Please do not change MAX_BLOCK_SIZE

[xcsler]

What the world desperately needs is a better store of value. We need a way of combating the inflationary policies of central banks worldwide.
Although gold has historically served this role, with the advent of the paper market, gold's true value has been obscured and manipulated by bankers. Bitcoin provides us with a new tamper-resistant way of saving our purchasing power over time.
Any increase in the blocksize will marginally decentralize and weaken this store of value property.
Personally, I'm all for storing my retirement savings "on the chain" and doing small, low-risk, day to day purchases "off the chain".

I'd love to hear what guys like Tuur D., Andreas A., Voorhees, and Surda think about the blocksize limit. I believe that ultimately this is more of an economic issue than a technical one.

Let's assume the transaction fees go up to $20 and Wordpress, Mega, Silkroad and all BitPay's clients sod off to LiteCoin or some other coin without this restriction, as they obviously would, so you were no longer able to actually buy anything with your Bitcoins.

Do you think the value of your Bitcoins would go up or down?

I'm thinking it would go down, and more than marginally.

There would be an exchange rate between BTC and LTC. Because of BTC's properties combined with a commitment to a limited blocksize I would guess that it would maintain its purchasing power much like its analog equivalent, gold.

[edmundedgar]

What the world desperately needs is a better store of value. We need a way of combating the inflationary policies of central banks worldwide.
Although gold has historically served this role, with the advent of the paper market, gold's true value has been obscured and manipulated by bankers. Bitcoin provides us with a new tamper-resistant way of saving our purchasing power over time.
Any increase in the blocksize will marginally decentralize and weaken this store of value property.
Personally, I'm all for storing my retirement savings "on the chain" and doing small, low-risk, day to day purchases "off the chain".

I'd love to hear what guys like Tuur D., Andreas A., Voorhees, and Surda think about the blocksize limit. I believe that ultimately this is more of an economic issue than a technical one.

Let's assume the transaction fees go up to $20 and Wordpress, Mega, Silkroad and all BitPay's clients sod off to LiteCoin or some other coin without this restriction, as they obviously would, so you were no longer able to actually buy anything with your Bitcoins.

Do you think the value of your Bitcoins would go up or down?

I'm thinking it would go down, and more than marginally.

There would be an exchange rate between BTC and LTC. Because of BTC's properties combined with a commitment to a limited blocksize I would guess that it would maintain its purchasing power much like its analog equivalent, gold.

That's a very - ahem - optimistic take.

It would be interesting to hear what those various people you mention think about it, but for example Falkvinge's piece on the target value of Bitcoin is specifically premised on being used as a payment system:
http://falkvinge.net/2013/03/06/the-target-value-for-bitcoin-is-not-some-50-or-100-it-is-100000-to-1000000/

[DeathAndTaxes]

What the world desperately needs is a better store of value. We need a way of combating the inflationary policies of central banks worldwide.
Although gold has historically served this role, with the advent of the paper market, gold's true value has been obscured and manipulated by bankers. Bitcoin provides us with a new tamper-resistant way of saving our purchasing power over time.
Any increase in the blocksize will marginally decentralize and weaken this store of value property.
Personally, I'm all for storing my retirement savings "on the chain" and doing small, low-risk, day to day purchases "off the chain".

I'd love to hear what guys like Tuur D., Andreas A., Voorhees, and Surda think about the blocksize limit. I believe that ultimately this is more of an economic issue than a technical one.

Let's assume the transaction fees go up to $20 and Wordpress, Mega, Silkroad and all BitPay's clients sod off to LiteCoin or some other coin without this restriction, as they obviously would, so you were no longer able to actually buy anything with your Bitcoins.

Do you think the value of your Bitcoins would go up or down?

I'm thinking it would go down, and more than marginally.

There would be an exchange rate between BTC and LTC. Because of BTC's properties combined with a commitment to a limited blocksize I would guess that it would maintain its purchasing power much like its analog equivalent, gold.

