), no rigorous analysis.

Yes, of course.  What's wrong with it ?


The behaviour of a piece of software will of course not stop anyone from using another piece of software if that is more advantageous to him.  Otherwise, there wouldn't be any consensus problem at all: "just use the right software" is not a trustless decentralized consensus mechanism.  

The reason why everybody HAS to stake on every chain it can, is that as long as both chains are in competition, you don't know which one will win, so the optimal strategy is to stake on both, to win with 100% certainty.  With PoW, this costs money, you can't, and you are betting on one or the other one (or dividing your influence).  But with PoS, you stake everywhere you can stake, of course.
 

Why would he forego his staking reward if ever it is the other chain that wins ?  If he receives both chains, why prefer one over the other ?  
  

What is 10 seconds more for Joe can be 10 seconds less for Jack, because of network propagation delays.  Even if they are "honest". Joe has to take into account that his preferred chain may lose against Jack's if it is propagated faster.

You see, it is not a matter of honesty.  It is a matter of not being able to reach consensus, because there are two equally valid histories, and you are now saying that their priority depends on the speed at which they propagate.  But that is not the same for all paths through a P2P network.  So the chain that is "winning" for you, is maybe losing for another part of the network *and there is locally no way to know who will win*.

I think you misunderstood the problem of nothing at stake.  It is not a matter of good guy and bad guy.  It is a matter of not reaching agreement on the past (no consensus).

I take your example again:

Joe: 50  Jack: 10  Alice: 10
Jake: 50 coins ; Zoe 50 coins


Jack makes block A and wins 1 coin.

Alice makes block B and also wins 1 coin.

Joe now makes block C (on top of A) and wins 5 coins in that chain ;
Joe also makes block D (on top of B) and wins 5 coins on that chain ;

Jake now makes block E (on top of C) and wins 5 coins on that chain ;
Jake also makes block F (on top of D) and wins 5 coins on that chain ;

Zoe now makes a block G (on top of E) and wins 5 coins ;
Zoe now makes a block H (on top of F) and wins 5 coins ;


....
 

What is now the good chain ?  A C E G or B D F H ?

There's no way to solve the riddle, until, 5000 blocks later, someone liking Jack and hating Alice, (Maybe Jack himself) doesn't add a block on chain 2.

I agree with you concerning ETH.  I don't like it because it is centralized.  I was fond of it because of its smart contract notion, until I understood the ridiculous and unnecessary disaster of Turing Completeness in such things.  But you are 150% right.  ETH has a lot going for it in the greater-fool market.  (but DASH too - I won't touch it with a stick even though I know I could make a fortune with it - I'm not interested in making a fortune for very evil personal reasons ; I'm only interested in living the good life I'm living right now).

That said, I'm losing faith in decentralized systems.  I think after all that Sun Tsu was right: the army that wins is the one that obeys its commander:


I can only think of small scale niche communities with a higher desire for decentralization than for benefit.