That's the guy who seemed to be claiming at Davos 2016 that the only money banks lend out is that money depositors give it. He conveniently 'forgot' about the 90+% of loan money that they zap into creation out of thin air by the very act of making the loan.

Oh the horror. It occupies $3.09 USD worth of disk space. Whatever shall we do?

http://www.newegg.com/Product/Product.aspx?Item=N82E16822236971

I thought you probably did. Sorry for the use of unfamiliar jargon.


No slight intended. I formed the impression that you thought the discussion over the max block size was a recent phenomenon. I just meant to point out that it has been ongoing for years.


By 'trajectory' do you mean to denote the rate of growth in the transactions per day? For there is nothing that _can_ be done about this curve. It just is, and reflects all the multitudinous decisions that scores of independent actors arrive at in their usage of the blockchain. It is a reflection of the actual amount of use that people are getting out of Bitcoin. Accordingly, there is nothing to be done about the trajectory. What we can do something about is the capacity for transactions per unit time.

When should such a change come about? I would advocate ASAP. The 50% bet, based upon a 2x/year extrapolation, is somewhat less than a year. But what if we get a surge in adoption? For we don't know when the next great influx will arrive. If it happens next month, we will have scores of befuddled newcomers utterly unable to comprehend where there money disappeared to. With attendant widely-published horror stories. Which runs a very good chance of shunting these -- and future -- newcomers onto some other system. One without such a stupid artificial restriction.

 

I'm thinking you must have thought something other than you typed here. A certain number of usages will indeed "bring us into this filled up capacity". This is pure math. The consequences that fall out of such an event can be debated, but the capacity is the capacity. The math is the math. It ain't even sophisticated math. Simple elementary school arithmetic will suffice.


First, the problem is not "a variety of attacks from a variety of angles". The problem is simple usage of Bitcoin, in the manner it was designed to be operated, absent a change in the max block size, is limited to 'about 350,000' transactions a day. Period. If more transactions than this are attempted, for a sustained period, those transactions will not be simply delayed. The excess over 'about 350,000' per day will never clear.

For the umpteenth time - the issue is NOT dust transactions. The issue is that, with max block size capped at 1MB, the system is simply incapable of handling more than ~350,000 transactions a day - no matter the value of those transactions. Quit changing the topic. Quit putting words in my mouth. Such is dishonest.

You can call it shoddily constructed. I'll call it precise enough for the purpose at hand.

Just so we understand each other, would you be happy if adoption stalls out at somewhat less than twice of where it stands today? (We'll get to following arguments as a follow on....)


Now who's advancing the debate by zero percent?

If you think the world has enough Bitcoin users as things stand today, you have a point. If you think that use of Bitcoin growing by a factor of more than several dozen percent is desirable, then your point is ridiculous.

Nonsense. If efficiency was the be all and end all, then we would not be communicating at this very moment over formatted HTML, which is hideously inefficient as compared to raw UTF-8 text.


Hmm. Interesting. By that measure, Bitcoin is not at the bottom layer. It is already sits atop the routing layer. It even sits atop the Transport layer, which in turn sits atop the Routing layer, which sits atop the Link layer. So raw bitcoin transactions already are already atop several layers of lower level protocols. Hmm. So what was your point again?

In order to have an informed position of the severity of the problem, yes.

But first, let me step back. I may have befuddled you with jargon, and jargon may differ from time to time and from place to place. But the concept is pretty simple, and I imagine you grasp it intuitively, if not intellectually by the jargon I employed.

In a two-dimensional plane, two lines that are not parallel will eventually intersect. This is a geometric mathematical law. Let us illustrate the concept with an example: https://blockchain.info/charts/n-transactions



What we are looking at is a chart of the number of Bitcoin blockchain transactions per day, plotted against the time axis of the last year. While it is very noisy, we can see that the trend over this year interval is that the number of transactions is clearly rising.

If we wanted to, we could engage in a process called 'curve fitting', where we kind of average out the high-frequency highs and lows, to show more of the trend, and less of the day-to-day variation. Even a moving 7-day average filters out a lot of spikiness: https://blockchain.info/charts/n-transactions?showDataPoints=false&timespan=&show_header=true&daysAverageString=7&scale=0&address=



As we filter out the day-to-day noise, we get a curve that we can use to extrapolate into the future, to make educated assumptions about what is likely to happen to that quantity as time passes.

Let us now perform a curve fitting to a straight line - while this may not have great fidelity to the past, in the absence of additional info, it is about the best we can hope for as a prediction tool. Eyeballing the graph, we see that we have gone from 'about 90,000' transactions a day one year ago to 'about 180,000' transactions today now. Or roughly doubled in a year.

So let us mark a point today at 180,000, and another point one year ago at 90,000, and draw a straight line through these points.

