Topic: Epic, monstrous post of Jihan Wu (AntPool) (Read 4881 times)

Sorry to interrupt all you "legendarys." But speaking as one who has to Google practically every post on this forum. Can us poor new era sops "since 2014" please get some definition of terms used here?
e.g. "Classic", "Sybil" etc...
Also, for my 2 cents, the real elephant in the room (and possibly silent partner to the gang of 3/4?) speaking as a Dev from the dark side (Microsoft) is the absolutely shockingly incomprehensible and ummaintainable code these "God like and secretive, ha!" core devs feel that they can get away with.
It is, to me, very excluding.

Please excuse the rant and the iPad keyboard syndrome but I have kept this frustration pent up for far too long.
Thanks for not flaming me in advance.

Thank you for your answer sturle...

I think satoshi someday mentioned that mining will be done by corporations at the end. It might be that he did not anticipate this becoming true that fast.

The exact details are over my head.  It is in some way possible to prove the sum of the outputs equal the sum of the inputs without revealing the amounts, and that's all you need to know the transaction is valid.  The amount can of course be decoded by the receiver's private key.

So when all we know is sum of Bob, Charlie and Alice change equeals to original Alice 1 BTC, how can we know how much Bob can spend? I assume we never know, but when I receive from Bob 0.8 BTC I get this info (how can I confirm he can spend 0.8 BTC?), and how can I confirm Charlie and Alice change can only spend the remaining 0.2 BTC and not more, like creating up to 3 BTC out of the initial 1 BTC

No, it is fundamentally different.  Bitcoin Core isn't trying to hard fork bitcoin.  If one of the many people who contribute to Bitcoin Core works for a bank, it simply doesn't matter.  He can't change the fundamental consensus, and even if he tries, such change would require support from 100% of nodes.  Hearn didn't contribute much in any case.  He didn't have commit access, and most of his ideas, like denying nodes to connect via Tor, were turned down.  Instead he forked off BitcoinXT before he left to work with some bank stuff.  Most of BitcoinXT has been merged with "Classic".

Yes, they keep it.  If the customer take a claim to the court, it can go either to the customer or to the state, but so far I haven't heard of a single case ending up in court.

Nobody have tried to calculate it as far as I know, but Coinbase is the third largest exchange by volume.  Due to their high fees, I doubt much of it is buying and selling within the exchange, like at BTC China or Bitfinex.  Both of them batch withdrawals, btw, so Coinbase may be the largest single source of transactions on the blockchain.

No.  Bitcoin Core isn't trying to hard fork bitcoin or take control over it.  They maintain Bitcoin within the original consensus.

I don't think adoption will be slow.  Adopters will have a strong economic incentive to do adopt segwit as soon as it has been accepted.  If it isn't, it just means fees are so low that people don't care.

There is at least one private sidechain enabling instant transfers between Kraken, Bitfinex, Xapo and more.  Two-way pegged sidechains are not exactly altcoins.

I think we will be deadlocked.  "Classic" demanding a hard fork berfore segwit (and probably won't implement segwit, or will implement some entirely different form of segwit), and we will have a deadlock where segwit can't gain the required 95%, and "Classic" can't get their ~72%.  Perhaps Classic can be thrown off by a coordinated effort where miners signal support for "Classic" for 1000 blocks, and then return to mine normal bitcoins.  Throwing off the fork without miner support will get rid of "Classic" for a while.

Define a "normal bitcoin user".  Transactions are mined by miners.  Users never had any other influence than trying to interest the miners with fees.  Anyone can be a miner.  Almost anyone can run a node (now).  If Bitcoin is going to scale by increasing the block size, running a full node will be off limits for more people.  Would that be pushing out normal bitcoin users as well, or not?

If we are going to increase the blocksize ten-fold to get the same capacity increase as LN with 10 transactions per block-chain transaction, with the same fee level as you suggest (average $2 per transaction), I won't be able to run a full node at home any more.  I would have to go with a less secure solution with less privacy.  Why is that better for normal users?

