RE: Time to End the Block-Size Blockade - by Roger Ver
Roger...this is just sad. I wanna run through some basic math with you here. I'm going to start with a few figures. First, it takes 5-10 milliseconds to verify the signature of a transaction. I'm going to be super optimistic and assume every transaction takes 5 milliseconds. An average, IDEAL, transaction is say 250 bytes. These figures I believe do not take into account gains from secp256k1, but even with those savings it will only give you so much head room.
So, breaking that down, it pretty much means you validate 200 transactions in a second, 12,000 in a minute, and 120,000 every ten minutes. Now this is a dangerous limit, this last number. If you push significantly beyond that, then statistically you will not be able to validate blocks as fast as they are being made, bitcoin will be a perpetually run away train that you can never validate to make sure the claimed current state is valid.
Now remember that figure about 250 byte transactions? An insanely overly optimistic figure by the way. That means that pushing up against that runaway train limit...is at 29 megabytes.
Yeah. 29 MEGABYTES. And you really only get a few orders of magnitude gains from secp256k1. Validation is the bottleneck Roger. Validation is also what gives it value, because I am validating and securing the integrity of my own money myself.
Stop attempting to use "a lack of understanding of economics" on the part of the devs as a BS excuse to scream for bigger blocks, proclaiming that will make bitcoin's fiat value increase.
It is more than just the technicals, though those alone make your proclaiming of GB and TB blocks out right insane, it is about maintaining the incentives. A big part of with is, I validate my money MYSELF. And that I can afford to do so, MYSELF. Paying hundreds or thousands of dollars a year on special purpose computers to validate my money is not cost effective, and not a sound and motivating incentive structure Roger.
I'm sorry, but you are the one who has a lack of understanding of how this all works, not any of the developers Roger.
EDIT for spelling
So by your own math, with today's hardware, Bitcoin can scale to 29 times the transactional capacity we have today. Thanks for making my case for me.
Wrong Roger, using layered approaches like Lightning Network and sidechains, it can scale thousands of times past todays transactional capacity while still maintaining the incentive balance by keeping validation costs low.
Okay, after mulling it over I think I found the source of your misunderstanding. It is by no means safe to just push up against that limit Roger. Keep in mind 10 minutes is the statistical average a block is found. The more you move towards that limit, the more of a potential problem this can cause. The bigger you make blocks, the more periods of high luck will cause problems because you are being sent a new block before validating the old one.
You cannot just push up to that limit, even more TOWARDS that limit starts to cause performance issues. Just fucking accept it Roger, you're a good guy at heart, but stop pretending you understand any of this when you clearly don't. Instead of being a sound public figure to advocate for the ideology beyond Bitcoin, and the social movements it will create, you are making yourself look like an idiot by speaking on matters you do not understand. Just stop it Roger. This is a 100% technical matter. The economics get involved when looking at how the system is FUNCTIONING. Economics will ALWAYS have to take a backseat to technical realities, and adapt to what is possible. Otherwise what you are suggesting is lets turn Bitcoin into a centralized service like Paypal, where nodes can refuse to relay your transactions for not using ID, being from somewhere they don't like, without paying a "membership fee" to use a nodes services. I do not think you want that, I don't want that, and I'm pretty sure most users don't want that.
Just be a man, step back a second, and realize you are WRONG here Roger. It does NOT make you a stupid person. It does NOT make you a bad person. It just makes you WRONG. And the sooner you realize that, the sooner we can all get back to being a productive community.
And also what is this comment supposed to mean? Your way of doing things only works so far until it hits problems, so just full speed at that brick wall anyway? You are just going to ignore technical reality and how the changes you propose weaken the quality of bitcoin that makes it valuable in the first place?
Umm ever heard of multi tasking? If verification is expensive why can't we just multi task it?
It is more than just the technicals, though those alone make your proclaiming of GB and TB blocks out right insane, it is about maintaining the incentives. A big part of with is, I validate my money MYSELF. And that I can afford to do so, MYSELF. Paying hundreds or thousands of dollars a year on special purpose computers to validate my money is not cost effective, and not a sound and motivating incentive structure Roger.
And why is 1mb correct? The point is the blocksize should be found as an equilibrium via market forces , not by design of some King. Even if he is satoshi or max or whoever controls core.
It is set by market forces. By the software people run to secure their coins. And right now it is voting 1 MB, and we'll see with the rollout of 0.13 if the market will support and prepare for Segwit rolling out to increase to a blocksize of 2-4 MB. Which is right up to the edge of what the network can handle in terms of latency increases right now.
Nope that is called hard coded. Market forces would mean no blocksize limit and miners deciding how much to put into blocks for what price.
He wrote " In the not so distant future" - perhaps you need to go back to school and retake the basic English reading comprehension class with the other 5th graders.
no, he thinks everything in this world runs on a giant CPU with 1 core and a single thread. lol
Your assumptions are off by quite a lot actually. sha256 chews through about 100MB/s. On 5 year old hardware. And single-threading.
So I'm wondering where that 5ms comes from.
That's not the only operation in tx verification in a block; the signatures have to be verified (the expensive bit), the inputs have to be checked against the utxo set, and the scripts need to be executed.
It's not that simple.
bitcoind is slow
Would have nice for you to bring a solution to the table and not just criticise. I'm all up for finding a solution to this this long standing problem so what's your idea if Rogers don't fit the bill???
Core's roadmap. We get a blocksize increase(bigger than the one Roger is pushing for), fix malleability, add compatibility for LN, and through raising the blocksize to a maximum of 4 MB increase the possible transaction throughput of the network to hundreds of thousands of transactions per second, not just doubling it to 14 tx/s.
Oh, and maybe you should ask Roger why he is intentionally disrupting testing on testnet? If he is soo enlightened and behind raising the transaction throughput, WHY IS HE ACTIVELY SABOTAGING TESTS ON THAT BY MINING ON TESTNET?