Bitcoin Forum

http://www.coindesk.com/purse-proposal-touts-extension-blocks-bitcoin-scaling-solution/

Just sharing, I know its been thrown around before but seems to beginning to get praise, Extension Blocks!

Bitcoin Core is saved ;)

We need segwit to do extension blocks. They've been a part of segwit for a long time.

I read they are ready for it, I understand why certain mining pools, Core, seg etc are in favour of such a thing as Extension Blocks but why would BU say they like it?

On the surface this seems like a sane approach compared to the core Segwit or a hard fork.

i think minners might seriously reconsider segwit, if they could get extension blocks shortly after.
my god, imagine 4MB effective coreblocksize + extension blocks + LN
bitcoin is going to scale up Up UP!!!!!!!! like a MOFO

I think the problem with most miners is that they don't understand with the future holds with all these changes coming to the space, fear of unknown

Sounds like it could have some promise.

My concern would be the timeline.

Is there any code written?  How's this
different from segwit?  Is there a whitepaper?

Remember, this debate
Is just as much about economics as it is
about technicals.

If the core devs have said they want full blocks
and a fee market, how much scaling will they
be willing to allow?
 
The last thing we should do is waste 2 more years coding something
and then get stuck in another debate on how many mb go in the extension block.

I say HF to 2mb NOW and that will buy us some time while
extension blocks are researched.

Yes the code has been written and is well tested. It's commonly called segwit. Segwit makes Bitcoin extensible and allows for extension blocks, sidechains and LN.

LOL,  :D

No you don't need segwit.

You need => https://bitcointalk.org/index.php?topic=1846625.msg18375513#msg18375513
Clif High's solution for Bitcoin to scale to infinity.


 8)


FYI:
BlockStream & The Banking Cartels need Segwit so they can make LN work, and rape you & your Children's Children by Financial Domination.
Just like they have the rest of the world with their fractional reserve system.
Segwit/LN is slavery , Wake up before it is too late.
Your Choice.

FYI2:
At the present time a minor block size increase or faster block speed would easily fix the current issues.
Segwit is the Trojan Horse that will infect the blockchain.

the practical application of the idea you link is Extension Blocks.

Why not just implement segwit?
Many solutions are arising and still none of it is being implemented. Why?
Because  it got stuck in the debate stage.
In many years to come many solutions will once again arise then it will be forgotten again like the others.

If only BU and the miners were not blocking advancement

Funny , no one remember the extension blocks until Clif's idea came out.
You're a good little minion and believe everything Core tells you.  :D
https://s-media-cache-ak0.pinimg.com/736x/48/bd/3f/48bd3f6e928d7cb4b8d499cb0f96b8a8.jpg


 8)

If by advancement , you mean blocking the ability of the Bankers to enslave the world and its children from a Fractional Reserve Crypto system.

Good Job, BTU.  :D


 8)

Nice try.  SegWit has an effective block size of 2mb anyway, and doesn't require a hard fork which would cause the price to tumble in the short term at least and be detrimental to the network (especially with continued opposition from any major Core supporters).

now that is a good proposal, which will not lead to any kind of more centralization, and if you think about it it's not very different from dynamic block

without directly change the main one, i'm more curious about how it work more precisely, they are a bit too vague about this or they can't explain it well

also there is some confusion if it need segwit or not? can someone more competent explain well this part? and why it need segwit?

Segwit does not extend block. This node only reduces the size of the transaction which means 1 block can include more transaction than usual. The size is still 1mb, no change at all. But it is really nice if they agree to activate segwit. I am looking forward to seeing litecoin activate their segwit. If that happens, bitcoin will follow litecoin's path

is good information if about block size issue is solved
so bitcoin unlimited and fork issue can die, and not effect again about bitcoin price

Fractional reserve banking is in any case unavoidable, it has nothing to do with crypto.  From the moment there's something that has an established value but may have real or fake, perceived difficulties in handling directly (complexity, transaction price, slowness, security, ...), nothing can stop an entity to give out an IOU of that asset that can acquire more fluidity than the original.  In order to keep that IOU in 1-1 relationship, you don't need a full reserve, you only need enough reserve to settle differences with other IOU emitters.
Read the first few chapters of "Treatise on Money" by Keynes.  He explains this (it is his argument of why banking must be regulated, but that's another story).

The only way to avoid fractional reserve banking, is to have perfectly liquid money at no difficulty of usage, at no security risk, immediate and without cost.  From the moment you introduce one of those frictions, you will give an opportunity to fractional reserve banking. 