I think this is an irrational fear.  A scheme like this will not get much support.  LN is useful for fast transactions, not for holding bitcoins for a long time.  If average people feel cheated, they can make a two-way pegged sidechain, and transact there.  The sidechain used by Kraken, etc, support confidential transactions as well, iirc.

The information doesn't have to be public.  There can be private sidechains where outsiders have no access to the blockchain.  With CT (confidential transactions) the amounts aren't public, but the outputs can be proven to match the inputs.  You can see the transactions between sidechains and Bitcoin, but you won't know which seidechain it was sent to unless you can monitor the sidechain as well.

Hopefully someone will find a way to plug sidechains into Bitcoin Core, and let the user transact on multiple chains without even knowing which sidechain or payment channel a transaction is going to.

Where is "should only be achieved by rising the amount of transactions" from?  Many of Satoshi's plans have turned out differently than he expected.  E.g. GPU/FPGA/ASIC mining instead of every bitcoin user mining on their computer.

Demand for bitcoin is higher than ever.  My turnaround in bitcoin is setting new records every month.

The image will just be uploaded once, then deployed multiple times.

The core devs have done one hard fork in the past.  It was so successful, the chain didn't actually fork.  As far as I know not a single block has been produced on the other side of the fork.  It fixed an unintentional bug due to a quirk in a database library, and was as uncontroversial as a hard fork can be.

Lauda



Well, you really are onesided. The thing you claim an error on the classic side is... of course, fully normal on the core side.


And you spoke about the importance of having to refer to worst case scenarious. It amazes me more and more how onesided you can watch at things.

You even seem to advertise LN again only after pushing away potential future risks as hyperbolic and so on. Well, it seems I will stop answering you in this thread too. It simply leads to nothing.


What did I say? I'm amazed.

Anyway... you are a true believer and being that it is as hard as with discussing with islamists, hardcore christians or antigermans.

Good luck again...

---

sturle



So they really keep it? They don't hold it as escrow? Well, that is a unique business model when the money is not given to authorities or something. I did not know that company is on such a low level.


But how high is the amount of transactions per block in average that belongs to coinbase? Is it a really big chunk at all? Well, sure they could change it and it might be true that classic developers would do what they want when they pay them. Though corruption like that could easily happen on the core side too, I think. The community would stand against though, I guess.

I'm not sure but I think the usefullness of segwit is accepted more and more. I mean segwit will have a positive effect all the time, even when blocks are higher. Segwit is able to raise the blocksize space on an prozentual level. Which is worth much and way far in the future. Well, of course segwit has to show that the predictions of it's usefullness is true. There are opinions that the target levels are way overexxagerated. At least segwit will come too slow to prevent the blocks being full. Adoption will probably happen too slow.


Hm, that certainly is an advantage though I thought that is already possible now. I did not realize till now that no sidechains exist yet. I thought it should be no problem since there are altcoins that are bound to the bitcoins blockchain security so holding bitcoins in escrow and releasing them to another user after the altcoins were moved did not sound impossible to me. Guess I was wrong then.


Hearn was someone who should have gone since a long time. His intentions did not look pure when one saw what he wanted to implement. I already got the idea he might work for an agency because of his stupid suggestions. Disabling tor, I think tainting coins was from him too.

Anyway, he was not the best example and when Andresen works for coinbase then I see the conflict of interest you refer to.

Well, we might see. If classic agrees to segwit then the theory could break. There should be no real reason against it except maybe gavin's ego. At the moment both sides look not fully convincing for me.


I think if LN gets reality and really would be used a lot while 1mb blocks are still held then it would be obvious that the transactions on LN, merged into one big bitcoin transaction, could have a higher value. Paying $1 on the bitcoin network for one transaction would be expensive. Paying $2 on the bitcoin network for 10 transactions on LN would be cheap. But still those transactions will get the priority, they are more value for the miners and could, if the amount of those transactions rise, push out normal bitcoin users because their fees would be too low and the higher fees would be too expensive.

