New video: Why the blocksize limit keeps Bitcoin free and decentralized

[becoin]

Video: http://www.youtube.com/watch?v=cZp7UGgBR0I

Website: http://keepbitcoinfree.org/

Pretty simple right now, but this is the beginning. For those of you at the 2013 conference, I'll be giving a presentation about off-chain transactions on Saturday as part of the tech stream.

Congratulations on the conclusions made. They are exactly as if reading directly my mind. I'm speaking about that for almost 2 years.

[ArticMine]

This is because the aggregate their transactions and perform them periodically.

Using an aggregator is not always possible and will not always happen. There could be two networks that have few transactions occurring between them. How are you going to transfer your currency from one network to another without a $20 network transaction in that case? Use Ripple?

Ignoring the fact that $20 was pulled right out of someone's ass at some point (and is doubled when it does not sound like enough to make a point...)

I mean they aggregate their internal transactions and square periodically with other entities when they need to.  If at the end of the day they have only a $100 imbalance, they don't even bother to square that day.  They'll just cover something like that with their own buffer of working capital.  In any event, the user see's a UI and can drag-n-drop funds as they please.  Again, with minimal fees.

What you are describing is nothing more than the clearing of cheques between competitive banks at the national level and it has worked well within a single country for well over a century. The trouble is that this does not happen effectively at all between competitive propriety e-money providers on an international level. Try moving money between say Perfect Money. PayPal, WebMoney, OKPay, etc and see how much you will end up paying in fees. By the way if this were possible in a cost effective manner there would little use for Bitcoin in the first place.

The key advantage of Bitcoin blockchain transactions have is that the end users can move the Bticoins themselves across international and sub national borders thereby freeing the respective MSBs from having to be compliant in, and develop trust in, multiple jurisdictions. This places Bitcoin in a unique competitive advantage over other forms of money transmission ranging from Hawala, to PayPal to Credit Cards to Bank transfers to Western Union etc.

[tvbcof]

What you are describing is nothing more than the clearing of cheques between competitive banks at the national level and it has worked well within a single country for well over a century. The trouble is that this does not happen effectively at all between competitive propriety e-money providers on an international level. Try moving money between say Perfect Money. PayPal, WebMoney, OKPay, etc and see how much you will end up paying in fees. By the way if this were possible in a cost effective manner there would little use for Bitcoin in the first place.

The key advantage of Bitcoin blockchain transactions have is that the end users can move the Bticoins themselves across international and sub national borders thereby freeing the respective MSBs from having to be compliant in, and develop trust in, multiple jurisdictions. This places Bitcoin in a unique competitive advantage over other forms of money transmission ranging from Hawala, to PayPal to Credit Cards to Bank transfers to Western Union etc.

Intermediary processors can realistically be no more free than the medium they process.  The costs of doing business with fiat processors is directly related to the fact that it costs a lot to comply with the the various regulations which they are forced to comply with.  Also, and hardly by accident, the barrier to entry is high fostering end-user gouging and stifling competition.

Off-chain Bitcoin processors would have none of these problems.  If Bitcoin itself can remain free, off-chain processors can as well, or at least have that possibility.  It is inevitable that open-source reference software which defines best-practice transparency and security will also develop and become the base platform which off-chain processors would use.  This would make the barrier to entry very low and their reliability to end users quite high from the start.

Of course off-chain processors who also wish to have a foot in fiat-land are a different story.  At least some of their fees will be much higher, and their life expectancy will not be very good.  I don't see off-chain transactions as the magic bullet for Bitcoin<-->fiat transfers, but I am hopeful that the need to do such transactions will decline.

I might add that if Bitcoin becomes a size when 'a few thousand' entities have the capability to operate the infrastructure, said entities will be at least the size of Dwolla, OKPay, etc.  And as history has demonstrated, they are far from immune from being pressured into whatever compliance is demanded.  Even if based in Russia.  I will have lost confidence in Bitcoin before it gets baited into that trap.

[tvbcof]

Sure, a blocksize limit is needed, we can't have blocks 10GB big, but the current limit of 1MB is definitely too low.

Anyway, i am checking the latest blocks, they are nowhere near the limit, why? most are 100/200kb big. But there are 19813kb of unconfirmed transactions.

