Block-size Solution Proposal

[Was]

Hello All,

Let me start off by saying how fortunate I feel to be apart of this forum and the bitcoin community as a whole. This really is an exciting time to be alive. Between Bitcoin and the internet, the implications of technologies (can be) endless.

However, It is my feeling that too many somewhat lack the sense of responsibility when it comes to ensuring that this payment network becomes what it inevitably was meant to be. We have been granted the opportunity to be present at the beginning of bitcoin. Therefore, the future of our world's financial system (and the world as a whole) lies in each one of our hands.

 I have an imperative proposal that I would like you all to consider with an open mind. I believe this proposal could reach consensus by the end of this year. (or as soon as people realize it's implementation is absolutely necessary, ie., huge backlog of txs.)

I believe that this proposal is very bipartisan, as it does not change anything immediately, until such time that it's improvements to the protocol are reflected by our blockchain.

I believe that we currently do not need larger blocks, because the average block is not even over 75% full. "If its not broken don't fix it"

However, I know that it is inevitable, as bitcoin continues to add users and liquidity, that we will need a larger block-size cap due to increased tx volume.

Therefore, my proposal is as follows:
 
Increase the block-size cap following a Fibonacci Sequence (1,2,3,5,8,13,21) dynamically according to the average full-ness (95-99%) of the last X amount of blocks. ( I am looking for input on the best compromise between quick resolution and required resolution)... ?

The Fibonacci Sequence is nature's mathematical equation for growth and it has proved infallible for all of life's existence. (It's actually really interesting, I suggest looking into it and keeping an eye out for it in nature)



Other proposals suggesting the "doubling" of the block size (Bitcoin XT) after set time periods show complete disregard for both the complexity of bitcoin itself and the ability of nodes to remain capable of relaying blocks of larger size.

This idea appeals to me because it allows gradual, then geometric growth of the network instead of exponential growth, which will not remain feasible in the long run.

This BIP will allow the block-size cap to remain at 1MB until such time that it is necessary for the block-size to be increased due to a long duration of essentially full blocks in order to prevent the build-up of backlogged tx's on the blockchain. This will begin to happen when we see more than 7tx/s. Which, at the latest, will happen Q2 2016.

I hope that this gave you all some food for thought, and I look forward to reviewing questions, comments and ideas from the community.

Peace and love

Was

[hexafraction]

Seriously, do I need to ask this again, since you failed to answer it in the previous discussion we had? Why fibonacci? What benefit does it have for Bitcoin itself? Sure, the constant phi (1.618) is cool and mathematically elegant and applicable to nature, but bandwidth and storage capacity do NOT grow as a geometric sequence with ratio phi.

Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes. This is obviously unsustainable. You can't grow the block size after only a day of fullness. It invites too much abuse.

[jonald_fyookball]

Seriously, do I need to ask this again, since you failed to answer it in the previous discussion we had? Why fibonacci? What benefit does it have for Bitcoin itself? Sure, the constant phi (1.618) is cool and mathematically elegant and applicable to nature, but bandwidth and storage capacity do NOT grow as a geometric sequence with ratio phi.

Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes. This is obviously unsustainable. You can't grow the block size after only a day of fullness. It invites too much abuse.

Yes, there is really nothing magical about Fibonacci that makes it automatically applicable to everything. You can apply the same idea (average fullness of blocks) with some other fixed ratio (like average of last x number of blocks) times 1.5 and it would work just fine as well.  

[Was]

Seriously, do I need to ask this again, since you failed to answer it in the previous discussion we had? Why fibonacci? What benefit does it have for Bitcoin itself? Sure, the constant phi (1.618) is cool and mathematically elegant and applicable to nature, but bandwidth and storage capacity do NOT grow as a geometric sequence with ratio phi.

Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes. This is obviously unsustainable. You can't grow the block size after only a day of fullness. It invites too much abuse.