I don't.  LTC (or some alternative) also would serve as a store of value.  So if given the choice between:

a) store of value with high transaction costs, low liquidity, and difficulty directly spending
vs
b) store of value with low transaction costs, high liquidity, and ease directly spending.

which is likely to be more popular?  Given the friction between crypto-currencies is very low (compared to crypto to fiat) I would expect a massive decline in the value of a relative to b.

TL/DR Bitcoin's value as a store of wealth is linked to its utility as a wealth transfer mechanism.  If the later if limited so if the former and the price will reflect it.

[d'aniel]

Bitcoin can always be forked into a block size limited version if we find censorship is becoming a problem with larger blocks. Artificially limiting the block size now is creating a problem to solve a non-problem.

Wishful thinking.

If censorship was a problem, that would be because hashing power was under the control of authorities, who would simply use that hashing power to 51% attack the censorship resistant Bitcoin. Changing the PoW algorithm doesn't help either, as new coins haven't been able to survive 51% attacks by people having fun, let alone focused attacks, for a long time.

We've got one shot at getting a decentralized PoW-based consensus system implemented. Don't screw it up.

So announce the PoW algorithm switch/reimplementation of the block size limit well in advance, giving ASIC manufacturers enough time to get their chips widely distributed.  New alt-coins haven't been able to survive 51% attacks because they've been worthless, and therefore don't command a significant enough mining force.  It can be assumed that this Bitcoin fork would not be worthless. and would thus command a significant mining force.

I'm curious though, suppose all of the reasonable hypothetical difficulties with reimplementing a block size limit could be addressed, would you then switch sides in this argument?

[edmundedgar]

BTW one of the mysterious things about people wanting to optimize Bitcoin for ease of running a mining node, at the risk of losing in the marketplace to another currency optimized for cheap payments, is that some other coins, notably LiteCoin, are actually already much better optimized for what _they_ are worried about than Bitcoin is.

The big issue here is the proof-of-work, which is far more important to the economics of mining than the block size. If you want a radically distributed network rather than one dominated by a bunch of big pools or companies, you need the economics of mining to be a good fit for a cottage industry, rather than a capital-intensive commodity business with large economies of scale.

This isn't an easy thing to design, but one of the main features would be that you'd be able to mine competitively on hardware that you already own, using spare CPU cycles, rather than needing to buy a lot of fast-depreciating specialized hardware, which you then need to install and operate highly efficiently to get a good return on your capital. This is something LiteCoin's proof-of-work is actually designed to do, and Bitcoin's isn't. It's not clear whether their current approach will work in the long-term, but in Bitcoin's case, as someone said upthread, the ship has long since sailed.

[Peter Todd]

d'aniel: Of course.

This isn't an easy thing to design, but one of the main features would be that you'd be able to mine competitively on hardware that you already own, using spare CPU cycles, rather than needing to buy a lot of fast-depreciating specialized hardware, which you then need to install and operate highly efficiently to get a good return on your capital. This is something LiteCoin's proof-of-work is actually designed to do, and Bitcoin's isn't. It's not clear whether their current approach will work in the long-term, but in Bitcoin's case, as someone said upthread, the ship has long since sailed.

Litecoin uses scrypt in a way that can be implemented on FPGAs and ASICs in a cost-effective manner. It's harder to do than the Bitcoin PoW, but that's actually a decentralization disaster because if one player ever does make a ASIC for litecoin mining it'll be much easier for them to get 100% of the market. Bitcoin's SHA256 PoW is really easy to implement on an ASIC, so we've already seen 4 different ASICS implementing it.

[solex]

Let's assume the transaction fees go up to $20 and Wordpress, Mega, Silkroad and all BitPay's clients sod off to LiteCoin or some other coin without this restriction, as they obviously would, so you were no longer able to actually buy anything with your Bitcoins.

Do you think the value of your Bitcoins would go up or down?

I'm thinking it would go down, and more than marginally.

Exactly. It is wishful thinking that Bitcoin can be parked in a state of suspended animation with "just the right amount" of market-share until the blockchain becomes a replacement for Fedwire, and off-chain solutions of the "just the right type" for all non-bank users mushroom around it.

Market forces mean it will either soar and succeed or crash and burn. There is no safe middle-ground.

