> "constantly" in this case means "once every 2 weeks"
No. Constantly means exactly that - revoked transactions for either party cannot be checked otherwise. I didn't even include the need to be online all the time, since otherwise payments don't work either as multiple interactions are required for each payment.
> You can also have a third party do it for you without having to hand over your keys.
And of course that 3rd party is totally trustworthy for some reason and offers their service for free? The 3rd parties still need to run scripts that have to work correctly. We all see how well that one works on a regular basis, what with the multi-million dollar thefts that happened recently due to flawed contracts and scripts.
>No. Constantly means exactly that - revoked transactions for either party cannot be checked otherwise.
Source? My understanding is unless the channel is cooperatively closed, there's a waiting period (ie. the two weeks) before the transaction is finalized to the blockchain. During that time that transaction can be challenged/overridden.
>I didn't even include the need to be online all the time, since otherwise payments don't work either as multiple interactions are required for each payment.
That seems to be a completely separate issue. Short of receiving random donations, I can't think of any situation where you want to make a payment, but somehow aren't online.
>And of course that 3rd party is totally trustworthy for some reason and offers their service for free?
You're making it sound like it's an issue, but I'm not seeing it. Thousands of hobbyists run bitcoin nodes for free, which costs storage and bandwidth.
>The 3rd parties still need to run scripts that have to work correctly. We all see how well that one works on a regular basis, what with the multi-million dollar thefts that happened recently due to flawed contracts and scripts.
ETH =/= BTC
It's far easier for the community to scrutinize one set of code than for everyone to scrutinize the implementation they rolled themselves.
Penalties for contract violations can only be enforced "if one is able to ascribe blame for broadcasting an old transaction."
Broadcasting transactions takes place on the BC, so identifying an old transaction is only possible by finding it on the BC.
The same is true for revocable commitment transactions. Both parties must observe the BC at all times in order to know whether the other party did indeed violate the contract by broadcasting a transaction. Contract violations cannot be detected otherwise.
> I can't think of any situation where you want to make a payment, but somehow aren't online.
Out here in The Real World(tm) I find the opposite to be the case. I cannot think of a situation in which I want to make a payment that requires me to be online. Neither my debit card, nor my credit card, and especially not cash has such requirement. I can buy petrol, groceries, clothes, pay the barber, eat at a restaurant, have a drink at a bar, pay admission for a ball game, etc. etc. all without being online.
Go ahead and call me a Boomer, but we seem to exist in very different realities.
> Thousands of hobbyists run bitcoin nodes for free, which costs storage and bandwidth.
That's orthogonal to the problem of Lightning network routing fees. It doesn't matter who operates a node and how. What matters is that unless you open a bi-directional payment channel, you'll route transactions through intermediaries that will take fees.
> It's far easier for the community to scrutinize one set of code than for everyone to scrutinize the implementation they rolled themselves.
I think you missed something here. Lightning isn't a single centralised implementation.
Just look at chapter 8.4 - once routing tables are in place, there's a wide open door for routing payment channels through profitable routes and tracking, for example.
Routing also requires keys to held online for intermediary nodes for latency reasons, which is another considerable risk since unlike with banking, there's neither oversight nor security regulations in place.
If you actually read the whole story, you'll quickly notice that a lot of Lightning's mechanisms are built on trust in scripts, incentives, and edge software that "does the right thing" (e.g. periodically make backups and checking the other party's state for honesty).
>Penalties for contract violations can only be enforced "if one is able to ascribe blame for broadcasting an old transaction."
>Broadcasting transactions takes place on the BC, so identifying an old transaction is only possible by finding it on the BC.
>The same is true for revocable commitment transactions. Both parties must observe the BC at all times in order to know whether the other party did indeed violate the contract by broadcasting a transaction. Contract violations cannot be detected otherwise.
I'm not really clear how your source contradicts mine. The source you provided is talking about how to ascribe blame, but then you magically turned it around to talk about having to be online 24/7? I'm not following the logic there.
>Out here in The Real World(tm) I find the opposite to be the case. I cannot think of a situation in which I want to make a payment that requires me to be online. Neither my debit card, nor my credit card, and especially not cash has such requirement. I can buy petrol, groceries, clothes, pay the barber, eat at a restaurant, have a drink at a bar, pay admission for a ball game, etc. etc. all without being online.
1. always-online mobile payment systems (eg. alipay/wechat) are quite successful in asia. while I agree being able to pay without internet access is great, internet access requirement isn't some sort of insurmountable barrier.
2. you do realize that even though the credit card doesn't require internet access, the terminal itself does? Also, payment terminals has bidirectional communication with the card via NFC. It's not hard to imagine some sort of NFC based protocol that allows limited information to be communicated between the terminal and the wallet software on your phone, so it can gather the requisite information to make the transaction.
>>>And of course that 3rd party is totally trustworthy for some reason and offers their service for free?