So there is one of our non-parallel lines. For the other, we need a representation of the max transactions per day that can be processed, based upon the max block size. Today, that number is 'about 350,000'. A year ago, and in all the intervening time, that number is 'about 350,000'. Before that, dating back to the time the 1MB limit was put in place as a temporary DoS avoidance scheme (back when bitcoins were cheap as pennies, BTW), that number was 'about 350,000'.

To represent this on our graph, let us draw a mental line straight across from left to right at the level above the current graph, where 350,000 would be if we preserved the Y axis scale. This is our other non-parallel line.

So with the max block size fixed at 1MB, yielding a hard limit of 'about 350,000' transactions a day, we can project the slope of our actual transactions per day line into the future. We find that they intersect at a point somewhat less than a year away.


Look - I get that you're relatively new here. But rest assured we have been discussing this very issue for several years. Without that perspective in the rear view mirror, you may feel that the urgency is exaggerated. But to those of us who have been tracking this incipient problem for years, the situation does indeed appear dire.

The issue is that, once the lines intersect, growth is limited at the 'about 350,000' transactions a day level. The upward trending line of transactions a day flatlines - limited at the intersect to the system capacity. She can' take any more, cap'n'

You may claim that nearly a year is enough time to work things out. And that is the ~50% probability (assuming the 2x per year is a good curve fit). But those who have been around this for a while know that this animal seems to get adopted in fits and starts. If we had a proportional spike like we did in Jul, or in Sep, or in Dec, the system could not handle the volume. Period. No capacity whatsoever. If it was a spike, the backlog would eventually clear out. But how would the newcomers feel about their money being in limbo for a month? And what if it would have been Bitcoin's eternal September - we would kill all prospects of an explosion in adoption like the Internet experienced in 1993.

note: this is a napkin analysis. Quibbling on the actual numbers does not affect the outcome of 'we are indeed running out of time'.

When blocks become persistently full, you will never be able to safely assume that a zero-conf transaction will ever get included in a block.

And that is a fact.

Do you understand the principle of non-parallel lines in a two-dimensional cartesian coordinate system?

Actually several years. We've been repeatedly pointing at the same stupid, unnecessary hard ceiling, and pointing out that the inexorable trend of increasing transactions has us on a clear intersect.

In the meantime, we have recently gone from things never sucking in regards to capacity, to things sucking for brief flashes of time. The issue is not how much one needs to pay to get a transaction through, the issue is that with the current block size, no more than about 350,000 transactions can be processed in a day. Period. No matter how much money is thrown at the transactions.

The miners will spin up as many nodes as are needed for the use of the miners. Compared to the cost of a viable mining operation, the cost of operating a node is lost in the noise.

Yes, users have the power to abandon the chain en masse. That is the only power users have. Non mining node operators (note: I would fit that description, so this is not meant as a mere insult) have zero-point-nothing power.

The power rests with the miners. They can do anything that does not cause the users to overwhelmingly abandon the chain.

Yes. It is. The clear trend is that transactions per unit time are increasing. At some point in the very near future, bitcoin will hit a ceiling that it has not before hit - the situation where blocks are persistently full.


I think BJA would argue against "everything fine". I'm in the camp of "things are gonna soon suck, and will continue to suck, until the pain rises to the point where most admit things need doing".

You know any significant miners that are not nodes? No, neither do I.

Here's the dirty little secret nobody wants to discuss: pure nodes have no power. None whatsoever.

What miners want, miners will get. The only counterbalance is the prospect that the users will abandon the coin en masse.

Maybe. Dunno. I refer to it as 'simple arithmetic'.


I suppose anyone who resists the change is at risk of losing value, yes.

ha. hahah. HahaHa. HAHAHAHAHa!

As measured how?

Well, no. If any more than 50% of the hash power decides to change Bitcoin, then Bitcoin will change. It is a certainty.

It may be painful as the longest chain waffles back and forth between the changed and the unchanged, but eventually, the chain built as per rules agreed to by the majority hash power will dominate. This potential to waffle is why the threshold for activating the new rules is put at a supermajority level - say... 75%.

Well, no. But I see you're going to do it anyway.

No - banks absolutely and unequivocally create most of the money they lend out of nothing. It is zapped into being by the mere act of making the loan. Given your long and august participation on this forum, I am surprised you do not know this already.

eta: reading the rest of the thread to this point, I must have misunderstood your statement? I hope so.

And I was so hoping we could discuss the following:


So just to be clear - do you maintain that block size increases are necessarily "Removing the rules against actions that the network protocol expressly forbids", and are therefore necessarily evil?


- Capacity increases for the Bitcoin system: https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-December/011865.html


For better or worse (I would say for worse), in this era of industrial mining, non-mining nodes have essentially zero power. Any viable mining operation has sufficient  resources to run a node of its own, and connect explicitly to other mining entities that share its philosophy. The only power outside of miners is the threat that users abandon the chain en masse.