Gold collapsing. Bitcoin UP.

[cypherdoc]

I like the looks of this. BU to the rescue!:

 

[awemany]

Possibly a stupid question but... I've been wondering about BU and nodes ability to set their max relay size. This is supposed to give nodes more of a vote in the max size by making larger blocks slower to propagate right? But miners have an incentive to mine larger blocks then the nodes may want because they collect more fees.

So what is to stop the miners from forming a sort of private network where they relay blocks directly between each other, bypassing nodes individual settings, forcing them to accept blocks due to their block height acceptance rules?

As another thought along those lines, what stops say, the top three miners from getting together broadcasting blocks between each other before the rest of the network, giving them a slight advantage over the rest of the network on starting the next block?

Good question. The answer is - nothing except for incentives. Incentive here is keeping the network functioning for a large amount of people, so that transactions can come in and the generated coins and collected fees actually have value and can be transacted.

Of course, a single large hash rate miner might go and create a lot of insane, PB-sized terrorblocks in his basement. He might even be so secretive to never tell anyone about it, yet he might have the largest blockchain, and all others are invalid and therefore doom 1!!1!!

That is another whole point in this debate which is always narrowly missed. (This might actually be due to too much adversarial thought?)

Bitcoin fundamentally depends on people wanting to interact and wanting to build a common and great payment system. The driving force behind all this is the human desire to build, create and make things better. You cannot explain Bitcoin just with prisoner's dilemmas and just a bunch of crooks and psychopaths outsmarting each other. The majority - the 50% - need to be sane and interested in a functioning system.

A lot of the pathological cases appear ridiculous as soon as you can see this fact.

 

[Zangelbert Bingledack]

As another thought along those lines, what stops say, the top three miners from getting together broadcasting blocks between each other before the rest of the network, giving them a slight advantage over the rest of the network on starting the next block?

Doesn't that delay just elevate their orphan rate?

 

[cypherdoc]

 

[_mr_e]

If this group formed > 50% wouldn't it act more like a selfish mining attack while bypassing nodes max block settings?

 

[cypherdoc]

Possibly a stupid question but... I've been wondering about BU and nodes ability to set their max relay size. This is supposed to give nodes more of a vote in the max size by making larger blocks slower to propagate right? But miners have an incentive to mine larger blocks then the nodes may want because they collect more fees.

So what is to stop the miners from forming a sort of private network where they relay blocks directly between each other, bypassing nodes individual settings, forcing them to accept blocks due to their block height acceptance rules?

As another thought along those lines, what stops say, the top three miners from getting together broadcasting blocks between each other before the rest of the network, giving them a slight advantage over the rest of the network on starting the next block?

Miners already do this through the relay network and SPV mining through stratum connections. I don't like it because it encourages bad behavior, as in non validation. Maybe with bigger blocks, more miner decentralization, and more value on the line, the relay network and SPV mining as a centralizing forces will get shut down one day.

 

[Peter R]

Possibly a stupid question but... I've been wondering about BU and nodes ability to set their max relay size. This is supposed to give nodes more of a vote in the max size by making larger blocks slower to propagate right? But miners have an incentive to mine larger blocks then the nodes may want because they collect more fees.

Well, bigger blocks would only increase revenue if Q* > Q_max, in which case I'd argue that limit should be raised. IMO, the limit in BU is mostly to put people at ease regarding "Terabyte Spam Block" attack.

So what is to stop the miners from forming a sort of private network where they relay blocks directly between each other, bypassing nodes individual settings, forcing them to accept blocks due to their block height acceptance rules?

Not much--only that it adds a bit of risk. Whether or not I'd expect someone to like BU, really comes down to his philosophy about the purpose of the block size limit. If one think the block size limit should serve as an anti-spam measure, then one will probably like BU. If one thinks it should serve as a policy tool to drive fees up (or to allow people to run nodes on low-bandwith internet connections) then one probably won't like BU.