This only under the premise that the fees on the bitcoin blockchain will rise that high that business models come up that say "bitcoin fee too high? Buy and hold your bitcoins on LN only. Never deal with the bitcoin blockchain. And safe a lot of money." Which would mean we have a somewhat centralized service. Which might be used because it is cheap and could be safer than nowadays exchanges. This company could offer to send the bitcoins anywhere you want. Through payment channels to known services and sites or to other users using the same or a similar service. Transactions into the bitcoin network would still be expensive.

At the end this scenario would mean that the free and anonymous money would not be for the average or poor man. It would be too expensive. Instead we would have companies on LN that could not offer any anonymity. Which would be no problem for most users since they do not do anything wrong. Which sounds like a bank. And the free internet money was taken away from the average people. Free the people from the banks power would have been a failed attempt.


But blocks were never as full as now. We did already raise the size of the blocks, not by changing the limit but because the amount of transactions grew. Iam not aware that we lost half of the nodes from the time where the blocks were only half as full until now.

There was a drop in full nodes. I think that might be normal after implementing the ability to host a normal node that does not store the full blockchain. People decided to safe on that point. If this development stops, then the blocksize is raised then another drop might happen. Though that would be only ideological since the blocks wouldn't be suddenly double that full. After that happened, when the average block size is double from now, then someone could say something about the effect on nodes.

Besides that. I think in the last years alot of bitcoiners left bitcoin. Disappointed about the price development. That might have lead to the node development too.

So there should be many things to consider.


Yeah, I think when segwit enables sidechains then that would be a great thing. Besides having to run another wallet. But it still would mean freedom.

Besides, wouldn't there be explorers about the connection from bitcoin to sidechain address as well as explorer about the sidechain transactions? Those infos should all be public, isn't it?


The hope of satoshi was to complement block halving with rising fees and that should only be achieved by rising the amount of transactions, as far as I know. So something in the plan already went wrong then.

I agree with the sidechain, that's a good thing. But demand going down for bitcoin sound sad. Bitcoin becoming too expensive to use so that demand even drops is not a nice vision.


Didn't know till now. I thought it will be a network that practically will eat bitcoin to use it as their backend database.

Well, let's see how usefull it will become.


Ok. Well, might be that uploading via disk would be the same for the vps's bandwith usage than downloading the blocks from the network. For the person installing it it would mean less work because the node would handle it on his own. Not using the installers home internet bandwith too.


You already got me doubting gavin. I think bigger blocks surely is not every nice thing that could get implemented. If I would create an alternative then I would try to add as many usefull features as possible. Well, he might not want it because supporters fear segwit partly and such.


I meant past hard forks were no problem though maybe people now learnt to doubt everything what the core devs do.

To most people, especially foreign customers, the cost of a lawsuit will be far higher than the amount which Coinbase confiscated.

You are getting there!  Raising the block size limit has no real value for Coinbase.  Coinbase can easily make a lot more room in the blocks by e.g. batching withdrawals in one transaction a minute or so, like almost everyone else.  What Coinbase want is control over Bitcoin.  If they are able to force through a hard fork, they have succeeded in both taking control over Bitcoin.  They will have the power to stop improvements they don't like, e.g. confidential transactions which depend on segwit.

With segwit the malleability problem is solved.  This means people can move their transactions to side-chains, payment channels, lightning, etc.  Many individual transactions will then be hidden for people doing blockchain analysis.  They can see a coin was taken out by me at an ATM, then returned to the chain by a merchant six months later, but they have no idea where the coin has been in the meantime.  Same as cash.  Segwit makes confidential on-chain transactions possible as well, after a soft fork.  You can see that Alice moved a coin to Bob and Charlie, but you can't see how much went to Bob and how much went to Charlie, or how much Alice kept as change.

Mike Hearn?  He hardly contributed anything.  Even less than Gavin did the past couple of years.

People will use LN to get instant confirmation, not due to fees.  LN will benefit from larger blocks as well.  LN is free software, btw.  There is no company monopoly.  You have to explain how companies can decide the fees to be paid.  You mean miners are organizing in some way, and plan to decide a minimum fee?