Hello! thanks for taking the time to reply.
I replied to your post on the other forum shortly after I wrapped up this post, but Im sorry if I didn't completely address your concerns.

I would ask you to re-read what I said, as I simply threw the 1 day correction period up in the air, for us to discuss.

What other mathematical equation do you suggest that allows initial gradual growth of the block-size limit??  It is my understanding that current proposals suggest a doubling 8MB block size limit every 2 years? How is this more desirable than a dynamic block-size of any formula?

"Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes"

This sentence hardly even makes sense, but it begs the general question: How would you plan to do this? I believe that in order to "dust" the blockchain with sufficient # of txs to raise the limit, it would take an increasing large amount of coin and, once you run out, you would be purchasing at an exponentially increasing rate and would just drive the USD/BTC price to the moon, further driving your ability to "abuse" my proposition into the ground.

I believe that we could benefit from a 2MB, then 3MB, then 5MB blocksize limit before we need to jump immediately to 8MB block caps.  Because, you know "bandwidth and storage capcity do NOT grow as a geometric sequence with ratio phi", but they don't grow in an exponential sequence either.

Nor would they need to, blocksize only needs to be dynamic in order to accommodate an ever increasing transaction volume driven by humans.

The reason the fibonacci sequence is so desirable for this issue and bitcoin as a whole is because it allows gradual growth of the blocksize limit while Bitcoin is in it's infancy, and then an accelerating growth in order to accommodate a transaction volume similar to Visa's and Mastercard's combined, while not disturbing the network

At the end of the day, I think its one of the best options, and if everyone is so against an infallible method of growth being applied to the bitcoin blocksize limit, go ahead and implement a static increase of 1.5x each time. You can keep your 1.5MB blocks.

Was

[jonald_fyookball]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect. 

[Was]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect. 

I see your point, so why can't this issue be handled before it becomes an issue and a threat to the network. Lets pass a BIP that allows Bitcoin Core to know when the network needs the additional headroom in order to make space for legitimate transactions.

Nothing has to change immediately, we can still keep the 1MB block in the short term, but when we begin to exceed 7tx/s on a daily basis, our clients and the miners will accept a 2MB block, until that threshold is reached, and so on and so forth as it gets exponentially more impossible for the network to be manipulated by dusting, and even if it was, it would be unsustainable.

If we can't reach consensus with a damn block-size increase when all it takes is math to understand the need for raising the ceiling, how will the community ever be able to deal with anything that pops up? Something that we didn't know was going to happen? I think we need a way for the community to vote on the blockchain for BIPs, and then maybe we'll be better able to deal with things like this...

Was

[hexafraction]

Seriously, do I need to ask this again, since you failed to answer it in the previous discussion we had? Why fibonacci? What benefit does it have for Bitcoin itself? Sure, the constant phi (1.618) is cool and mathematically elegant and applicable to nature, but bandwidth and storage capacity do NOT grow as a geometric sequence with ratio phi.

Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes. This is obviously unsustainable. You can't grow the block size after only a day of fullness. It invites too much abuse.

Hello! thanks for taking the time to reply.
I replied to your post on the other forum shortly after I wrapped up this post, but Im sorry if I didn't completely address your concerns.

I would ask you to re-read what I said, as I simply threw the 1 day correction period up in the air, for us to discuss.

What other mathematical equation do you suggest that allows initial gradual growth of the block-size limit??  It is my understanding that current proposals suggest a doubling 8MB block size limit every 2 years? How is this more desirable than a dynamic block-size of any formula?

"Also, if I spammed for 15 days, I'd get the block size up to a gigabyte every 10 minutes"

This sentence hardly even makes sense, but it begs the general question: How would you plan to do this? I believe that in order to "dust" the blockchain with sufficient # of txs to raise the limit, it would take an increasing large amount of coin and, once you run out, you would be purchasing at an exponentially increasing rate and would just drive the USD/BTC price to the moon, further driving your ability to "abuse" my proposition into the ground.

