Topic: The Barry Silbert segwit2x agreement with >80% miner support. - page 55. (Read 120030 times)

Then they don't use it.  In fact, "off-chain" is a perfect way to connect to exchanges.  LN is perfect with exchanges as central hubs, and exchange customers having their holdings in a channel with their exchange, which protects them from theft, and renders the exchanges trustless.  It would turn the current exchanges in the new banks.


It most probably won't, unless, of course, in the "exchanges as new banks" world ; bitcoin's main application is "greater fool game", and of course, we need a lot of greater fools, but on the other hand, I don't know how many of them will be gullible - though there's hope, given historical records of frenzy.  When "average Joe" gets into bitcoin, sell everything !


Actually, it may be simpler to get "new masses" to bitcoin by giving them a central authority (their exchange) and giving them classical banking with off-chain linking to their exchange.  It would be more secure too for them, they would feel much less at a loss, it would be standard banking.  There may even be a law that only off-chain transactions are officially allowed, and that you need a licence for on chain transactions. 


Bitcoin is broken beyond repair.  The first thing to fix in bitcoin, is the emission curve.  A decreasing coin emission renders every coin unstable and/or centralized.  The second thing to fix is proof of work consensus.  These are the two principal value propositions of bitcoin, and they render bitcoin unfixable if you don't modify them - which goes against bitcoin's belief system.
Only a temporary fix, that is changed regularly, can keep it afloat for a while, and this might very well be sufficient, because bitcoin doesn't maybe need to scale to very very large sizes.


Sure, but the LN could transform bitcoin into classical "banking" (but rather, classical clearing house transfers).  Bitcoin is not money, but is a speculative betting asset, and that betting is mainly done on clearing houses (exchanges).  Now, the number of players may become large, but has nothing to do with the size of payment networks.  Bitcoin being more like "complex derivatives for the people", the LN could be a way to tie yourself to your clearing house and render the relationship more trustless.

It's funny that the rationalisations used on either side gets accused of being based on tiny, limited scopes =).

As someone who will never be fully initiated into the realm of "knowledgeable" I have to say that both you guys say what is genuinely on our minds. What we do understand of this allows us to want a way forward, but we prefer solutions that say: "this is how we will move forward, and this is what we will do when these likely issues come up" instead of "we won't do that because that sucks".

Amen on realpolitik. No one is deluded to think there is one final solution.

I'm still not sure you understand the basic premise of what LN is. LN only works if you have a "thin client" that creates a valid Bitcoin transaction. I'm at a loss as to why you, and so many others, want to create new clients to do exactly what current clients do. The only advantage that LN provided to that average user is the rare instance where they want/need to make voidable transactions (which is more anti-Bitcoin than I care to discuss). It is 100% pointless to use LN to make an immutable bitcoin transaction (since the end result of any finalized LN transaction is, in fact, the same Bitcoin transaction that one would have made using Bitcoin).

I'm still not sure why you're stuck on these insane ideas that:

• The protocol should be limited so that people that don't need to run nodes can run them anyway using dialup and a PC less usable than an Apple Lisa
• LN has anything to do with normal Bitcoin transactions

The Btc1 repository seems to contain the coding work that is being done. I've not checked it but according to other users, the code based on the agreement is progressing.

(It seems unavoidable that in Bitcointalk the thread degenerates in some form of "small blockers" vs. "big blockers" and sometimes vs. "moderate blockers" discussion, but I estimate that in this forum only very, very few development decisions are being taken.)


So current pruned Bitcoin nodes do not have to download & verify all blocks anymore? OK, that would be new for me. Or are you talking again about the (irrelevant) HD space?


I am convinced that we need a form of sharding (splitting the Bitcoin transaction data in smaller sections, so not all full nodes must verify all data) - be it sidechains, childchains, extension blocks, or a traditional sharding solution like Ethereum is planning (if it's not vapourware), if we want to scale up. I, for myself, don't want a coin where only datacenter and rich users could run a full node. And yes, I want a LN-style micropayment solution to pay for coffee and other small things. But I also don't want LN as the only second layer as some Core Maximalists like it.

I'm not even against Bitcoin Unlimited, but I would really like to see EC tested in the wild - as an altcoin or a "cleanly forked" Bitcoin (based on a UTXO snapshot) to see like it behaves, until it is introduced as an option for the main BTC client.