If you scale by increasing the block size to accommodate all spam, only corporations will be able to run full nodes.  It is putting Bitcoin in the hands of corporations and bank like businesses.  If you read Satoshi's paper and old postings, this was in fact the initial plan for bitcoin.  He thought that most people would be satisfied by a less secure SPV wallet.

If you enable sidechains and lightning instead, anyone can opt out of high fees and corporations by transacting on a sidechain.  It will still be bitcoin.  100% convertible between bitcoin and sidechains.  You can even have different rules for the sidechain, to enable e.g. free microtransactions.  You can even add decimals on the sidechain, as long as the transactions back to the Bitcoin blockchain are compliant.  Which means the extra decimals will only exist on the sidechain.  And normal people can still have the security and privacy benefits of running a full node at home.  The fee will not increase more than people are willing to pay, of course.  When demand goes down, and the supply stays the same, price will go down as well.  The most fundamental of all economic theory.  If you agree with Satoshi's initial plan, you are free to transact on a sidechain controlled by a bank or corporation, but Bitcoin itself will be as uncontrollable as it is now.

Lightning isn't one network.  You can have as many lightning networks as you want.  I assume most of them will be among people and businesses who transact often with each other.

When you deploy a full node on Amazon, you will of course deploy it with the blockchain pre-loaded on disk to get it up quickly.  Still those nodes seem to ask for blocks quite often.  It makes sense if the nodes are downloading the blockchain, or if it is a request forwarded by a pseudonode.  I haven't done any deeper analysis.

Do you mean Core?  Yes, there is talk about a hard fork to fix a few things.  There is an old wishlist here.  I doubt it will be successful unless it is entirely non-controversial or adds really useful stuff like Schnorr signatures.  Schnorr signatures are a way of combining many signatures into one, taking up much less block space for transactions using multiple inputs.  Why doesn't "Classic" do that instead in their hard fork?  It would be a real benefit, like cheap dust swiping, cheaper coinjoin transactions, and is less controversial.  Ah, of course..  Coinbase doesn't like Coinjoin.  It makes their chain analysis harder.

Since a hard fork will have to be deployed by 100% of users to be successful, I think it is obvious it won't happen as long as people disagree.

Even if they did, it doesn't matter. Core consists out of a lot of contributors, employed by various companies.

This is just hyperbolic, fear-inducing, speculation. The thinking behind this is flawed. Why would anyone even consider using Bitcoin if a transaction costs $10 (i.e. when there are much cheaper options)? The Lightning Network is a amazing 'feat' when it comes to scaling Bitcoin. There is no mainstream adoption without it (or some other, yet to be discovered, features), regardless of whether you want to believe this or not. Bitcoin is not efficient and can not process a lot of transactions on the main chain unless you harm its decentralization.

Baseless propaganda. May I ask who slipped this nonsense into your mind?

Yeah, at least fees should be calculated by the size of transaction. Would be even better when taking the mempools into account.

Though I mentioned "Spamming would not be usefull with zero fee transactions" because 2mb blocks would not make it cheaper to spam. Assuming people would not suddenly lower their fees a lot then spammers would still have to pay the same fee as before too because otherwise the normal transactions would take priority.

So I think spamming should not really become considerably cheaper. I think it would become alot more expensive since the amount of space they would need to fill is way higher. It would cost less again when the 2mb blocks are nearly full.


Coinbase probably can't make money with the frozen accounts. That would be outright theft and they would face lawsuits. Or is the problem that US customers can't sue them because they would have to admit that they gambled on illegal websites?

If that is the case then they surely would be low.

Well, coinbase is creating two transactions per withdraw. I don't really see the need for the second one yet but to assume that they do this to raise the blocksize sounds a bit far fetched. They would have no advantage.

You mention that segwit makes blockchain analysis way harder. How can that be? All the details about every transaction still would be available to everyone wanting to do a blockchain analysis. Or what do you refer to?