Hell, bitcoin itself is fractional reserve banking for fiat in a certain way !  You hold bitcoins which you think are going to be "re-imbursed for fiat" in exchanges.  Only if you use bitcoin in a closed circuit, you are not using bitcoin as fiat fractional reserve banking !  But most people think of bitcoin as "representing fiat value" and they look at coinmarketcap to see how much fiat they "hold" in their "bitcoin account", thinking that at any moment, they can "withdraw fiat" for it.

That would be very good if that is true but if not then we will still be suffering from slow confirmation and many will shift to ethereum. But aside from increasing the blocksize I hope segwit could do something about the miner fees which is very high right now. But anyway if 4mb effective blocksize will be implemented successfully then it will be time for bitcoin to be used in shops not only online.

I like this part

Hard fork will be disaster for bitcoin network. Haven't gone through all technical details behind this new solution but if it is only soft fork than it is far better than BU.

So here is a question, in order to move forward with Extension Blocks, what needs to be done and who needs to agree?

Why will it be a disaster if everyone is in consensus about a HF?
 

because that consensus will never happen.

purse and bitpay..

hmm
DCG again.. hmmm

though extension blocks is another backdoor implementation.. it still does not address the issues of native key users.

IF everyone agrees then it can happen, so this is not a reason.

extension blocks arent going to get passed, the devs seemed to be pretty split on them and feel like they are sacrificing security for community consensus.

The problem with them (miners) is the fear to lose incentives since with SegWit, they think it will take away lots of transaction fee from them and LN will just kill them.  Due to personal preservation, they gone blind on what is the best approach to bitcoin scaling.

Whats funny is BU plans to add LN too.

if somthing "kills" miners its toxic to bitcoin...

if bitcoin is secured by the incentive to mine 2.5BTC every time minutes then it is very INSECURE

as mining rewards drop in half it becomes absolutely mission critical to maximize fee revenue.

The miners are screwing themselves over by proving to be bad actors.

best way to do that is probably increase the user base, because increasing fees per transaction
is greatly limited due to competition. Iow, why would I (or anyone interested in saving money)
pay more with Bitcoin when another method is cheaper?

the best way to do it is using some basic economics 101
making blocks that have a size which creates some fee pressure, will net alot more fees in total.
even not considering altcoin competition its clear that a any static blocksize limit will fail to produce a healthy fee market.
1MB = fees pressure so high that poeple simply cant TX
50MB = fees pressure so low that everyone TX can get in the block for free.
both do nothing to maximizes fees/block, and so both lead to less security for bitcoin.
we NEED a dynamic blocksize limit which will yield the most optimal  pressure fees, or bitcoin will be very insecure in the future, and therefore useless and therefore worth 0.

It actually does not need a block limit. Miners can dynamically produce their own block sizes to cope with mempool transaction pool demand pressure and by artificially delaying lower paying transactions fees. If bitcoin price rockets out of the solar system, big blocks full of 1 sat transaction fees and no coinbase reward will work perfectly fine.

Seems like there could be an enforced minimal fee to prevent spam and past that a sliding schedule that controls how soon the transaction gets put in a block.   While it would be artificial it could increase fee revenue.    After all why should buying a car and driving it off the lot be the same as buying a cup of coffee?   If you needed quicker confirmation you should pay a higher fee for that.

There are several dynamics that can be played around with:
1) Size of transaction --- basically what we have now  ... if you use a lot of data it costs more
2) Speed of the transaction --- Seems logically that there should be a fee schedule and low fee payers simply wait longer even if the wait is artificial.
3) Amount of the transaction --- Maybe it does make sense to have some very small fee the scales on on transaction size.  Say something like 0.01%  <== Okay I know this is evil     

You make a lot of sense but i'm not sure we need a blocksize limit to do this. 
I'm curious if you've read Peter Rizun's whitepaper about the fee market sans blocksize limit.

yes i read it.
I reference it in my essay:
https://medium.com/@adamstgbit_25789/bitcoin-unlimited-to-bring-stability-to-bitcoins-fee-market-6b5a4f882fc0

I'm slightly confused on your position because it sounds like you agree with Peter that EC can solve the issues...on the other hand, Peter
is advocating that we shouldnt and neednt have a limit that is below market demand, while you are saying blockszie cannot outpace adoption.
I'm not sure if there's an actual disagreement or you're just describing a different property of the natural fee market.

 

and many will shift from ethereum to tezos due to dictatorship and centralization not to mention the hardfork problem that leads to ethereum classic and also the unsecure javascript unverified code

I agree that there is a natural cost to miners for including TXs in blocks. But i do not think this will be the limiting factor in determining Blocksize. Peter says we dont need a block size limit at all because miners won't produce blocks that have a bad risk/reward chance-to-orphen/fees.