[d'aniel]

...
I'm curious though, suppose all of the reasonable hypothetical difficulties with reimplementing a block size limit could be addressed, would you then switch sides in this argument?

d'aniel: Of course.

Excellent.  I think it's important for the small blocks crowd to enumerate these hypothetical difficulties then, so they can be addressed.  Were you convinced by my response to the one you already brought up?

[tvbcof]

There would be an exchange rate between BTC and LTC. Because of BTC's properties combined with a commitment to a limited blocksize I would guess that it would maintain its purchasing power much like its analog equivalent, gold.

That's a very - ahem - optimistic take.
...

I can only speak for myself, and I don't think that it should be a big factor in this decision, but...

I will definitely be liquidating a fair fraction of my BTC holdings if I don't feel that the solution is on a trajectory to withstand very significant growth and very significant attacks.  It is in fact something I've been actively planning but a serious discussion about block size and off-chain frameworks has given me new hope.

Now, it is fair to say that I'm an unusual thinker in my system analysis.  I was buying with both hands in late 2011 based on analysis and intuition, and with a few exceptions it seemed like many/most of the people on bitcointalk.org were kind of on death watch.

[oakpacific]

If we scale up to the limits of a beefy desktop PC with a RAID array and >1Mbps network connection, we wouldn't be pushing anyone out from actually using the network.

Except from the ones who don't know what RAID array and >1Mbps network connection is.

In fairness, one does not really need to know what such things are to use them.

A much bigger concern in my mind is that one may need, say, 1Mbps of bandwidth which is free of disruption by commercial network carriers.  These players have little ability to act autonomously of government security policies even if they had a desire to do so.

I think it hopeful in the extreme to assume that any political leadership is going to stand idle by while the power inherent in controlling a currency system is eroded.  If distributed crypto-currencies never reaches the stage where it is a eats into existing monetary solutions and supplants some of their function then yes, we are probably safe from such attacks.  It is far from clear to me that such solutions will not achieve such a capability.

In the future, we may need a lot of ASICMiner style USB miners which can run all day at high hashrate, yet doesn't cost you anything noticeable in electricity, yes it will not give me a steady income of mining revenue, but should be a worthwhile investment to secure the network if the power cost can be driven to low enough that you can just plug it into your laptop and stop caring.

[Peter Todd]

...
I'm curious though, suppose all of the reasonable hypothetical difficulties with reimplementing a block size limit could be addressed, would you then switch sides in this argument?

d'aniel: Of course.

Excellent.  I think it's important for the small blocks crowd to enumerate these hypothetical difficulties then, so they can be addressed.  Were you convinced by my response to the one you already brought up?

Heck no. How on earth do you think you are going to manage to come up with enough funds to outspend potential government level attackers with a brand new ASIC and hashing algorithm, and on top of that, have people trust you that the effort is real?

[tvbcof]

In the future, we may need a lot of ASICMiner style USB miners which can run all day at high hashrate, yet doesn't cost you anything noticeable in electricity, yes it will not give me a steady income of mining revenue, but should be a worthwhile investment to secure the network if the power cost can be driven to low enough that you can just plug it into your laptop and stop caring.

I hypothesis that the competitive parties in mining will be 'the masses' who, like you say, don't use enough power on an individual level to consider, and large entities with a big enough footprint to harvest high value business intelligence.  The latter can subsidize their power costs.

New as Bitcoin is, we've already seen that power costs can influence mining effort, so it is not a wild theory at this point.  I do suspect that small mining operations will vanish unless they can monetize the data they can obtain and sell it up-stream to entities who can use it.  Servicing SPV clients or offering stratum server services may be one way to do this.

I strongly agree that 'ASICMiner style USB' devices in the hands of ordinary market participants are a powerful and positive thing.  I do have some qualms about buying sha256 ASIC gear myself or inducing others to do so however.  It may not happen this year like Kaminsky tried to claim (or ever), but it could be that a different and possibly varying proof of work algorithm is in the interest of a 'free' Bitcoin.

Relatedly, I've been musing about whether it might be possible to favor work done by demonstrably 'network edge' participants.  These players have the 'strength of numbers' which make the concept of voting based on choosing software work.  I hope that smarter persons than myself have been pondering that as well.