Nobody really knows why Satoshi set things at 1MB.  He didn't leave a comment.

But an interesting thing happened on the IRC channel a few days ago.  The high priests got to scratching their heads about things and noticed that 1MB just happens to put things right near the edge of what is practical to competitively mine off of Tor.  That's how I read the conversation.  Coincidence?  Who can say...

Probably that number was chosen considering the current technology level. As internet connections get faster, it can be happily increased.

That is not clear.  It depends a bit on the attack surface that Tor exposes and how threats to it will be mitigated.

But if you wish to propose basing the block size limit on what technologies are proven to be able to work around coordinated network level attacks, it would be a metric I would be very interested in.

[amincd]

This is because the aggregate their transactions and perform them periodically.

Using an aggregator is not always possible and will not always happen. There could be two networks that have few transactions occurring between them. How are you going to transfer your currency from one network to another without a $20 network transaction in that case? Use Ripple?

Ignoring the fact that $20 was pulled right out of someone's ass at some point (and is doubled when it does not sound like enough to make a point...)

I mean they aggregate their internal transactions and square periodically with other entities when they need to.  If at the end of the day they have only a $100 imbalance, they don't even bother to square that day.  They'll just cover something like that with their own buffer of working capital.  In any event, the user see's a UI and can drag-n-drop funds as they please.  Again, with minimal fees.

What you are describing is nothing more than the clearing of cheques between competitive banks at the national level and it has worked well within a single country for well over a century. The trouble is that this does not happen effectively at all between competitive propriety e-money providers on an international level. Try moving money between say Perfect Money. PayPal, WebMoney, OKPay, etc and see how much you will end up paying in fees. By the way if this were possible in a cost effective manner there would little use for Bitcoin in the first place.

The key advantage of Bitcoin blockchain transactions have is that the end users can move the Bticoins themselves across international and sub national borders thereby freeing the respective MSBs from having to be compliant in, and develop trust in, multiple jurisdictions. This places Bitcoin in a unique competitive advantage over other forms of money transmission ranging from Hawala, to PayPal to Credit Cards to Bank transfers to Western Union etc.

That's exactly how I see it, and what I've been trying to articulate to tvbcof.

Intermediary processors can realistically be no more free than the medium they process.  The costs of doing business with fiat processors is directly related to the fact that it costs a lot to comply with the the various regulations which they are forced to comply with.  Also, and hardly by accident, the barrier to entry is high fostering end-user gouging and stifling competition.

Off-chain Bitcoin processors would have none of these problems.  If Bitcoin itself can remain free, off-chain processors can as well, or at least have that possibility.

I'm genuinely trying to understand why you believe BTC banks and clearing houses would be any different than fiat ones, and struggling to understand your reasoning.

How does the medium they process being free prevent them from being subject to the same inefficiencies, over-regulation, and centralization you see in other industries? Gold is free, but trading gold through an exchange is not low cost.

To me, it seems obvious that it's their size and immobility, which comes from the requirement for trust-based relationships between the clearing house and its members, that makes them choke points in a market, not whether they deal in bitcoin, dollars, euros or gold.

Any type of credit-based market will naturally centralize, as a result of people and companies gravitating toward the party with the most already existent connections, and that central point then becomes a critical and easily disrupted point in the industry.

It is inevitable that open-source reference software which defines best-practice transparency and security will also develop and become the base platform which off-chain processors would use.  This would make the barrier to entry very low and their reliability to end users quite high from the start.

It's inevitable? That's a pretty confident statement, which again, I can't understand the basis for.

Whether there is quality open-source reference software for payment processors and clearing houses is not even that important to my critique of your vision, just an example of one of your claims that I see as pure conjecture.

Of course off-chain processors who also wish to have a foot in fiat-land are a different story.  At least some of their fees will be much higher, and their life expectancy will not be very good.

If there's a major clearing house that tens of thousands of BTC-banks use, it will be subject to the same regulatory obligations as a clearing house that fiat-banks use.

I might add that if Bitcoin becomes a size when 'a few thousand' entities have the capability to operate the infrastructure, said entities will be at least the size of Dwolla, OKPay, etc.  And as history has demonstrated, they are far from immune from being pressured into whatever compliance is demanded.