I'm saying Peter's theory determines an upper limit, but blocksize won't get close to this limit at all because miners will set a limit which will create fee pressure. This will be like 2MB with today level of adoption and maybe 2.2MB 6 months from now. ( obviously no where near peters limit)

miners are in the business of selling blockspace
they can:
 A) offer lots and lots of blockspace for cheap ( near peters cost per byte )
or
 B) offer a very limit amount of blockspace at a high price

which will yield the more revenue?

well... this is a classic economic problem with a straightforward way of determining how to price your good ( or in this case how to size your blocks )

https://cdn-images-1.medium.com/max/800/0*KG5PUsKo6YLvQYih.

this optimal amount of block space the miners should be offering, depends on fee paying TX demand, which isn't limitless, and very much tied to adoption.

miners are incentivized to set an Excivies Blocksize (EB) which doesn't go over this optimal block size limit.

miners won't create massive blocks unless there is Massive TX demand.

TX demand will take YEARS to get anywhere near needing 16MB blocks.

https://s11.postimg.org/7k7jutvhf/Untitled.png
i made a few edits to the stock graph to better illustrate what i'm saying

as block size goes up fees go down, but fee revenue is at its highest at 2MB, because science.

Very nice Killerpotleaf, do you happen to teach?

i agree, i intuited what you are saying before, but credit to you for writing about it!

but obviously there is also the constraint of what other
payment types besides Bitcoin are offering right?

we can safely assume we are not in competition with paypal or visa, crypto is more compelling than them on many levels, if fees end up a little higher than paypal its not going to impact our success much. however other crypto are in direct competition with us, and while they are subject to the same Hardware limitations ( internet speeds, storage etc.) they are not bound by any silly self imposed limits like 1MB blocks.
we have seen bitcoins market share drop drastically now at scary levels, because of our inability to upgrade.

and off-chain scaling ISN'T GOOD ENOUGH...

on-chain scaling isn't important  because" lower fee = more adoption ", on-chain scaling is important because "  higher miner revenues = more security ".

what's best for miners ( and by extension blockchain security ) is a dynamic blocksize which can adapt itself to the ever growing fee paying TX demand.

if bitcoin doesn't upgrad on-chain, ether or dash will, and 30 years from now dash's blockchain could be more secure than bitcoins!

There is a ton of crazy business interest schemes with SegWit.  Why not do what has always been done?  Either make blocks 1.2MB or be big time like LTC and go to 2MB.  

It's a simple thing that has got twisted ...

After much consideration, I have finally decided to point my miners to a BU pool. 

  

30 years?  if we don't upgrade, we might not last 30 months.  things move fast in the cryptoworld which is only 7 years old.  look what happened in less than a year with the marketshare...95% to 66%. 

Its the potleaf, he meant 3 ;)

I agree, should fee's, time and functionality/sustainability for parties as a whole continue to move in the wrong direction, 30 months may even be a stretch.

the thing the network effect is strong, and segwit will buy some time, then LN will buy more time, but after a few more block reward halvings, if the onchain TX fee revenue isn't up significantly...

we can continue to go down the worng path for a long long time, avoiding the inevitable. thats why i'm thinking the consequences are years away.

altho admittedly 30years was a bit much.

sorry but your and blockstreams high school level understanding of economics is the failure

if a block can only handle ~2500 native transactions
then this is the result:
if blockreward * current bitcoin price leaves a deficit of say $2500 (EG bitcoin price went from $1200-1000 = $15,000 to $12,500)
then that means fee's need to be ~$1 to combine to make up for that deficit.
if blockreward * current bitcoin price leaves a deficit of say $7500 (EG bitcoin reward went from 12.5-6.25 = $15,000 to $7,500)
then that means fee's need to be ~$3 to combine to make up for that deficit.

the result is a LIMITED amount of people (due to only ~2500tx native key users) will get peed off with having to 'just pay more'
the funny part is
segwit with its 75% discount. means although there 'could be' 4500tx segwit users per block IF block was 100% segwit key use. the fee's are then at best 50cent-$1.50 for segwit.
but that means $2-$6 for native key users(undiscounted)


however.
we have a couple years before needing to increase onchain capacity to ofset the next block halving (without having to rely/hope for the bitcoin price to save the day again)

EG imagine the price was stagnant for the next 20 years at $1,200.. and you will see that trying to squeeze more blood out of 2500 people every 10 minutes, wont work.