Ok, and didn't some core dev lately switch to work for a bank? Besides that, one could see a conspiracy theory behind lightning network and the negation of a raise of the blocksize too. Since what will happen when a minimum transaction will cost $1, $10 or more at one time? I mean LN plans to make bitcon the settlement layer for LN. Which makes companies will decide the fees to be paid in the bitcoin network. The more transactions can be put into one single bitcoin transaction the higher the value of the bitcoin transaction and the higher the fees on the bitcoin network can rise.

So what will happen once the fees rise very high with bitcoin? Nearly no normal user will be able to afford it. Which means it won't take long until we see offers like offers from a bank offering to buy bitcoins directly in the lightning network. Only pushing bitcoins around on there, using a service connected to lightning network.

I mean how does this sound? Like bitcoin being practically in the hand of corporations and bank like businesses? At least it sounds like the very opposite of the initial plan for bitcoin.


Hm, that sounds like a hint at least. Did you check if the traffic is something like "loading blocks to build a blockchain" or does it only look like the behaviour you described for fake nodes? Otherwise it might be that they only were sat up lately and tried to get up to date. Though I guess you already checked.


Ah ok, I was of the impression that is somehow a free service that allows a full or normal node on that space.


Wasn't there a hard fork needed in the near future for classic too? Besides the soft fork for segwit?

Besides that... a fork only on core would surely not bring such uproar like having to go against core. The uproar only comes from people disagreeing with the way the things are handled. A fork for a fix that is not needed yet would be useless of course now. But when other important things would have to be implemented anyway then implementing that fix would be nothing that would bring up trouble.

I have heard it in discussions as well, but always as a straw man.  Never anyone actually making the claim, which I find interesting.  Eligius has the most aggressive spam filter among the pools, and Eligius mine free transactions with high enough priority.

Some pools block the spam, but if there was an official script it would be trivial for the spammers to avoid it.  The picture is quite complicated as well.  The hash spam, ending with OP_RETURN, can easily be camouflaged as financial transactions instead.  The OP-RETURN transactions are prunable, the fake financial transactions are not, and will make matters worse by adding unspendable transactions to the UTXO set.

There are right ways to do it, and there are wrong ways to do it.  Satoshidice did it wrong for a long time.  Making each bet a separate blockchain transaction.

I primarily include services like Proofofexistence.  There are many other as well, as you can see here.

No, free transactions are still mined.  (Given enough priority, of course.)

Fees are calculated per kB.  Some stupid wallet implemetations use a fixed fee, which is either very high or very low depending on how the transaction ends up.  Blockchain.info used to be one of them, and may still be.  A 252 byte transaction and a 1000 byte transaction will pay the same fee of 0.0001 BTC.  If a spammer sets his fee per kB to 0.00011 BTC, she will  pay a higher fee than those wallets would calculate for a transaction of 910 to 1000 bytes.  Effectively blocking seemingly (to users of stupid wallets) random transactions for ever at a cost of maximum 0.11 BTC per block.  The real cost is lower, of course, unless all transactions in a block are from the spammer.

No, you can freely move bitcoins between the blockchain and a sidechain, and back.  No need to exchange.  With a little work on a client, you can plug in a sidechain and use them interchangeably.  If you have coins in the sidechain and want to send on the normal blockchain, it will be two transactions.  One on the sidechain and one on the normal blockchain.  If you have coins on the right chain, it will just be a single tx on the right chain.  Here is a good explanation of how sidechains work.

Coinbase is the only noteworthy business supporting "Classic".  Coinbase makes much of their money by doing blockchain analysis and confiscating the funds of people using gambling services, etc.  Just google for "Coinbase froze my account", and see their terms of service.  Coinbase knows that segregated witness will destroy this business model by enabling technologies which make blockchain analysis much harder to impossible.  Instead Coinbase wants a larger blockchain, and they use all kinds of dirty tricks to make that happen.  There is evidence of Coinbase spamming the blockchain as well, to generate artificial need for larger blocks.

Gavin Andresen works for Coinbase.