I think that is my final post to the big/small block issue because I think we are desvirtuating the thread.

I was tired of any, and all, LN arguments even before they were made. LN and nodes have virtually nothing to do with each other and the existence, or non-existence, of LN is moot to any part of this conversation.

• Given that a pruned full node is @ or < 2GB, your math is way off.
• Be careful of your disdain for alts, because if some folks in this thread have their way, NotBitcoin is the new path forward.
• Even Moore has said that the end of the "law" has come (not that I ever use it as a claim).
• Moore's Law was an observation of something that has no relevance (not that I ever use it as a claim).
• I didn't mention CPU and RAM, because per required unit they are even less expensive than storage.
• No one in the random 3rd world country you are positing needs to run a full node.
• No one in the random 3rd world country you are positing needs to care about the hardware requirements to run a full node.
• Your argument about nodes vs the "underprivileged" (for lack of a better term) is still moot, because users of Bitcoin don't need to run full nodes.
• If you still can't grasp the last one, then you likely don't understand Bitcoin as a protocol.

I am seeing this is back on the front page. Doesn't seem like this went anywhere... Unfortunately.

Where does this stand now, anyone have a good link or info?

Last time I thought to myself I will believe it when I see it, I still don't see it happening.

What more can we do beyond making posts? Are we ready to start firing up miners to get some percent into a pool and push it over the top?

People that don't run a full node can also use LN for their transactions. In terms of "trustlessness" it's comparable.


Altcoins do not count. If you run Bitcoin + 14 other coins, the restant 14 altcoins (even if ETH is among them) would probably not even eat more bandwidth than one single Bitcoin node, so you are barely using the same bandwidth needed by full Segwit blocks with 1,7-2,2 MB.

Try to run 8 parallel Bitcoin full nodes, then your argument becomes more credible. And don't cheat interchanging the blockchain files And: Even your humble 10 MB are only a dream for >70% of the world population.

And you didn't mention CPU and RAM. (Moore's Law argument incoming ...)

Keep in mind that not every potential user needs to run a full node and that you're trying to make a bandwidth argument to someone who will never have more than 10MB down (or 0.75MB up) anytime in the next 15 years, runs a full node of 15 different coins, streams HD video 16+ hours every day, and who's ISP's legal name is "Hillbilly Wireless" (because he lives in a barely-connected small town of ~600 in a county of less than 15,000).  

HDDs are the least important problem. I think you know that

The problem is the initial blockchain download (which consumes bandwidth, CPU and RAM usage) and the block propagation/keeping in sync. Keep in mind that not every potential user lives in highly-connected global cities in the first world.

Perhaps they shouldn't be continuing to buy 17 year old Quantum Fireball 4500RPM 40GB HDDs? Beyond that, I still fail to see the legitimacy of such a claim.

Come on, that really has been discussed ad nauseam here

The danger (I implicitly mentioned it) that common people - and above all, small businesses - could not longer be able to host a full node with consumer-grade hardware is enough for me. We would hinder people to use Bitcoin in a trustless way -  they would be forced to use SPV clients or even centralized online wallets to participate. In this case we could also switch to LN, it would be the same level of "trustlessness" (I don't know if that's an English word, but I think it's understandable what I mean). There is no need for conspiracy theories about evil miners and banksters behind them to come to the conclusion that big blocks alone aren't that good.

The scalability problem is not trivial to solve. That's why there are so many discussions about it. And now we're caught in a stalemate. That's why I favor a short-term fix if there is no consensus regarding a long-term fix, rather than discussing the long-term fix the next years.

And they are....

...
Which goes back to my post that offended you; dreams, wishes, and delusions are what support a short-term fix rather than an actual fix. It's delusional to think that the dreams of longevity are supportable by the wishes of breaking it all again in the future to almost fix it now.

I partly agree - I had favoured a conservative, but dynamic approach like this one for future blocksize scaling. I also agree that LN is not the Messiah some are preaching about - it's a solution suitable for certain kinds of micro- and minipayments, not more. I regard it as the equivalent to a (general-purpose) prepaid card in the cryptocurrency world.

But the dangers of too big blocks cannot be ignored. I know the importance of full nodes is sometimes exaggerated, but for the individual users they are the most secure (and the only really trustless) option to participate in the Bitcoin network.