more transactions per second onchain = more people sharing the burden = less of a burden overall

segwit wont 'buy some time
bcause the network effect of users feeling the change. does not occur due to just activating segwit. it occurs by people voluntarily moving their funds over to segwit keys.

the burden of moving 46mill UTXO will create MONTHS of constant mempool bloat of native key users moving across. or years of slow switching over of keys just to even try attaining the ~4500tx a block hope.

the other failure of segwit is by offering a 'discount' it make the fee continue to rise, which then begins to impact native key users 4x(due to segwits 75% discount). EG a 10% segwit fee rise is 40% native key rise

the end result is native key users will eventually be paying more to move their UTXO than their UTXO is worth.
we are already seeing that now.
which just results in people not moving their UTXO and creating dead UTXO set of millions of what then becomes 'dust' due to being under say $6 value.

However as the number of transactions changes the size of the blocks would need to change to keep the price optimal.   In effect you are just determining what optimal block size is for the current number of transactions.   The only way something like this would work is if the block size was changing almost every block.   

this is correct!
optimal block size is dependent on TX demand, which is known to be going up, in a crazy unpredictable manner.
and so your right every single block would have a different "optimal size"
but thats not the point... the point is there is an optimal size given X amount of demand.
Miners ( and Non-mining nodes too , mine is set for 1.5MB right now :D) will need to smooth out the day to day ups and downs and approximate TX demand for that period of bitcoins history, and set their EB appropriately, each of them will do this individually, and collectively a blocksize limit will be determined through EC.

i think we'll see nodes always kinda lagging behind, and we'll mostly always have blocks that are slightly smaller than what the optimal size should be for that time. but thats a hell of alot better then being 2-4x above or below optimal size as a 1MB or 8MB static limit is sure to produce.

So that is what BU allows but SegWit wouldn't?   If that is the case it makes SetWit a really hard sell. 

segwit is an EASY sell, i'd buy into it in a heartbeat. IF its blockweight/effective blocksize thing is removed...

The error in this plot is that you think that fees are elastic.  Fees are VERY INELASTIC, meaning your blue curve goes to the sky on the left.  If only 10 transactions are possible, people will be paying HUGE fees in order to obtain one of the 10 places.  The guy wanting to transact 10000 BTC for a huge cocaine deal will need to get the transaction through, no matter what, or they'll get him, so he'll pay 200 BTC of fee.
10 of these transactions, and you earn 2000 BTC per block in fees.  From the moment that you relax, fees will drop drastically per transaction, much more so than your straight line suggests.

And then you see that the optimum is on the left side of the plot, because fees rise much faster than 1/block size, and drop much faster with block size.  

Disagree entirely.  i use bitcoin as a paypal alternative for remittances.  anything that
saves business owners money will be "a thing".  This is the easiest use case.
Merchant adoption requires more user adoption but saves business even more
on payment processing.

i think its hard to incress fee pressure without increasing TX demand and simply limiting blocksize.
at one point fees are so high a lot of potential TX can't happen anymore, and this removes some fee pressure, so the higher you push fees the harder it will be to incress fee pressure, its kinda like rocket science, you need more fuel to launch more weight, but the fuel itself weighs somthing so you need more fuel to add more fuel.

its hard for me to entirely disagree, because not core-fan-boy, but bitcoins usefulness is not only as a medium of exchange, so while we could stand to benefit by being more competitive to paypal, we don't NEED to be. ( i say this but part of me doesn't agree at all, its what i heard others saying, seems to make some sense..)

Is sacrificing bitcoin's utility in the name of keeping full node cost very low, worth it? I think not! but some would disagree.

core wants prunned, no witness modes.. and a cesspit of downstream nodes that dont get all tx data and dont relay segwit transactions
 which is a lot more damaging to the node count than anything else..

dynamics does not mean 8mb by midnight, nor 8gb by midnight.

nodes put in their useragents to what they can COPE with. and pools only make blocks to the amount the majority can cope with. thus no harm to nodes

we are not in the 2009 days of:
3g mobile internet/ADSL landline
Raspberry Pi1 min specs
libsec validation

we are in 2017
4g mobile (approaching 5G)/fibre landline
Raspberry Pi3 min specs
libsecp256k1 validation...

meaning the 'speed' of validation propagation etc is better by between 5-20x compared to 2009's '1mb safe'
its been estimated that 8mb is safe for the network. even core know this and went with their 4mb 'supersafe' weight.
so can we give up all the foolish excuses of needing 1mb as a 'safety barrier'

most likely the bottom 15-25% of EB values will set the "limit"
so, only a few nodes and miners need to see the profit in keeping blocks small to create some fee pressures, which will maximise fee revenue, and thus add extra incentive to add more security to the blockchain.
So while you think blocks will be as big as most can COPE with, what will actually happen is blocks will be as big as fee TX demand can push it up to.
nodes and miners need to understand the very real benefit of making sure blocks fit just right given X amount of TX demand.

