Har. He was on the right track, but the devil is in the details, eh?

If you have successfully demonstrated  that unbounded block size cannot reach an equilibrium outside a majority-collusion environment, I have missed it. If your argument is that there is no equilibrium in a majority-collusion environment, then your fear is as valid for today's Bitcoin than it is under unbounded blocksize. Nakamoto consensus is -- as far as we know -- dependent upon 'a majority of honest [miners]'.

And as a practical matter, Bitcoin operated just fine for multiple halvings with no practical bound on blocksize. This may be partially explained by a relatively amount of mining centralization, but such does not seem monotonically increasing. Certainly, we've had actually naked over half hashpower in a single entity before.

So if I read this correctly, your insight is that a collusion of parties can force their will upon the network? Yes - we've known this since 2009.

I don't consider your 'paranoia' unwarranted. However, I think keeping the transaction processing capability down works against your stated goals.

It would seem to me that your stated goal for 'decentralization' is an increase in the number of non-mining nodes. I think we can likely agree that the only population that would have any interest whatsoever in being a non-mining node would be Bitcoiners. I reason from this that increasing the number of Bitcoiners would add the the population of people that may self-select in becoming non-mining nodes. Indeed, it would be my expectation that increasing the number of Bitcoiners will -- all else being equal -- lead to more non-mining nodes.

It seems axiomatic to me that making Bitcoin applicable to more use cases, or more usable for the use cases it has, or downright cheaper to use, will lead to more usage, and more Bitcoiners. And that increasing transaction capability and decreasing transaction cost both serve these aims.

Long story short, I believe making Bitcoin more capable -- even at some real increment of per-node operational cost -- will lead to an increase in the number of non-mining nodes. I realize others may differ. And creating a controlled experiment is likely out of the question.

Of course, I really don't think non-mining nodes add any value to the network. Only to their operators. Again, I realize opinions on this matter differ as well.

Well, no. In addition to the protocol, the coin is also the ledger.

I don't believe you are right on your immutability claim. Plate tectonics may be analogous. Two forces are working in opposition. As the stresses get higher and higher, eventually something snaps, and a new equilibrium is formed.

Hey - it's an analogy, not a model.

While this is a reasoned observation, the extent of the powers of the organization extend only to the activities of the organization itself. It is merely an overlay to Bitcoin, which remains permissionless. At best, it may lead to quicker resolution on community-wide decisions. At worst, it is simply irrelevant.

Refuted? WTF are you talking about? I did not make an assertion. I asked to see the model upon which the conclusion was drawn. You can't say that a request for info is refuted. That's not defensible for any definition of the word 'refuted'.


That's possible. I asked you a couple questions regarding your assertions on another thread. Perhaps your answers will lead to me understanding. If that is the thread when such questions were asked, let us continue the dialogue there.

Well, no, it does not make it clearer. However, it is peripheral to my point. Which is that it is unlikely that BU miners will creep up initially. They will likely go straight to 2MB. For as long as it takes to clear the transaction backlog. For such would deliver a demonstrably better user experience (faster cheaper transaction processing) relative to the other chain.

Yes, further increases may be likely to be more incremental in nature.

No. That is not the most widely-discussed plan, AAR. Which would be >=75% for a full difficulty period, and another difficulty period (not dropping below 75%) for other wallets to upgrade. Only then create the first large block. Presuming, of course, they stick to that plan.

Of course, some other large blocker may jump the gun. But a premature large block would likely be orphaned off by all the other BU miners.


Umm... presumably you are aware that, even if there were patents or copyrights on Bitcoin preventing its free use, they would be unenforceable in a permissionless environment, right?


Well, no. Not really. They can slow it down, as difficulty is predicated on current network hash power. If 50% of the miners start hashing blocks that core will not validate, the core (and fork) block interval will be slowed to 20 minutes. If the fork occurred at 75%, core blocks would slow to 40 minutes. Which would increase competition for block space, driving up fees. BU blocks would slow to ~13 minutes, but have increased transactions per block, so throughput would actually be increased on the BU fork. Decreasing competition for block space, thereby driving down fees on the BU chain.

But yes, the likely result is that all but an intransigent few would migrate to BU. Because it would be functional, and the core chain would be largely not.

So if I understand you correctly, you assert that the fact that not everyone is willing to pay the same transaction fee rate in BTC/B, this is somehow demonstrative of a failure of free markets? Cause that's what it looks like you are claiming.

You seem to have successfully demonstrated the central argument that each miner is incentivized to include as many high-BTC/B transactions up to the point where transaction fee = marginal orphanage cost. So what is the problem? Why do you consider this as 'not free market'?

I see nothing in this that suggests BU is worse than core in this regard. Help me understand.

Sorry - I did not mean to offend. That was the rhetorical 'you'. I know you run a node - you've told me so, and if not, I would have assumed you did from your other postings. As do I.

I am more referring to the notion that node operational costs must be kept to an absolute minimum. In my opinion, if a potential operator is not willing to have some skin in the game, they will be worthless as a node anyway.

Though I reiterate my main point above: As long as there is no force barring you from firing up a node of your own, your criteria of access to trustlessness is satisfied.

Sorry. dinofelis is correct. Non-mining nodes have essentially zero power of enforcement.