[d'aniel]

Were you convinced by my response to the one you already brought up?

Heck no. How on earth do you think you are going to manage to come up with enough funds to outspend potential government level attackers with a brand new ASIC and hashing algorithm, and on top of that, have people trust you that the effort is real?

The government level attacker would have to match that spent by the new miners, which is roughly proportional to the expected value of their block rewards, which is roughly proportional to the demand there is for the censorship-preserving fork.  Are you doubting this demand?  The effort can be trusted to be as real as the ASICs produced by chip makers, which would be being produced well in advance of the fork, and purely as a for-profit enterprise based on expected demand for the fork.

If demand for the censorship resistant fork is high, but the government level attacker is still willing to outspend the miners, well then the problem is completely orthogonal to the issue of block sizes.

And just to be clear, all of this matters only after 1) all governments across the world have successfully colluded to eradicate all non-hidden, censorship-resisting mining pools everywhere, and 2) mining pools aren't able to find any technical means to hide the location(s) of their machines, despite the funding available from parties (good and bad) with financial interest in resisting censorship.

Not to mention there are other, non-technical, ways to deanonymize miners: "Just register with us, the government, to collect your mining subsidy."

[piotr_n]

I think it's more likely that the community would let a fork go, rather than trying to beat governments on hashing power, resources, or (lack of) anonymity in order to run a node, or a miner...

People can adopt easily and they can easily switch to another branch, if they find an actual need to do so.
In fact, knowing life, I wish they didn't, but they probably will.
But I don't think it will be the end of the world.
If we managed to gather so much power to build this net, we can as well manage to move this money into the right branch, after a potential fork.
When the chain splits, the hashing power must also split - so the risk of a 50+% attack gets actually lower for you, because you get to choose to go with a branch that you personally consider less vulnerable.

[ecliptic]

Can you use/fork altcoins (maybe a temporary blockchain?) to handle microtransactions with minimal impact on the blockchain?  Is this completely insane?

It sounds like any solution for microtransactions AND not bloating block size requires some degree of centralization.  Like, microtransactions to company A require that all people who want to microtransct with A use a pooled, shared BTC account.  then it batches together transactions for many people, keeping the data of who gets what off the blockchain.

but it would be one-way, only many people to one company.  one company to many people (ASIC miner shares for example(?)) seems to be a pain in the ass

The risk of a 51% attack on a lower hashrate network is mitigated by the fact that the transactions on said network are significantly less valueable.


i.e. the Y $ invested to pull off a 51% attack on the main network produces X $ expected value from a double-spending attack, X/Y is the same for the alternate chain.  Of course calculated in BTC for these purposes

it just seems like altcoins (especially with faster blocks, and for transactions where the value is low that you don't really give a damn about the [hopefully] small chance for a 51%/double spend attack ) would be a good solution for micro-transactions.

[Anon136]

Can you use/fork altcoins (maybe a temporary blockchain?) to handle microtransactions with minimal impact on the blockchain?  Is this completely insane?

It sounds like any solution for microtransactions AND not bloating block size requires some degree of centralization.  Like, microtransactions to company A require that all people who want to microtransct with A use a pooled, shared BTC account.  then it batches together transactions for many people, keeping the data of who gets what off the blockchain.

but it would be one-way, only many people to one company.  one company to many people (ASIC miner shares for example(?)) seems to be a pain in the ass

The risk of a 51% attack on a lower hashrate network is mitigated by the fact that the transactions on said network are significantly less valueable.


i.e. the Y $ invested to pull off a 51% attack on the main network produces X $ expected value from a double-spending attack, X/Y is the same for the alternate chain

this is precisely why i own some litecoins 

[ecliptic]

Can you use/fork altcoins (maybe a temporary blockchain?) to handle microtransactions with minimal impact on the blockchain?  Is this completely insane?