IMO the appropriate level of fee pressure is when your relatively sure a 10cent fee will get you in the next block, but users put 20cents just to be sure.

do you disagree? do you want blocks to be big enough so that fee pressure is non existent? do you buy the idea that there is "unbound" 10cent fee TX demand?

the OPTION of so called witness nodes, as i understand them, sounds good. similar to SPV wallets, like what i use on my phone (android bitcoin wallet) that doesn't the full blockchain on hand, right? please correct me if i'm wrong.

but forcing that, making it such that running a full node is unmanageable for the average techy geek... that worries me.

and yea, the average techy geek would have the best home user internet they could get, and and excess of computing power and storage space. i do, and i'm not even employed right now.

ok, except the internet part... stupid middle of no where... but hey, my phone doesn't have a data limit. it mostly works.

there are FAR better ways to allow full-nodes to run on home PC without losing witness data
nodes can limit number of connections.
or go as far as enable blocks only mode.

if you're going to loss the witness data, you might as well go full SPV

well, yea. but people are dumb and lazy and...

ok, i got nothing.

Demand comes from people "locked into" bitcoin, and NEEDING to transact (the only thing bitcoin is good at).  If you HOLD bitcoin, the only reasonable thing you can do with it, is transact, sooner or later.  That's the demand.


Uh, no, that is exactly the pressure.  All TX have to happen.  If they can't happen, or aren't needed to happen, then the bitcoins involved in them are useless. 

If you mean, however, this will limit bitcoin's usage to only big transactions, you're perfectly right.  Bitcoin with small blocks is not a payment system that can be used for small transactions, only for big ones, at low transaction rates. 

In as much as this has an incidence on the bitcoin market value, this will probably not even matter.  If the fees increase 10-fold, the market cap must fall 10-fold before this starts to be negative for the miners.  So squeezing fees out of users, even if this crashes the market price, is still interesting for miners if the inelastic wall of "I absolutely need to get my big transaction through" is steeper than the price crash.

you have to be shitting me....
most poeple think 90% of the bitcoin TX are spam
they complain that "spam" is creating fee pressure.
remove the "spam" and you think we'll fill block with nothing but Big TX willing to TX at anycost?
ya right,at that point poeple are sending there hoards to poloniex so they can buy the digital cash they were promised.

Spam is nothing else but a covert way of the miners to lower the effective block size.  It costs them, on average, nothing, if they spam according to their own hash rate (they pay the fees to themselves).  If miners spam, it means that they are of the opinion that the blocks are already too big for their optimum profit, and need to lower the available user block room.  Note that miners are the only entities that can spam without a cost.  All others that would spam, have a cost, the fees of the spam.  Miners don't, because they get them back.

Note that spam is a self-regulating block size limiter that settles dynamically between miners.  A miner that spams more than his fellows will pay net fees to his fellows ; one that spams less will receive fees from his fellows.  So spam is only interesting as a "common good" for miners if they can settle on the right amount of spam to squeeze the maximum of fees out of the users.

This is BU, but on the low side.  

Bitcoin's game theory is one big hell of a clusterfuck.

You mean, miners send many transactions and it clogs up the network?  Is this what people mean by spam?  


Not really.  I have sent one transaction with a higher fee that was picked up by a miner on the pool I point towards.  I don't think I really got much of it back on a PPLNS payout and 0.04% share.


This could be true for some.  However, I received daily payouts and would put them into four different wallets at payout.  Two different ones for electricity (home, hosted - not cloud), one as a "hot" wallet and another to cold storage.


If you use the same origin and destination address, anybody can send 20 transaction without a fee and one with a large fee.  If they are to and from the same addresses it picks up all the transactions.  However, most aren't using bitcoin as frequently as miners to and from the same addresses as frequently as miners.
 

How do you make a 'junk' or useless transaction with Bitcoin?  I don't understand this concept of spam.  Perhaps there is either a change in the quantity demanded - or we are entering a totally new demand curve.  I think as we move from innovators to early adopters using BTC, it's going to move out the demand curve and not just change the quantity demanded.

New name, same thing. 

They have not because they think they are going to lose money with it, it is that simple, but what they don’t realize is they could earn a lot more by allowing bitcoin to scale and get a lot more users, maybe they will suffer in the short term but in the long term they will be better off.

So where are Extension Blocks now? Are they looking to use extension blocks with segwit?