Gold collapsing. Bitcoin UP.

[sickpig]

@ladoga I think he is pretending, really. Both technically and on an economic level.

Firstly, I can't believe his not aware of QoS and traffic shaping in general.He's bfgminer lead dev.

Secondly he is in bitcoin since the beginning, he managed a pool all along. Due bfgminer development he has received multiple miners gear free of charge, i.e. ROI guaranteed. Just to say that even if he's not filthy rich by now he should have enough resource to have a dedicated internet connections for his nodes.

 

[awemany]

great to see you back @awemany. esp since you were one of the original proposers of the BU idea.

some of the concerns with the configurable setting is the unknown complexities of such a thing. @rocks has elucidated some of them and i worry about them as well. my main goal is introducing an option that has the greatest chance of acceptance by the community. the experience with XT imo is that the community worries greatly about any change to the code that might introduce unpredictability. this is understandable. in that vein, the fewer the changes, the better. also, personal projects also introduce worry. those are to be avoided.

this would be the thinking and reasoning for simply lifting the limit in BU and leaving it at that. one could then simply make the argument that we are returning the code back to Satoshi's original vision and no more. that would avoid any pretences of COI and complexity.

Well, as I said, I'd be fine either way. I think the important thing now is to get the various parties who have assets like /r/BU or bitcoin-unlimited.com together. Is there a separate thread for BU discussion that I missed somewhere already?

 

[theZerg]

it's time for fiat to be diverted into the currency unit itself:

By investing in Bitcoin startups, VCs were betting that a specific application will succeed regardless of technology. But these same investors should have been betting that Bitcoin will succeed regardless of application.

 

[cypherdoc]

@theZerg

precisely.

the real killer app was hiding directly under their noses for a long time.

i personally know investors who spread their money around way too much looking for diversification. in theory, not a bad idea. but when you understand the implications of a fixed supply currency by studying history, bitcoin is where it's at.
[doublepost=1445955712][/doublepost]

Well, as I said, I'd be fine either way. I think the important thing now is to get the various parties who have assets like /r/BU or bitcoin-unlimited.com together. Is there a separate thread for BU discussion that I missed somewhere already?

i own btcu.io and bitcoinu.io for what it's worth.
[doublepost=1445956090][/doublepost]let's try this again:

 

[Justus Ranvier]

"If blocks were unlimited, miners would accept every transaction, no matter how small the fee (2), which would set the per-tx fee to ~1 satoshi (3). Rational miners will thus limit the number of txs they accept (1) to ensure fees don't go to zero."

Sometimes I think the biggest threat facing Bitcoin is otherwise well-meaning people drawing incorrect conclusions from spherical cow game theory models.

Since bandwidth is not free, and since there's no way for a miner to a) know that a transaction exists, and b) signal to the network that they are including that transaction in their block without consuming the bandwidth that they have to buy, then there's always going to be a minimum production cost, and if they accept transactions that pay a fee lower than that cost they are going to lose money.

I guess some people are looking at the protocol that since the current protocol does not give miners any way to filter out the unprofitable transactions before spending bandwidth to download them, then it never will so they'll be forced to accept any fee-paying transaction at all without the benevolent economic dictators giving them a protocol-enforced production quota.

If, on the other hand, you assume that miners are sentient then you have to assume they'll notice their business becoming unprofitable and, like every other non-failed business, will adjust their behavior to make it profitable again.

This means that if the current communication protocol is unprofitable at higher transaction rates, then the current communication protocol will not be used at higher transaction rates.

 

[cypherdoc]

that's a severe non-confirmation on the part of the $DJT:

 

[Melbustus]

Sometimes I think the biggest threat facing Bitcoin is otherwise well-meaning people drawing incorrect conclusions from spherical cow game theory models.

Since bandwidth is not free, and since there's no way for a miner to a) know that a transaction exists, and b) signal to the network that they are including that transaction in their block without consuming the bandwidth that they have to buy, then there's always going to be a minimum production cost, and if they accept transactions that pay a fee lower than that cost they are going to lose money.