The central clearing house that your free BTC-credit transfers require will be orders of magnitude larger than these entities, which by the way is an exaggeration, as parties much smaller than Dwolla would be able to operate a high-bandwidth node, will be much harder to replace if taken down, is a much more attractive target for governments to attack, and will have a much larger network effect advantage that leads to increasing centralization than those peer-to-peer nodes.

[tvbcof]

I'm genuinely trying to understand why you believe BTC banks and clearing houses would be any different than fiat ones, and struggling to do so.

I just explained it to you in what I thought were very clear terms.

Processors who deal with fiat are encumbered by all of the baggage that fiat systems and their regulatory apparatus bring along.

Processors who deal directly with a free Bitcoin blockchain simply don't have these problems to pass along to their customers.

The central clearing house that your free BTC-credit transfers require will be orders of magnitude larger than these entities, which by the way is an exaggeration, as parties much smaller than Dwolla would be able to operate a high-bandwidth node, will be much harder to replace if taken down, is a much more attractive target for governments to attack, and will have a much larger network effect advantage that leads to increasing centralization than those peer-to-peer nodes.

You can try to employ various annoying rhetorical devices like labeling off-chain processors 'BTC banks' or 'BTC-credit transfer', but in point of fact, it is technically feasible to enforce just about any model one likes.

Other than that, you statement is hard to parse.  As best I can tell, the most legitimate answer to your concern is that, again, the powers that be picking off a second-tier processor is whack-a-mole and does not threaten the entire Bitcoin economy.  This is particularly true if pretty much anyone can bootstrap to being an off-chain processor by running some freely available software which I see as highly probable.

It would be most common, I suspect, that off-chain processors are incapable of inflating or stealing customers funds by use of multiple signature mechanisms and the like.  This contrasts sharply with our current banking model under fiat systems which are forced by law to give customer's funds back on demand.  Until the law changes as we saw in Cyprus.

That said, other flavors of off-chain processors are perfectly viable including those which could be legitimately labeled 'banks' or 'credit transfer agents'.  As I said, under the right set of circumstances which protect my interests I'd be happy to patronize them myself since I don't believe that they are inherently evil or whatever.

From a customer's standpoint, the choice of which processors to make use of shifts from an economic decision to more of a political one.  I'll be looking for diversity in terms of operational jurisdiction, transparency, customer privacy, promotion of mutual interests, etc.

And least we forget, a vast bulk of my assets would be sitting undisturbed and gathering dust in plain old vanilla Bitcoin with a paper wallet in a very safe place.

[amincd]

I just explained it to you in what I thought were very clear terms.

Processors who deal with fiat are encumbered by all of the baggage that fiat systems and their regulatory apparatus bring along.

Processors who deal directly with a free Bitcoin blockchain simply don't have these problems to pass along to their customers.

And I explained why I am struggling to understand your reasoning:

If there's a major clearing house that tens of thousands of BTC-banks use, it will be subject to the same regulatory obligations as a clearing house that fiat-banks use. All credit has fundamentally the same technological properties and reliance on trust, regardless of what the underlying asset backing it is. Any credit redeemable in a major global currency, which is what we're hoping BTC becomes, will be seen as equivalent to money, and handling it subject to the same financial regulations fiat currency is.

The central clearing house that your free BTC-credit transfers require will be orders of magnitude larger than these entities, which by the way is an exaggeration, as parties much smaller than Dwolla would be able to operate a high-bandwidth node, will be much harder to replace if taken down, is a much more attractive target for governments to attack, and will have a much larger network effect advantage that leads to increasing centralization than those peer-to-peer nodes.

You can try to employ various annoying rhetorical devices like labeling off-chain processors 'BTC banks' or 'BTC-credit transfer', but in point of fact, it is technically feasible to enforce just about any model one likes.

It's annoying to you because it exposes the kind of centralized, traditional financial system you want to replace Bitcoin transactions with.

Other than that, you statement is hard to parse.  As best I can tell, the most legitimate answer to your concern is that, again, the powers that be picking off a second-tier processor is whack-a-mole and does not threaten the entire Bitcoin economy.