I believe that we could benefit from a 2MB, then 3MB, then 5MB blocksize limit before we need to jump immediately to 8MB block caps.  Because, you know "bandwidth and storage capcity do NOT grow as a geometric sequence with ratio phi", but they don't grow in an exponential sequence either.

Nor would they need to, blocksize only needs to be dynamic in order to accommodate an ever increasing transaction volume driven by humans.

The reason the fibonacci sequence is so desirable for this issue and bitcoin as a whole is because it allows gradual growth of the blocksize limit while Bitcoin is in it's infancy, and then an accelerating growth in order to accommodate a transaction volume similar to Visa's and Mastercard's combined, while not disturbing the network

At the end of the day, I think its one of the best options, and if everyone is so against an infallible method of growth being applied to the bitcoin blocksize limit, go ahead and implement a static increase of 1.5x each time. You can keep your 1.5MB blocks.

Was

You still haven't understood the point of what I've been saying. I have made it clear that I do not support a static increase. However, a growth factor capped at 1.6 per DAY is VERY excessive. What's wrong with requiring a week or a month of full blocks to raise it, and only raising at a 33% or even 10% increase?


Your idea is a sound restatement of already discussed concepts, but your specific parameters (increase factor and time needed for an increase) are absurd. Unless there's something mathematically beneficial in the Fibonacci sequence that there isn't in a normal geometric sequence, the 1.618 growth factor per day is absurd.

[solex]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect.  

Correct.
I spent all of 2013 and 2014 in the dreamworld that Bitcoin's best and brightest devs would come up with a cool elegant solution that would allow volumes to grow with demand while fending off the worst of spam, and encouraging organic fee growth.
The only obvious dissenting voice was Peter Todd who said right at the start that the limit shouldn't change. Well, for all that time I disliked him because of that message but now I respect him for being honest up-front and calling it how he saw it.
I think he is wrong and my very long experience in IT dealing with systems which approached many types of limits gives perspective (not systems like Bitcoin - but the principles are universal).

The block size limit will seriously cripple ecosystem growth, and trying to force fees higher too soon is a crazy schoolboy error which will hand precious ecosystem momentum to alternative crypto.

There is little point in coming up with new proposals although there might be mileage in seeing if Pieter's proposal can be modified. If his unnumbered BIP was to begin ramping at Jan 2011 instead of Jan 2017, and in the 20-25% range then this might be enough long-term.

[jonald_fyookball]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect. 

I see your point, so why can't this issue be handled before it becomes an issue and a threat to the network. Lets pass a BIP that allows Bitcoin Core to know when the network needs the additional headroom in order to make space for legitimate transactions.

a very sensible idea and i agree, so it's baffling why they don't.

[Was]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect. 

I see your point, so why can't this issue be handled before it becomes an issue and a threat to the network. Lets pass a BIP that allows Bitcoin Core to know when the network needs the additional headroom in order to make space for legitimate transactions.

a very sensible idea and i agree, so it's baffling why they don't.

"They"?

[jonald_fyookball]

It's not an argument worth having because this (whether to increase blocksize based on a deterministic formula or whether to increase it as a function of usage, or any such combination) doesn't appear to even be the issue at hand. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect. 

I see your point, so why can't this issue be handled before it becomes an issue and a threat to the network. Lets pass a BIP that allows Bitcoin Core to know when the network needs the additional headroom in order to make space for legitimate transactions.

a very sensible idea and i agree, so it's baffling why they don't.

"They"?

we need consensus, those that are against a block increase are the ones causing an issue IMO.

[tsoPANos]

This is the most silly proposal I have ever heard.
Fibonachi blocks will skyrocket within a few years.
If it matches some nature models it doesn't mean we'll implement.
It's useless...
Incredibly silly...

[DumbFruit]

.. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect.  

Correct. ..

we need consensus, those that are against a block increase are the ones causing an issue IMO.