Of course.  It is supported by the fact that the "Classic" nodes on Amazon seem to use a disproportional amount of my outgoing bandwidth, but that is not proof.  I have worked around the problem by throttling my outgoing bandwidth to "Classic" nodes to make my ADSL line usable again.
If you buy enough storage, then yes.

The only other "fix" I know of is to impose a hard limit transaction size, which would take a hard fork to lift.  Considering the amount of noise the current hard fork attempt is making, why do you want to create reasons for more?  Segwit is a much better solution.  It allows for faster validation overall, and it doesn't add extra hard limits on transaction sizes.

I never thought someone would doubt that so I don't have a clue who answered that to me but I heard that pretty often in discussions around this topic. Though I'm not sure if some bigger person claimed that. Like a pool owner or something.


If it is so easy then why didn't circulate a script around pool owners to block these transactions? It could have rendered the spam useless. But it didn't happen.

Gambling as spam? They are normal transactions or what makes them spam?

Datastorage... don't you include thinks like colored coins into that too?


I'm not sure, isn't the minimum fee already reality?

Besides that, the spammer having a lower fee than the normal users with their transaction would not fill the blocks and block normal transactions.

Regarding your ADSL line... there are possibilities all the time. Doesn't matter how big of a block size.


A side chain would mean having another wallet and other coins, right? You have to exchange bitcoins and so on. I'm not fully sure but what is moved on the sidechain can be called a bitcoin IOU at max, right?

Coinbase and classic? I have read that some times now. Sounds like a conspiracy theory. I don't see anything pointing to classic getting something implemented that hurts anonymity and such.

Besides that... how should coinbase gain control at all.

Well, doesn't sound convincing. Care to give some details to convince me?


Hm, sounds like one would need a pretty big amount of time, work and money to do this. So the theories that practically all classic nodes are pseudonodes are nothing but a theory yet?


Ah that... though it is possible to run a full node or a normal one, right?

I wonder if it's worth at all anymore to use these hosts for mining. I doubt somewhat since it should be a VPS, so you will have not a lot of cpu power.


Well, if coredevs do not want to include a fix then yes, it is not fixed. But the fix exists and only need to be implemented so anyone bringing out 2mb blocks is safe. Not worth to mention a problem when the solution already exists.

Who are those people?  I have only heard people claim that other people claim this.  I have never once heard anyone claim transacctions with low fee are spam by default.  Ever.  Your straw man doesn't exist.

Spam is pretty easy to spot.  The same coins moving in circles, usually using a fee adjusted just large enough to block a few percent of transactions from wallets paying a low fixed fee.  Then there is the gambling type of spam and the transactions which are pure data-storage, which aren't moving any coins.  Just a fee and an amount of data followed by OP_RETURN.

But why should I allow spammers to fill up my data storage and ADSL line in the first place?  This is the non-solution of trying to pay yourself out of the problem by buying more disk space.  This non-solution will only allow the spammer to pay lower fees.  The total amount of fees the spammer pay to fill a block will not increase, unless you suggest to enforce a minimum fee.

If I could make my transactions on a sidechain instead, however, I could avoid the problem completely.  If the "classic" supporters think large blocks is a good idea, why not create a sidechain using very large blocks, merge-mined with the bitcoin blockchain, and suggest everyone just move their economic activity over to the sidechain?  It will work with the same security as Bitcoin, they can hard-fork to a different block-size, lower transaction size limits, and support whatever features the community wants whenever they feel like it.  They can even fork to multiple sidechains.  Why isn't this even discussed in the "Classic" community?  Simple: The hard fork has nothing to do with block size.  It has everything to do with Coinbase wanting to gain control over Bitcoin to implement their KYC stuff.

A pseudonode is very hard to detect, unless you have a sybil attack running, and manage to get the suspected pseudonode to make an outbound connection to one of your nodes.  In that case you can make an inbound connection to it from another IP, ask for data from the suspected pseudonode, and check if it sends a request for the same data to one of the nodes it has outbound connections to.  If it does, it is probably a pseudonode.