Which ignores every argument I made with a handwave. Second-tier processors (which in reality are BTC-banks) are not all going to have credit relationships with each other. They will use clearing houses to avoid $20 on-chain transaction fees. Clearing houses have a tendency toward centralization, for reasons I explained:

Any type of credit-based market will naturally centralize, as a result of people and companies gravitating toward the party with the most already existent connections, and that central point then becomes a critical and easily disrupted point in the industry.

They become a choke point in the market, that is very difficult to replace, and easy to disrupt.

[tvbcof]

I just explained it to you in what I thought were very clear terms.

Processors who deal with fiat are encumbered by all of the baggage that fiat systems and their regulatory apparatus bring along.

Processors who deal directly with a free Bitcoin blockchain simply don't have these problems to pass along to their customers.

And I explained why I am struggling to understand your reasoning:

If there's a major clearing house that tens of thousands of BTC-banks use, it will be subject to the same regulatory obligations as a clearing house that fiat-banks use. All credit has fundamentally the same technological properties and reliance on trust, regardless of what the underlying asset backing it is. Any credit redeemable in a major global currency, which is what hoping BTC becomes, will be seen as equivalent to money, and handling it subject to the same financial regulations fiat currency is.

Ah, you are legitimately confused then.  If off-chain processors wish to avoid BTC transactions themselves when practical, they can probably trade directly with other off-chain processors on an exchange or via individual relationships.  That's a little tangential, but it is not critical to their operations nor particularly visible to their customers other than that it will further reduce fees and increase performance and feature set.

An exchange or 'clearing house' formed to facilitate such interactions is tiny and vastly more conducive to operating smoothly than one who deals with fiat.  They can easily lurk in the shadows.

Customer's of the off-chain processor still enjoy the protection offered by the Bitcoin blockchain and mathematics.  Off-chain providers who leverage such a 'clearing house' risk only what they have on balance since the last time they squared on the block chain.

tl;dr, the Bitcoin block-chain is THE clearing house.  Other auxiliary ones are perfectly possible.

[ArticMine]

I'm genuinely trying to understand why you believe BTC banks and clearing houses would be any different than fiat ones, and struggling to do so.

So am I. Especially after reading the FinCEN guidance http://fincen.gov/statutes_regs/guidance/html/FIN-2013-G001.html

I just explained it to you in what I thought were very clear terms.

Processors who deal with fiat are encumbered by all of the baggage that fiat systems and their regulatory apparatus bring along.

Processors who deal directly with a free Bitcoin blockchain simply don't have these problems to pass along to their customers.

The FinCEN guidance above makes it abundantly clear that this is not the case. It does not matter if the underlying asset is gold, USD, Bitcoin or any other form of money. Set up a MSB and one is subject to regulation.

[ArticMine]

Ah, you are legitimately confused then.  If off-chain processors wish to avoid BTC transactions themselves when practical, they can probably trade directly with other off-chain processors on an exchange or via individual relationships.  That's a little tangential, but it is not critical to their operations nor particularly visible to their customers other than that it will further reduce fees and increase performance and feature set.

An exchange or 'clearing house' formed to facilitate such interactions is tiny and vastly more conducive to operating smoothly than one who deals with fiat.  They can easily lurk in the shadows.

Customer's of the off-chain processor still enjoy the protection offered by the Bitcoin blockchain and mathematics.  Off-chain providers who leverage such a 'clearing house' risk only what they have on balance since the last time they squared on the block chain.

tl;dr, the Bitcoin block-chain is THE clearing house.  Other auxiliary ones are perfectly possible.

This does not address the critical question of multi jurisdictional regulatory compliance and the sky-rocketing costs for the processors involved. Ever wonder why M-Pesa does not work in the US or Dwolla in Canada or Interac in China?

[tvbcof]

I'm genuinely trying to understand why you believe BTC banks and clearing houses would be any different than fiat ones, and struggling to do so.

So am I. Especially after reading the FinCEN guidance http://fincen.gov/statutes_regs/guidance/html/FIN-2013-G001.html

I just explained it to you in what I thought were very clear terms.

Processors who deal with fiat are encumbered by all of the baggage that fiat systems and their regulatory apparatus bring along.

Processors who deal directly with a free Bitcoin blockchain simply don't have these problems to pass along to their customers.