And sometimes, I think a little bit of realpolitik is needed. It's obvious that 2MB+Segwit (or Segwit alone) isn't a final solution. It's a dirty hack, but it solves most short-term problems we actually have (that were my previous posts about). We will very probably (no mathematical proof, but I'm 99% sure) have some air after it to find a better governance model and a long-term scaling solution.

Remember that time, oh so long ago (like last year), when there weren't enough transactions to fill all 144 blocks with 1MB worth of transactions every day......
The problem with so many of the rationalizations I see are all based on tiny, limited scopes and the fundamental change of everything that is Bitcoin.

Segwit + 2MB + LN = all problems are fixed and the world is a happy, shiny new place where growth will no longer occur.
Not!

What if the current masses don't want their transfers "off-chain"?
What if Bitcoin grows to something even in the earshot of "mainstream"?
What if the new masses understand the tech don't want their transfers "off-chain"?
What if we fix what's broken about Bitcoin, in a way that it remains Bitcoin, and the fix isn't just some half-assed bandage that's going to need to be changed in 6 months, 1 year, or 2 years?
What if we say, "Core devs we accept no half-way measures, any fork must be future-forward looking"?


Edit: Have I mentioned, at any point, that LN is an external service; not needed (and even pointless) in the vast majority of current Bitcoin usages; and, in the end- still requires the "final" transaction to go into a Bitcoin block the same as a "standard" transaction?

Yeah, it would take a couple of blocks to clear the current backlog. But as I wrote in the previous post, the 2MB blocks with the actual transaction rate (3,5 tps) would be only full for 60%. Let's add 10% more for the "transactions that [..] aren't happening because of congestion" (I repeat: the 3,5 tps are the transactions added to the mempool including transactions that are never confirmed).

So we would have 600 kB per 2MB block to reduce the backlog. Assuming that per MB of block capacity about 2000 transactions can be processed (at an average size of ~500 bytes), a backlog of 100.000 transactions would need about 150-200 blocks to be cleared. Bitcoin processes ~144 blocks per day. So basically in a day and a couple of hours the backlog would be reduced to zero.

If there is a fallacy in my argumentation, please help me to find it
 

OH, you mean the post where you ignored that it would take 50 2M blocks just to clear the current mempool if no new transactions ever occur, that "everyone" claims there are transactions that don't occur because there's a backlog, and where "everyone" ignores the amount of transactions that they claim aren't happening because of congestion? 

Well, here I think I provided strong arguments to back my claim that it is very probable to almost sure that in the case of an introduction of Segwit users would pay less transaction fees (I know it's not a mathematical proof, but that's impossible if we're trying to predict human behaviour).

Your next post could be interpreted as an accusation to me to base my stance on "dreams, wishes, and delusions".

I did what, to who, about what? 

Yes, you're right here. It's even getting worse measured in USD: As Bitcoin has doubled in price since the end of 2016 the USD value of the average transaction has doubled.

But I'm sure that if we looked at the median, it would be much lower than the average (does anyone know the charts/numbers here?). I often observe a lot of transactions in the 0.01-0.1 BTC range. 0.1 would be a little high for LN, but the lower end (0.01-0.04, as LN has, if I remember right, a limitation to 0.042 BTC) could be covered with lightning transactions.

The current high average transaction amount may also to be related to the fact that businesses like exchanges may bundle a higher number of transactions together than they used to do before.


Agree. But there is a large subset of the OP_RETURN transactions that could use LN (for example, Counterparty asset and other coloured coin transfers). A large number is using just the "immutable" feature provided by on-chain transactions (e.g. Factom), that's true.

Well, looking at blockchain.info, one can see that the number of daily transactions has been rising slowly until about 300 000 transactions per day, and that the average estimated daily transaction VOLUME is about 250 000 BTC daily since quite a while.  So we see that the *average* bitcoin transaction transacts about 1 BTC.  I have no idea of its distribution, but apparently a transaction on chain is around 1 BTC : are there 240 000 transactions of micro-amounts, and 10 000 transactions of several tens of BTC, or are all of them turning around 1 BTC, I don't know.  But the average transaction doesn't seem to be a micro transaction.
Of course, the bitcoin block chain is also used as an immutable ledger for other applications than transactions ; these will not go away.