I guess some people are looking at the protocol that since the current protocol does not give miners any way to filter out the unprofitable transactions before spending bandwidth to download them, then it never will so they'll be forced to accept any fee-paying transaction at all without the benevolent economic dictators giving them a protocol-enforced production quota.

If, on the other hand, you assume that miners are sentient then you have to assume they'll notice their business becoming unprofitable and, like every other non-failed business, will adjust their behavior to make it profitable again.

This means that if the current communication protocol is unprofitable at higher transaction rates, then the current communication protocol will not be used at higher transaction rates.

Agreed. Such linear-extrapolation thinking is what leads guys like Adam Back to declare Bitcoin mining to be a tragedy of the commons situation (one that they must "fix" via introduction of careful and precise incentive/economic management through code....bleh).

The obnoxious thing is that if you try to argue points like the above with some people, the conversations just get increasingly arcane. There's no appreciation for letting the market figure it out; it's a completely deep-engineering-first mindset.

 

[cypherdoc]

honest question:

how do the transport vs consensus layer theorists reconcile @rocks argument that configuration is just moving block size consensus from miners out to full nodes?

 

[Justus Ranvier]

There's no appreciation for letting the market figure it out; it's a completely deep-engineering-first mindset.

The central planner's fallacy: "Everyone except me is a deterministic robot. If I don't do their thinking for them, then it won't get done."

 

[cypherdoc]

Such linear-extrapolation thinking

wow, that's one i haven't used in years.

i used to describe miscreanity that way when we argued about gold vs Bitcoin from 8/11 to 12/11 in my first gold thread: https://bitcointalk.org/index.php?topic=35956.0

you're absolutely right; Adam is exactly the same way. the way he talks you'd have to believe he'd keep his foot on the pedal while driving towards a brick wall.
[doublepost=1445961236][/doublepost]

The central planner's fallacy: "Everyone except me is a deterministic robot. If I don't do their thinking for them, then it won't get done."

he actually says this multiple times in the Trace Mayer/Adam Back interview last month.

they don't even bother hiding it at this point.
[doublepost=1445961562][/doublepost]i've been watching the $DXY and the $DJI closely for months. i'm convinced they're moving together as i have been warning about while the TLT is moving counter to both of those. that has severe implications for the US economy, imo.

 

[Peter R]

"If blocks were unlimited, miners would accept every transaction, no matter how small the fee (2), which would set the per-tx fee to ~1 satoshi (3). Rational miners will thus limit the number of txs they accept (1) to ensure fees don't go to zero."

I'll make my routine comment that miners will only add another transaction if the extra fee revenue is greater than the marginal cost of including that transaction--so no, fees won't drop to 1 satoshi any time soon.

But there's another point to be made about what the statement "if blocks were unlimited" really means. Sure, blocks may be unlimited in theory, but if the largest block ever mined was, for example, 2.1 MB, then as a miner I'm probably not going to attempt to publish a 100 MB block because I know that a practical limit will exists somewhere. Furthermore, as a miner, I can only really calculate my orphaning risk cost for blocks similar in size to the blocks I've produced in the past. I could extrapolate to what my orphaning risk might be at 100 MB, but, since I know I could be very wrong, it would be hard to ever justify mining such a block.

 

[Zangelbert Bingledack]

how do the transport vs consensus layer theorists reconcile @rocks argument that configuration is just moving block size consensus from miners out to full nodes?

It seems to me that as long as mining is centralized enough that most full nodes are not also mining nodes, we will always have the problem of most full nodes not having a real stake in the decisions despite having a voice. It might have to wait until mining fully commodifies.

However, I'm not sure I see how it's possible to cook up a bunch of full nodes to block larger blocks. Wouldn't it be that as long as there are enough full nodes willing to transmit the blocks they will go through the freely relaying nodes and just bypass the blocking nodes? There may be some leeching or spoofing that the blocking nodes can try, but it seems there could be a way to thwart that. If that is impossible due to Sybil attack type issues, then perhaps we really do need to wait for full mining commodification.

 

[theZerg]

honest question:
how do the transport vs consensus layer theorists reconcile @rocks argument that configuration is just moving block size consensus from miners out to full nodes?