It sounds like any solution for microtransactions AND not bloating block size requires some degree of centralization.  Like, microtransactions to company A require that all people who want to microtransct with A use a pooled, shared BTC account.  then it batches together transactions for many people, keeping the data of who gets what off the blockchain.

but it would be one-way, only many people to one company.  one company to many people (ASIC miner shares for example(?)) seems to be a pain in the ass

The risk of a 51% attack on a lower hashrate network is mitigated by the fact that the transactions on said network are significantly less valueable.


i.e. the Y $ invested to pull off a 51% attack on the main network produces X $ expected value from a double-spending attack, X/Y is the same for the alternate chain

this is precisely why i own some litecoins 

yeah, move several BTC (or more!) into LTC/etc altcoin on an exchange.  takes one transaction in the block.  You now can do microtransactions on your altcoin block with no influence on BTC.  when you're done, move LTC/etc altcoin back to BTC on an exchange.  again, one transaction in the block.

And going from virtual currency <-> virtual currency i'm pretty sure was explicitly listed as not subject to regulation.

(which is funny, because it could easily serve as a coin mixer, which i think they tried to call illegal.)

[DeathAndTaxes]

And going from virtual currency <-> virtual currency i'm pretty sure was explicitly listed as not subject to regulation.

Dubious as it may be FinCEN believes their oversight authority covers the exchanges of virtual currency for other virtual currencies.

[d'aniel]

I don't see how having one currency for 'store of value' and one for 'medium of exchange' is 1) possible to centrally plan, and 2) helps to do anything except dilute mining power and liquidity of both.  Bitcoin is a valuable medium of exchange because it commands significant mining power, it commands significant mining power because the block reward is valuable, and the block reward is valuable because it's a good store of value.  (Yes, 'significant' and 'good' are relative terms.)  And even if the currency functions could be separated this way, how would the 'medium of exchange' currency have avoided the very censorship problem it was designed to solve in the first place for the 'store of value' currency?  It's pretty difficult to move back and forth between the two if either one is censored.

[edmundedgar]

OK, so now I see some bitcoin war having a chance to pop up in a near future. People vs corporations?
It's not good, and I surely hope to stand on the right side.

It's actually quite interesting topic to consider.
Splitting the coinbase into forks, where one stays at the limit, while the other one goes unlimited in size, or there are more branches, with different limits...

I think technically that could be done quite easily. Almost naturally - you just let a fork go and see how the market acts upon it.
BTC would split into BTC1 and BTC2 and from that moment each of them would have a different exchange rate. At the beginning, probably both below the last common price, but later - obviously either of them is still better than any other concurrency..
Maybe that could actually work

There's no particular need to have different exchange rates for different chains. What we're talking about is just sharding - it's a common way of growing systems that run into some natural scaling limitation. Without doing anything particularly clever:

When you got to a particular size you'd split the existing outputs into two shards, which would split in turn when they got too big. So some of your coins would be on Shard 1 and some would be on Shard 2. Your client would have to be able to connect to all the shards you had coins on. Most people would use SPV or whatever, but if they wanted to they could run full nodes for some, all or none of the shards.

For most purposes this would look identical to the end user to what they see at the moment. Just as your client currently looks at the amount of money you want to spend and figures out the appropriate collection of spendable outputs, it would gather together enough outputs on the same shard to make your payment.

The hitch is that if you didn't have enough coins on a single shard to make your payment, you'd have to make two different transactions, one on each shard. The client could display this as it would a single transaction, except that you might have something like "65% 3 confirmations, 35% 1 confirmation". In theory there could be cases where the 35% didn't confirm at all. The vendor would need a bit of extra logic to watch for their payment coming through in multiple transactions.

If you needed the whole payment to go through in one transaction - maybe it's a fancy escrow contract or something that needs to be atomic - you'd have to exchange coins on one shard with someone for coins on another shard.

There's no particular reason why people should ascribe different values to different shards - they're all Bitcoins (or whatever alt-coin), they can all be paid to and from the same addresses, and they all look the same in the client. But if you were worried that people might end up valuing different shards differently for some strange reason, you could head the problem off by making a method to move coins across shards. It wouldn't have to be fast or cheap - you'd just need to be able to reassure people that every shard was worth the same, and if anyone got the idea that they were worth different amounts the difference could be arbitraged away.