The FinCEN guidance above makes it abundantly clear that this is not the case. It does not matter if the underlying asset is gold, USD, Bitcoin or any other form of money. Set up a MSB and one is subject to regulation.

Of course a responsible off-chain provider (or any entity) will be complying with all regulations which apply within their jurisdiction.

It is perfectly possible, however, that at some point regulations might be promulgated exclusively for the purposes of protecting one's benefactors in an an environment rife with cronyism.  I, and I think others, seek a solution which could realistically bow out of participating in the aspects of such an arrangement which serve no useful purpose for in protecting the market participants and general welfare of the citizenry.

[amincd]

An exchange or 'clearing house' formed to facilitate such interactions is tiny and vastly more conducive to operating smoothly than one who deals with fiat.  They can easily lurk in the shadows.

This is my main point of dispute with you: as far as I can see, there is no reason to assume these clearing houses will be tiny and would lurk in the shadows.

Clearing houses almost by definition are central points in a trade network, which gain efficiency in aggregation with scale (more volume = larger aggregation and lower transaction fees), giving the largest clearing house a natural advantage over the others.

For this reason, I think a Bitcoin economy reliant on traditional payment aggregation to keep fees down will end up with a clearing house that is as critical for straight BTC transactions as MtGox is for BTC-fiat transactions.

tl;dr, the Bitcoin block-chain is THE clearing house.  Other auxiliary ones are perfectly possible.

The blockchain is merely the network where the transaction takes place. The clearing house has to aggregate payments before clearing them.

[tvbcof]

The blockchain is merely the network where the transaction takes place. The clearing house has to aggregate payments before clearing them.

Nope.  Aggregation will happen largely within the off-chain processor itself.  That is indeed their primary function and why they are helpful to the Bitcoin network.

If/when it is desirable to clear between processors it could be either through messaging or interaction on one or more light-weight exchanges and squaring on the Bitcoin blockchain, or potentially squaring on some other agreed upon medium.  The exact nature of the solution(s) employed depend a lot on the type of attacks (if any) they need to deal with.

[amincd]

Aggregation will happen largely within the off-chain processor itself.  That is indeed their primary function and why they are helpful to the Bitcoin network.

I'm referring to aggregation of payments between BTC-banks (off-chain processors), not within them. These will go through a large central clearing house.

If/when it is desirable to clear between processors it could be either through messaging or interaction on one or more light-weight exchanges and squaring on the Bitcoin blockchain, or potentially squaring on some other agreed upon medium.

Which goes back to my main point of dispute with you, which I've already detailed: I see no reason to assume these clearing houses will be "light-weight". Every comparable real world market instead suggests a few, or a single large clearing house, that will handle almost all payment aggregation between banks. A centralized point where all inter-BTC-bank transfers go through would be neither light weight nor "lurk in the shadows".

[tvbcof]

If/when it is desirable to clear between processors it could be either through messaging or interaction on one or more light-weight exchanges and squaring on the Bitcoin blockchain, or potentially squaring on some other agreed upon medium.

Which goes back to my main point of dispute with you, which I've already detailed: I see no reason to assume these clearing houses will be "light-weight" or numerous. Every comparable real world market instead suggests a very few, or a single large clearing house, that will handle almost all payment aggregation.

If they are attacked successfully, they will splinter.  Just as will happen if/when Mt. Gox get's shut down.

Unlike Mt. Gox who needs to assume the overhead of operating in fiat-land, such (mainly optional) clearing exchanges don't.  This makes them vastly less capital intensive to operate.

Secondly, off-chain processors operate with some reasonable buffer and, to the extent that they need to use such messaging systems at all, can do so at an irregular and relatively low frequency.

Remember that what we are talking about here is not a layout of all transactions in the economy, but rather an aggregation of them which occurred withing the off-chain processor's space.  Even if they find it necessary to comply with cumbersome regulations within their jurisdiction, the only 'customer' then need to 'know' is the off-chain processor.  They simply don't deal with individual users.

[amincd]

But there's no reason to assume it will be attacked. Instead the clearing house could turn into a traditional clearing house that operates under the auspices of the regulatory bodies of multiple jurisdictions, the same way as MtGox is turning into a traditional financial institution.