Because it use a neglible amount of resources and no diskspace.  You can use the rest of the resources of your Amazon instance to mine for an altcoin or something.

Which means it isn't fixed.  Yes, segwit fixes it.

The problem is how to define spam. There are enough people that claim transactions with a low fee are spam by default. With that attitude you have paypal fees in no time and 1 of the few bitcoin advantages is gone. Plus it being a big hit for adoption.

And bigger blocks are a solution to spam when it means you have to create way more transactions, including higher fees per block, just to spam these blocks. The recent spams only worked because the blocks reached their limit nearly. So a spammer would have deep pockets and a good reason to spam away his bitcoins to fill bigger blocks.


I always wondered about that point. Are these pseudonodes not detectable through some requests? Should be easy to find out.

And is it not possible to host real nodes on amazon? I mean if you do the work anyway then why not create a real node from the start. Why use a fake node?

By the way... that attack vector is already fixed. There is a BIP that only would need to be implemented in core. Developers didn't till now.

No evidence is needed in this case as you want to scale according to an observation (Nielsen's law). You are drawing to conclusions based on the "evidence" (from the past) gathered from a small region (?). Internet prices and speeds to to variate quite heavily depending on region and country. We need not be designing a system with a best-case scenario in our mind. Bitcoin needs to be scaled very cautiously and prepared for the worst, because as soon as something happens and one of those prediction fails for a period then you are spiraling into a 'doomsday scenario'. It is only a matter of time before things such as Nielsen's and Moore's law die. They are not laws in the traditional sense, they are observations.

Answered by sturle. This is the problem in the community. People are voicing strong support for something/someone when they lack a lot of knowledge/facts.

Neither. I'm disagreeing with the logic used to draw to a conclusion. You've picked out a single aspect that is 'shared' by both proposals. Segwit is superior as it carries many benefits with it which a 2 MB block size limit does not.

Answered by sturle.

That would be your personal definition, and not a general property of settlement layers.  Not all transactions are transparent in Bitcoin either, btw.  It has been shown how you could transact on the side, e.g. A->B->C->...->Z on a sidechain, and commit A->Z to the blockchain.  Coinjoins are commonly done to obfuscate the real partners in a transactions.

And why on Earth do you think a larger blocksize would solve the spam problem?  It will only make the consequenses worse.  Bitcoin clearly needs a way to transact unobstructed outside the blockchain, and only commit to the chain when needed.  If miners want to promote normal transactions over spam, they should use spam filters.  Some pools do, e.g. Eligius.  If spam is the problem, you should attack the spam, not allow more of it.  When my mail server got attacked by spammers, I didn't run to the shop to buy more disk.

There is an attack on non-segwit transactions where you can make a 2 MB transaction which take more than 10 minutes to validate.  "Classic" has "solved" this by introducing a new hard limit to Bitcoin ("Classic" likes to brag about a limit they lift, not so much about extra limits they add), where a transaction can't be larger than 100 KB.  The largest transaction mined to date was 1 MB, and it takes 20 seconds to validate on a modern fast CPU.

Pseudonode is an example of a fake node.  It can emulate all popular versions, and some unpopular like "Classic" and "BitcoinXP", using almost no resources.  It just acts as a relay to fetch the requested data from another node which (hopefully) stores it.  Unless it is another psaudonode, which fetch it from another node, etc.  Perhaps there are only a handful real "Classic" nodes on the network.  We will never know.  The operator who run hundreds of "Classic" nodes on Amazon to attack the Bitcoin P2P network probably want to do it as cheaply as possible, and Pseudonode is the obvious solution.

I disagree. every bitcoin transaction will show up on the blockchain for anyone to see. If I were to send BTC to address "A" and then subsequently send BTC "from" address "A" to address "B" then even if both transactions are confirmed in the same block, then both of these transactions will be visible in the blockchain and the transactions will appear in a specific order. If I were to have a balance of 10 BTC in my coinbase account and then send 2 of those BTC to someone who also has a coinbase account via email transfer, then I never really had 10 BTC, I only had IOU's for 10 BTC, and transferred those IOUs, not actual bitcoin.