Argument from Ignorance. Don't bother to learn why the other side objects, they're just insensible.

https://en.wikipedia.org/wiki/Argument_from_ignorance

[Was]

This is the most silly proposal I have ever heard.
Fibonachi blocks will skyrocket within a few years.
If it matches some nature models it doesn't mean we'll implement.
It's useless...
Incredibly silly...

They will skyrocket because bitcoin use will skyrocket. Paypal alone handles around 115 transactions per second. Bitcoin will eventually have to sustain over 100,000 tx/s, so why not set a cap in small, then large increments (fibonacci sequence) in proportion to the need of the network, but prevent it from outpacing Moore's law with a time-based equation?  At such point free market takes over and provides further incentive to develop nodes and miners and bandwidth that would support larger transaction volumes and therefore, a greater amounts of tx fees.

[hexafraction]

You are still dodging my question. Why is 1.618 per day reasonable? I'm just heading unrelated crap about 'patterns in nature' that have nothing to do with computational resources or txn volume.

What is wrong with waiting a week or month for growth? Your current rate is excessive. I'm not buying the claim that we even have a chance of needing terabyte blocks within only one or a few months. I favor something more reasonable such as 1.5 factor per week at most.

[jonald_fyookball]

.. It seems that certain people are against ANY block increase in short term, no matter how sensible or perfect.  

Correct. ..

we need consensus, those that are against a block increase are the ones causing an issue IMO.

Argument from Ignorance. Don't bother to learn why the other side objects, they're just insensible.

https://en.wikipedia.org/wiki/Argument_from_ignorance

My argument isn't based on ignorance simply because I don't provide a complete dissertation and demonstrate mastery of the subject matter in my post.  I have seen some of the arguments but can't say I'm convinced.  Letting the transaction bandwidth become intentionally filled to promote a market for fees is at least a principled reason to eschew a temporary increase, but is no doubt dubious, or at the very least, debatably foolish. 

[Kazimir]

+61% per day is absolutely excessive.

Currently, XT does +100% (i.e. doubling the block size limit) every 2 years. I think it's too much. And too static, why not a more flexible, adaptive approach (like the difficulty).

[Was]

You are still dodging my question. Why is 1.618 per day reasonable? I'm just heading unrelated crap about 'patterns in nature' that have nothing to do with computational resources or txn volume.

What is wrong with waiting a week or month for growth? Your current rate is excessive. I'm not buying the claim that we even have a chance of needing terabyte blocks within only one or a few months. I favor something more reasonable such as 1.5 factor per week at most.

I never solidified a 1.618x increase day over day, just threw it up in the air, however, I don't see why if the network demanded it we couldn't implement it.

 If Bitcoin use doubled within a week because of some event similar to Greece, we need to ensure that Bitcoin remains dynamic enough to handle such genuine influx. Or risk giving a bad impression on new-adopters.

I don't see the point in getting caught up in all of this, really. We can always implement an equation on top in order to prevent transaction volumes from out-pacing Moore's law, and introduce as a result, a fee market, but in the end : Such a market would further incentivize people to develop capable hardware of handling the excess transactions that the network would otherwise be incapable of handling.

If there's money to be made, it will be made. Just look at computers, there was a ton of money to be made, so we now carry the decade ago equivilent of super computers in our pockets... casually.

And on the other hand, there isn't much money to be made in developing alternatives to the current banking system (short term), or Gasoline.... or Cures instead of life-long treatments. So it doesn't happen.

Guys, please try to move away from the argument-like vibe of this topic, we're all trying to achieve similar things, so we should be collaborating and discussing, not bashing other people's ideas or notions. If someone is factually wrong, feel free to correct them, but stay out of your feelings please.

[funkenstein]

Nice pics   However nobody has argued for quantitative relationship between available block size and bitcoin use / user base / anything that might be called BTC growth.  So it's kinda hard to comment on an incomplete argument.