Unlike Mt. Gox who needs to assume the overhead of operating in fiat-land, such (mainly optional) clearing exchanges don't.  This makes them vastly less capital intensive to operate.

Any centralized party handling credit will be subject to the overhead of operating in fiat-land. Fiat-land suffers overhead because it is credit-based and happens on centralized, market critical conduits, as a large BTC-credit clearing house would. Fiat activity that is not credit-based and does not go through these central clearing houses, like cash transactions, doesn't have the same overhead.

Remember that what we are talking about here is not a layout of all transactions in the economy, but rather an aggregation of them which occurred withing the off-chain processor's space.

We're talking about how users would move their BTC from one BTC-bank to another when network transactions cost $20. If the lowest cost option is for BTC-banks to use a large traditional financial institution for clearing BTC credit positions, then the BTC economy will look like the modern financial industry.

[tvbcof]

Remember that what we are talking about here is not a layout of all transactions in the economy, but rather an aggregation of them which occurred withing the off-chain processor's space.

We're talking about how users would move their BTC from one BTC-bank to another when network transactions cost $20. If the lowest cost option is for BTC-banks to use a large traditional financial institution for clearing BTC credit positions, then the BTC economy will look like the modern financial industry.

Not at all.

Firstly, it is somewhat rare to move value from one processor to another vs. buying one's morning Java.  It's common enough for a processor to wish to optimize however.

Secondly, all of the people who did such a transfer in, say, one day would be balanced against those who did the opposite transfer.  If there is a notable difference in value which the off-chain processor does not feel like carrying, he makes basically a single transaction on the block chain which covers all of the day's activity.  This one (supposed) 0.15 BTC cost is then split amongst all of the customers in that time-frame.

Customers win by sharing the fees of a blockchain transaction and Bitcoin wins by lowering the data-rate.

---

You may claim that because this is somewhat similar to how banks work, it means that the entities will be just like banks.  I reject that because I claim that the things we don't like about banks are more associated with their dealing with fiat in todays climate than they are with the properties which naturally fall out of the search for efficiency.

[amincd]

Firstly, it is somewhat rare to move value from one processor to another vs. buying one's morning Java.  It's common enough for a processor to wish to optimize however.

It's not a processor, it's a bank. A processor merely processes your transaction, not provides you with BTC-credit in exchange for your BTC.

In BTC-land, it's not that rare to move your BTC from one e-wallet to another. If network transaction fees are $20, then you can't do it cost effectively unless it's a huge sum. The other option is a clearing house. Any sufficiently efficient clearing house would have to be large.

In a scenario where BTC is a major global currency, I see almost zero possibility that a large BTC clearing house used by thousands of BTC-banks would not be a traditional financial institution. We're not talking a few thousand transactions a day. We're talking hundreds of millions of transactions, worth hundreds of billions of dollars.

This would make a BTC economy with $20 transaction fees very similar to the modern financial system, where the lowest cost option for transferring BTC is established credit relationships between established financial institutions.

[tvbcof]

Firstly, it is somewhat rare to move value from one processor to another vs. buying one's morning Java.  It's common enough for a processor to wish to optimize however.

It's not a processor, it's a bank. A processor merely processes your transaction, not provides you with BTC-credit in exchange for your BTC.
...

Not true (nullifying the arguments I snipped.)  Just because the payment processor is using some of his own capital to make transactions efficient does not mean that there needs to be a credit relationship between he and his customers.  And he could eliminate any unwanted credit relationship between he and his peers with a bond if they like.  This still allows all of the functions I described.

Again, I personally am happy in some situations to patronize processors who DO have some credit component if that is how they choose to operate, but they would not likely be my primary store of even my spending money.

[amincd]

Just because the payment processor is using some of his own capital to make transactions efficient does not mean that there needs to be a credit relationship between he and his customers.

I'm having trouble understanding your explanations now. They seem to be more convoluted and vague than they need to be.

A credit relationship exists as soon as a customer deposits BTC with the bank, and is given demand deposits in exchange. The demand deposits are credit, that others accept as BTC, because they trust the bank to be good for them.

And he could eliminate any unwanted credit relationship between he and his peers with a bond if they like.  This still allows all of the functions I described.

Again, having trouble understanding what you mean.