The configuration proposal simply uses the power that nodes already have. So this power WILL be used someday; best we use it for positive results rather then wait for someone to use it to attempt to destroy the network.

Right now miners have all the explicit control. So why don't they just all mine 0-txn blocks? Because they understand that they need the community that makes a bitcoin valuable. Right now they simply guess as what the community wants (or read reddit and talk to people). Having full nodes with configuration options allows the community to express their choices within the protocol. Checks and balances.

There is clearly the danger of sybil node attack, and retrospectively it would have been great if Bitcoin made it harder to do this. However, sybil attacking DOES require some resources (in this case a bunch of VMs), and a full node's ability to influence the network is minimal. A node's abilities does not allow it to directly gain monetary advantage, it simply allows it to influence the long term evolution of the network. And finally, someone with the resources and drive to sybil attack the network:

1. Will not be deterred by having to modify the bitcoin core client himself (it would be necessary to significantly modify it to mount a cost-efficient sybil attack anyway).

2. Will mount much more effective attacks than adjusting the block size. For example, he could attempt to partition the bitcoin network entirely, or blacklist certain classes of txns.

 

[cypherdoc]

@Zangelbert Bingledack& @theZerg

let me be clear.

it's not that i think user configurable block size settings is a bad idea. i just think adding it could kill BU's chances of being accepted. we know for sure that core dev will make up all sorts of shit as to why it's bad; perhaps enough to kill adoption. that's my worry.

but then again, you didn't address my original question. the configuration setting, to me, seems to simply push block size consensus out to the nodes from miners. core dev will certainly pick up on this inconsistency in our arguments and use it against us.
[doublepost=1445964521][/doublepost]i also agree that the Sybil attack is highly unlikely in the presence of configuration. but that's not my point.

mine is one of political palatability even for just an unlimited BU w/o other features.

 

[AdrianX]

honest question:

how do the transport vs consensus layer theorists reconcile @rocks argument that configuration is just moving block size consensus from miners out to full nodes?

Actually with BU (as in simply Unlimited - not limited my node operators) full nodes would still have the final say. BU just would just allow miners to optimize block size as they're the ones who need to make it work. As opposed to the node operators who have an incentive to reduce block size to minimize overhead.

 

[Justus Ranvier]

he actually says this multiple times in the Trace Mayer/Adam Back interview last month.

they don't even bother hiding it at this point.

In that case, it's probably worth spending some time to deconstruct that error.

Here's a video apparently intended to make fun of iPhone users, but it actually demonstrates a point different from that which the author intended:


Why is the salesman wrong in this video?

 

[cypherdoc]

big, big, badass move DOWN by $DJT
[doublepost=1445966545,1445965626][/doublepost]

 

[theZerg]

@cypherdoc let me try again: Even with max block configurability full nodes cannot determine consensus. The full nodes will ultimately accept excessive blocks if the blockchain containing the excessive block grows. And that blockchain can only grow via the actions of other miners. So miners decide whether an excessive block should be build on. All the nodes can do is slow down the propagation of excessive blocks across the network, giving another competing block the chance to be discovered.

I actually think that configurability would help acceptance. Because otherwise how will you counter the argument that a "rogue" miner could create a 2GB block and destroy (let's imagine that a block that size causes all sorts of issues with people's nodes) the network? Even if a competing reasonable block is issued quickly, IIRC forks are still processed by full nodes.

 

[cypherdoc]

@theZerg

remember that i said it's not that i don't think configurability isn't a good idea (it might be a bad idea due to this muddiness of the ultimate acceptability of chain growth). i don't know the answer.

my concern is one of political palatability. less could be more in this situation.

I actually think that configurability would help acceptance

the way i've come to look at this is that block size was only made a pseudo-technical issue by core dev b/c of the political issue of COI. if we remove the limit completely, it no longer becomes an issue at_all since it should be completely handled by the free mkt. therefore making configurability a non-issue. it would certainly play well with the masses (and even ourselves) as we can always say "hey, look at us, we're offering block size configurability!" but does it really matter given how the issue evolved politically in the first place and that we always could set it at the CLI?

 

[Peter R]

Why is the salesman wrong in this video?

Listen to your customers!