I would define a settlement layer as one that all transactions are not transparent. With a settlement layer, for example if coinbase customers were to send 150 BTC to bitpay customers, and bitpay customers were to send 53 BTC to coinbase customers then there would be a single transaction in the amount of 97 BTC from coinbase to bitpay, and not all individual transactions will be clear, nor will they show up on the blockchain.

The average transaction fee increased to something closer to $0.06 to $0.10 per transaction during what was alleged to be spam attacks on the blockchain/nodes. While this is still a nominal amount, both in absolute terms and when compared to the costs associated with using the likes of PayPal and credit cards, it is a significant increase when compared to what transactions have cost in the past. As Seb has pointed out, when blocks do become full and stay full for even the medium term (and possibly the short term), then transaction fees will increase at an exponential rate, and will continue to increase until transaction fees have increased so much so that users are discouraged from sending additional transactions (or until the maximum block size is increased).

Don't give any evidence, or make any kind of factual argument when refuting my points.

Roughly 20 years ago, it would cost a residential user roughly $45 per month for a 28.8 kbs internet connection (for both a dedicated phone line and the fee the ISP charges). Today, I can spend roughly the same amount of money for a 8 mbs internet connection. There are 10 two-year periods between this timeframe. This works out to available internet speeds to residential customers increasing by roughly 75% every two years for the same price, however it should be noted that the US (where I am located) is generally "behind the curve" in terms of broadband access so the increase would likely be greater for the rest of the world.

According to the lighting white paper, for Bitcoin to be able to handle as many transactions per second that Visa has ever had to handle per second, the maximum block size would need to increase to roughly 8 GB. It should be noted however that Visa has not had to handle that many transactions in roughly 2.5 years and usually handles much less transactions per second. If we were to increase the maximum block size to 2 MB today, and subsequently increase it by 75% every two years, then the maximum block size would increase to over 8 GB in 15 two-year periods, or in 30 years.

With my current internet connection of 8 mbs, I would be able to transmit an 8MB block to 10 peers in 10 seconds, and would be able to transmit a 2MB block to 40 peers in 10 seconds.

Don't give any kind of evidence, links or anything like that. I am not aware of any kind of attack vector that you are claiming.

What part of my statement do you disagree with? Do you not think that SegWit will increase the potential amount of data that will need to be transmitted with each block? Do you think that the maximum block size should decrease once SegWit is implemented?

How exactly do you define a fake node? How exactly are you able to claim that the nodes that are reporting as being classic are in fact "fake" by your definition?

Bitcoin is a settlement layer by design.  If you use it for something else, you are doing it wrong.

Totally agreed. Core Devs are the ones brought us here and now, after Satoshi.


As long as now is now...and not in 6 months.

Yes, it either stops adoption or kills bitcoin by making it unuseable.

I somewhat doubt that bitcoiners will jump on some altcoin, I would not count bitcoin forks to altcoins there, altcoins have not the best reputation with most bitcoiners so it probably won't happen.

And surely the lightning network won't take over. By now everyone knows the critics on LN and when bitcoin really becomes unuseable and everyone knows a fork would fix it then LN looks like the evil to even the rest in doubt.

Besides that... LN is not Bitcoin, it is another network which means new things to install and whatever. It won't happen. Things like that happened already alot and even when they were innovative, they don't really played a big role in the end. It's worse for LN when it really rides on the wave of dying bitcoin when it looks like it could have been avoided but wasn't for reasons.

Your right about the exponentially rising fees only if there is no alternative to Bitcoin. In reality people will stop using Bitcoin before fees go to paypal levels. Theymos told it nicely, if ever fees go very high like over 1 USD, increasing blocksize limit can start to be discussed. The point is such high fees cannot happen because it is self regulating - Increased Bitcoin usage means increased fees which means some people stop using Bitcoin which decrease fees - some equilubrum has to be reached, fee size and number of Bitcoin users wanting to pay such fees. Thats why promoting artificially limited blocksize is direct attack to cripple Bitcoin.