As another thought along those lines, what stops say, the top three miners from getting together broadcasting blocks between each other before the rest of the network, giving them a slight advantage over the rest of the network on starting the next block?

They can do this already. This is the selfish mining attack describe by Eyal and Sirer. I believe my subchain proposal will reduce the effectiveness of selfish mining (but I can't prove it yet).

 

[bucktotal]

it seems like most (if not all) of the devs are academics. as such, i can't fault anyone for working on problems they find interesting. After all, its mostly free work. It may not always jive with certain ideals regarding bitcoin, but its hard to argue the problems aren't theoretically interesting. perhaps debatable what people should be working on, but thats just how it goes.

 

[Zangelbert Bingledack]

I agree, but it does give a clue as to who knows what they're talking about economically. The thing to do to actually verify it would be to go to all those that have altcoin involvement and see what they say about it. If they say something like, "Freicoin is better than Bitcoin because Bitcoin encourages hoarding," or if they hold more than a smallish amount, that would be definitive in my book.

 

[Zangelbert Bingledack]

Interesting line of discussion for forking in general and BU specifically:

This seems very natural, makes my inner economist happy, unlike the feeling I get from Core's 1MB cap and tightly managed bonsai economy. Let miners take responsibility for their own connectivity and cost/benefit analyses. Rather than devs having to think about orphans to divine one clunky blocksize cap for all times and all individual situations, let miners be the ones to think about orphans and how they will manage their blocksize tolerance. I assume if you have very cheap energy but are poorly connected, you will want to mine and preferentially relay different sized blocks than someone who has expensive energy but is well-connected, and then different sized blocks at different times of day and days of the week as fees offered would vary.

If we imagine Bitcoin had always been like this, it would seem downright amateurish to not have this kind of market incentive in the system. Of course, it is: it's the same kind of mistake economic newbs make all the time. We'd have people coming in telling us we need a blocksize cap or Bitcoin will fail sooner or later due to this and that, but under BU miners will become experts and really take ownership of maximizing their profits by understanding how the network works and how their orphan probabilities weigh against fee revenue for their own unique economic situation and dynamically through time.

This is in sharp contrast with the bobblehead 90% mining power at the Scaling Conference who seem to have little conception of such things and need everything spoonfed from Core. "Please tell us what to do!" It's like Core is the State and the miners have become sheeple because of the nannyism.

 

[solex]

The most difficult hurdle to jump in getting BU across to everyone is that unlimited does not mean "no limit" it means an "emergent limit" (in this case consensus).

So, the analogy is that a dot is a like a Bitcoin full node, and the "8" is an effective block limit emergent from all dots.

Unfortunately, emergence is one of the most important attributes of nature which is least discussed. Anyway, I'm thinking that these are good pics to give an easy visual for why BU is a blue hat solution to the block hard limit deadlock. This augurs well for its approach to other scaling problems with technical and economic considerations.

Some people can't see one, so here's two

 

[cypherdoc]

ha. those are Ishihara Color Vision Testing plates i use in my office.

color blindness is common in young men; esp geeks (i made that up)

 

[JVWVU]

Good work to both Peter T and Zerg, while I dont necessarily support a BU yet, I would trend that way over what we currently have.

I would like to see the person whose input I trust the most (NewLiberty aka haq4good) really be more active here to quanitify what BU would do.

 

[Zangelbert Bingledack]

Ludwig von Mises wrote the following in 1949. Consider it as relating to Core's interventionist/governmental approach to keep the Bitcoin node-automatons in line and BU's "laissez faire" approach to blocksize that recognizes node operators are people who respond to incentives, exist with various and changing circumstances and priorities, and can think for themselves:

As the interventionist sees things, the alternative is "automatic forces" or "conscious planning." It is obvious, he implies, that to rely upon automatic processes is sheer stupidity. No reasonable man can seriously recommend doing nothing and letting things go as they do without interference on the part of purposive action. A plan, by the very fact that it is a display of conscious action, is incomparably superior to the absence of any planning. Laissez faire is said to mean: Let the evils last, do not try to improve the lot of mankind by reasonable action.