When faced with a block, a node has two possibilities. It can accept that block and forward it, or it can consider it an invalid block and not forward it. That is all. What it cannot do is prevent that block from getting to another miner that is perfectly happy to create another block on top of it. If miners are extending the chain, demonstrably accepting those blocks by building other blocks atop them, there is doodly-squat that non-mining nodes can do about it.


Yes. BU plans a malleability fix. Later, after more pressing problems (e.g., the hard cap on the transaction throughput production quota) are fixed. Whether or not that fix ends up being SegWit is still a matter to be discussed. At this point in time, the other leading proposal would seem to be FlexTrans. But there is nothing endemic to SegWit that prevents it from being ported to BU.


Not ignore, exactly. Just that it is better development practice to not introduce several new major features at the same release.

Jeebuz.... They're not meant to be the President and Secretary of _Bitcoin_, just of the confederation. Their powers are outlines specifically.

President: makes sure BUIP numbers are assigned properly
Secretary: maintains a website

Get a grip!

Well, what it really says is that all these exchanges plan on listing BU.

Disclaimer: at least two of the signatories have since announced that the text was changed after they assented to signing.


Yes, but it would also gratuitously make BU 'not Bitcoin'. So it likely will not be implemented by BU. Clients already have all tools necessary for avoiding replay at their disposal. No protocol change needed.


Well, no:


source: https://blog.gdax.com/https-medium-com-adamlwhite-bitcoinhardfork-3f6ac05ceec6#.9eyge9yg0


Over 40% of the last 24 hour's blocks are signaling large block compatibility. 12 hours ago, it was even over half. Seems resistance to BU in the mining community is nowhere near universal. (https://coin.dance/blocks)


I'd like to know what reasons there are to believe that "Satoshi is adamantly against BU".


He just as plausibly might help BU by dumping on core. Not that I think either eventuality is very plausible.


Possible, but I doubt it. Part of the fork dynamics is dependent upon BU delivering greater value. If BU forks at 75%, then the 1MB chain will be about a 40 minute average block time, delivering 25% of its pre-fork throughput. With a corresponding rise in fees due to the rise in blockspace supply. The BU chain will be about 13 minute average block time. If BU stayed near its pre-fork block size, its throughput would be on the order of 76% of its pre-fork throughput. If, however, BU went straight to 2MB, then it would deliver 150% of its pre-fork throughput. With corresponding drop in fees. Blowing through the backlog in relatively short time. Cheaper transactions, more transactions: greater value. Leading to increased incentive for fence-sitters to throw in to the BU side.

Future increases may be more probing, and therefore incremental. But I would expect the 2MB immediately. They'll fall after the backlog is cleared.


Well, yes. That's exactly what it is. It would certainly be dramatic, price would tank, effect would be temporary. Kind of definitional, that.


Yes, but you are arguing against religious fundamentalists.


Um... by retroactively claiming that what the miners have done is exactly what the nodes wanted all along?

Really? Then you of course will not mind providing a link to him stating such.

I'd be interested to see your mining cost model, that allows you to claim that moderately larger block sizes invariably leads to greater centralization.


Why should BU add prevention for transactions to be mined on competing chains? To do so would be to ensure that there could never be a healing of the forks. This is simply bad policy. Hence, no plans to do so.

Besides, clients are easily able to build transactions in a manner that avoids this issue altogether. Protocol changes are not needed.


B does not follow A. Exchanges will deal in the majority chain, or be bankrupted by others that do.

While your nightmare scenario is technically possible, so are hash collisions.

Nonetheless, any node is free to use the existing invalidate block mechanism to ensure it will never follow the temporary loser eventual overtaker.


Are you asserting that nodes have the ability to cause one miner's solved blocks from reaching other miners?

The large block miners will not require small block nodes to propagate their solved blocks. Small block nodes will be routed around. Node operators can choose to follow the operational chain that is processing the bulk of transactions, or they can choose to follow the essentially non-operational chain which is nearly stalled. Nodes have that right, and they have that ability. Other than that, non-mining nodes are powerless.


Miners do not ask anyone if the block they have mined is valid.

How are those confirmations made? By miners mining additional blocks on top of them. Non-mining nodes have nothing to do with it.


Then users will abandon what Bitcoin has become, the price will drop asymptotically toward zero, and the miners will be left with row after row of rack after rack of hasher after hasher suddenly worth zero, having lost hundreds of millions of dollars in the process.

Nakamoto consensus. It's A Thing.


That would have been a defensible position 18 months ago, but that particular ship has long since sailed.


Not at all. I have consistently advocated this viewpoint. Lie much? Perhaps inadvertently, but a lie nonetheless.

As long as there is no force barring you from firing up a node of your own, your criteria of access to trustlessness is satisfied. If you are unwilling to spend the rather negligible amount of money required to do so, that is not a fault of the system.

BU is just Bitcoin. It is a fork of pre-segwit core, which has been modified to allow the market dictate max block size. The first time a miner mines a block larger than 1MB, it will result in the forking of the Bitcoin chain into two. Whether or not both forks have a viable future is currently unknowable. Whichever ends up the majority fork will no doubt be considered the real Bitcoin by the world at large. The expectation among the BU community is that the BU branch will be the overwhelming winner.


RAM is cheap and getting cheaper. Bandwidth is cheap and getting cheaper.


The expectation that the future health of the network depends upon people who will not spend 0.2 BTC on their machines is ludicrous.


Well, no. 8MB will happily validate larger blocks up to the 8MB limit. That'll likely be several years of BU compatibility. IOW, 8MB is counted into the tipping point.