This is utterly fallacious talk. The argument advanced for planning is entirely derived from an impermissible interpretation of a metaphor. It has no foundation other than the connotations implied in the term "automatic" which it is customary to apply in a metaphorical sense for the description of the market process. Automatic, says the Concise Oxford Dictionary, means "unconscious, unintelligent, merely mechanical." Automatic, says Webster's Collegiate Dictionary, means "not subject to the control of the will, ... performed without active thought and without conscious intention or direction." What a triumph for the champion of planning to play this trump card!

The truth is that the alternative is not between a dead mechanism or a rigid automatism on one hand and conscious planning on the other hand. The alternative is not plan or no plan. The question is whose planning? Should each member of society plan for himself, or should a benevolent government alone plan for them all? The issue is not automatism versus conscious action; it is autonomous action of each individual versus the exclusive action of the government. It is freedom versus government omnipotence.

Of course the tired response will come that this wouldn't work because "Bitcoin is a consensus system," but this no more implies that the blocksize need be ordained from on high than that the price of grain needs to be ordained in order for there to be consensus on it. Like the price system that coordinates the price of grain through communication and price feedback that emerges through the individual actions of millions of people focused only on their own cost/benefit analyses in bidding on and setting ask prices for grain, the blocksize can and should be coordinated through the dynamic communication and orphaning feedback that emerges through the individual actions of thousands of miners focused only on their own cost/benefit analyses in choosing what sizes of blocks to relay or build on (as well as perhaps other more direct signaling).

 

[Mengerian]

Bitcoin is often thought of as a technology: a payment system, network protocol, decentralized consensus system, open source software project. These are all valid and important aspects of Bitcoin. But I would argue that at its core, Bitcoin is a social agreement, and the technology is a practical means of implementing this agreement. I believe that this subtle but fundamental difference may explain much of the current schism.

The market is what ultimately ensures the sanctity of Bitcoin. Everyone who holds or transacts in bitcoin is here voluntarily, by free choice. We agree to rules as to what constitutes a valid transaction because we understand that these rules will lead to system that brings us value as individuals. The technology is a tool used as defence mechanism against non market, ie. coercive, interference.

With this in mind, technological development of Bitcoin should focus on two overarching goals:

1. Empower users to do whatever is in their self interest. For example: lower costs of transaction, make commerce easier, publish irrevocable timestamps, etc.
2. Defend individuals against coercion. For example: provide fraud proofs, help people protect their private keys, protect privacy, etc.

Bitcoin is a market phenomenon and the system works because we agree to a system where our individual incentives are aligned to give value to the whole system. We need to trust these market forces. Any type of top-down “collective defence” will not work if the means to secure that defence sacrifices individual interests, or involves coercion. We are here voluntarily. A system that sacrifices individual interests at the altar of collective benefit sows the seeds of its own destruction.

It is debatable whether or not the “big blocks = centralization” idea is valid. But if defending against this threat requires miners sacrificing potential profits, or transactors sacrificing cheap transactions, then it is untenable. Free markets have a way of postering innovations and balancing tradeoffs to maximize overall value without harming individual interests. People want security and they also want low cost to secure their transactions. We need to trust that a suitable balance between these interests will emerge in a manner aligned with individual interests.

 

[theZerg]

didn't @theZerg say he was releasing a BU whitepaper soon?

I plan to hit reddit with this tomorrow. You guys can have a sneak peak:

http://www.bitcoinunlimited.info/1txn

I think that it forms a solid theoretical basis for BU, when combined with Peter R's fee market and network latency -> orphan analysis.

Basically the paper analyses the current network and extrapolates it in the case of no block limits. I find a maximum network transaction commitment throughput (60Kb/s) and this therefore defines an effective maximum block size.

I then go on to address how 1-transaction mining can be used to essentially vote for the network block size. There is one major "hole" in the situation though and that is what if a miner produces and exa-block (very large block). I use game theory to show that in this situation the other miners will be more profitable if they agree to ignore it. I construct the miner profitability equation based on the idea that miners will switch to a competing block if the time to validate it is less than the time to validate the rest of block they currently have -- if a miner switches he will be able to mine transaction-filled blocks faster, which results in more profit.

It turns out that miners become unprofitable at around 30MB no matter the block size, and that right now (with today's txn fees and coinbase) transactions actually decrease profitability a tiny bit -- but really, not enough to worry about for blocks < about 5MB. But when it starts taking other miners minutes to validate your block, the opportunity to switch to a short block (and therefore orphan the large block) become so large that it will start making sense for other miners to do so.

Its interesting -- it shows the much-maligned 1-txn blocks in a better light. They are the way the network slows itself down. This use of 1-txn blocks during the new block validation phase only is ofc very different than SPV-mining which never validates. SPV mining is a pretty bad idea...

 

[VeritasSapere]

Node #19 up and running.

I set it up on a pretty nice Linux server inside of a datacentre, with more then enough processing power, bandwidth and storage, so bring on the big blocks I say I can take em. Lets bring so many nodes online so that we reverse the trend in node count and ease the fears of the small blockists.

I have not had the time to catch up on all of the great writing on this thread, I plan to however and I feel fortunate that we have such a great community here. The official release of Bitcoin Unlimited is cause for celebration, it has been a bright light in the darkness of this ongoing struggle, today it gave me hope and put a smile on my face. Thank you to everyone involved, lets continue these good works and do what is right and just. The original vision of Satoshi will triumph, I do think that we are on the right side of history.

I hope that everyone has a great Christmas, whether you believe in that or not, I wish everyone peace, love and happiness.

 

[cypherdoc]

I'd say the market is celebrating the release of BU.

 

[Peter R]

I plan to hit reddit with this tomorrow. You guys can have a sneak peak:

http://www.bitcoinunlimited.info/1txn

I think that it forms a solid theoretical basis for BU, when combined with Peter R's fee market and network latency -> orphan analysis.

Basically the paper analyses the current network and extrapolates it in the case of no block limits. I find a maximum network transaction commitment throughput (60Kb/s) and this therefore defines an effective maximum block size.

I then go on to address how 1-transaction mining can be used to essentially vote for the network block size. There is one major "hole" in the situation though and that is what if a miner produces and exa-block (very large block). I use game theory to show that in this situation the other miners will be more profitable if they agree to ignore it. I construct the miner profitability equation based on the idea that miners will switch to a competing block if the time to validate it is less than the time to validate the rest of block they currently have -- if a miner switches he will be able to mine transaction-filled blocks faster, which results in more profit.

It turns out that miners become unprofitable at around 30MB no matter the block size, and that right now (with today's txn fees and coinbase) transactions actually decrease profitability a tiny bit -- but really, not enough to worry about for blocks < about 5MB. But when it starts taking other miners minutes to validate your block, the opportunity to switch to a short block (and therefore orphan the large block) become so large that it will start making sense for other miners to do so.

Its interesting -- it shows the much-maligned 1-txn blocks in a better light. They are the way the network slows itself down. This use of 1-txn blocks during the new block validation phase only is ofc very different than SPV-mining which never validates. SPV mining is a pretty bad idea...

This is coming along very nicely!

Two quick questions:

#1. Is the data in Figure 5 real or simulated? It illustrates the effect very well!

#2. Do you want feedback on this and for us to point out typos? Is there a thread setup anywhere to make comments? (I've been keeping notes as I go through it.)

 

[albin]

@Zangelbert Bingledack
What a great excerpt, Mises exactly nails down what block size as "emergent" means in that passage.