>>You're making it sound like it's an issue, but I'm not seeing it. Thousands of hobbyists run bitcoin nodes for free, which costs storage and bandwidth.
>That's orthogonal to the problem of Lightning network routing fees. It doesn't matter who operates a node and how. What matters is that unless you open a bi-directional payment channel, you'll route transactions through intermediaries that will take fees.
Are you losing track of the thread, or are you trying to move the goalposts? We were previously talking about how watchtowers would be operated/funded, now you're talking about how transaction fees for intermediary nodes?
>I think you missed something here. Lightning isn't a single centralised implementation.
As it relates to scrutiny, how is an issue? Unless there's some fatal flaw with the protocol itself, you should be secure against loss of money, even if the peer was malicious. This is different than difi-project-of-the-day that have one implementation and one protocol (the two are essentially the same thing in ethereum).
>Routing also requires keys to held online for intermediary nodes for latency reasons, which is another considerable risk since unlike with banking, there's neither oversight nor security regulations in place.
What's your objection here? That you can get hacked, or that your intermediaries can get hacked and somehow cause you to lose money?
>you'll quickly notice that a lot of Lightning's mechanisms are built on trust in scripts
yes, that's how cryptocurrencies are supposed to work - trust in systems rather than trust in people. Which is better has already been debated to death so I'm not interested in discussing that again here.
> always-online mobile payment systems (eg. alipay/wechat) are quite successful in asia.
I don't live in Asia.
> you do realize that even though the credit card doesn't require internet access, the terminal itself does?
I do, that's why I emphasised that 'I' don't need to be online. I don't care about the payment processor's requirements.
> Are you losing track of the thread, or are you trying to move the goalposts?
No, I'm not moving goal posts. You said that hobbyists are running nodes, I say so what? That has zero to do with transaction fees.
> As it relates to scrutiny, how is an issue? Unless there's some fatal flaw with the protocol itself, you should be secure against loss of money, even if the peer was malicious.
Have you actually read the resources? The protocol itself explicitly does not protect you against loss of money - it can't. It has to rely on trust in intermediaries and software trying to give security guarantees that the protocol itself cannot provide (hence the need to monitor the BC).
> What's your objection here? That you can get hacked, or that your intermediaries can get hacked and somehow cause you to lose money?
The issue is that the intermediary is inherently insecure (and shouldn't be trusted by design) and in contrast to the regulated finance industry, there's nothing protecting customers from fraudulent or extortive behaviour of 3rd parties. Manipulated routes (no hacking required) can lead to increased fees or delays without any transparency or protection for users.
You might find that acceptable as The System is a sufficient safeguard, but even the Lightning team admits that there are unsolved problems (i.e. monitoring of the BC is required to mitigate these flaws, which brings us back to initial point).
> trust in systems rather than trust in people.
That's an interesting assumption given that systems are created and operated by people. It's trust in people with extra steps minus regulatory and legislative safeguards. Brave new world.
The point is that always-online payments works. At most the problem you described is a minor annoyance.
>I don't care about the payment processor's requirements.
But you otherwise don't have any objections if we can come up with a NFC protocol that allows an offline phone to make lightning network transactions?
>No, I'm not moving goal posts. You said that hobbyists are running nodes, I say so what? That has zero to do with transaction fees.
1. "hobbyists are running nodes" were in reply to the question of who is going to be running watchtowers. You sneakily switching to a separate topic (transaction fees) as a reply to that is moving the goalposts.
2. what's wrong with intermediaries taking fees? Credit cards have fees. Debit cards have fees. ACH has fees. Cash has fees (resulting from costs associated with handling it)
>> As it relates to scrutiny, how is an issue? Unless there's some fatal flaw with the protocol itself, you should be secure against loss of money, even if the peer was malicious.
>The protocol itself explicitly does not protect you against loss of money - it can't. It has to rely on trust in intermediaries and software trying to give security guarantees that the protocol itself cannot provide (hence the need to monitor the BC).
My original point was relating to having to trust peers. As long as you do your part correctly, ie. keeping your keys secure and checking every two weeks (or getting someone to do it for you) if you have channels open, then you don't have to trust them. I agree the requirements are more onerous than on-chain transactions, but I never claimed they weren't, and I think it's a fair trade-off to make for the benefits.
>The issue is that the intermediary is inherently insecure (and shouldn't be trusted by design) and in contrast to the regulated finance industry
The same "regulated finance industry" that allows you to withdraw money from your account with just your account numbers, and it's up to you to wrestle with bureaucracy to get it reversed? If that doesn't count as "inherently insecure" to you I don't know what is.
>there's nothing protecting customers from fraudulent or extortive behaviour of 3rd parties. Manipulated routes (no hacking required) can lead to increased fees or delays without any transparency or protection for users.
1. As opposed to "increased fees or delays" in the form of ATM fees (just one example) that are present in the "regulated finance industry"?
2. With regards to the problem of suboptimal/"extortive" routes. Is this an theoretical problem or a practical one? Are lightning users getting scammed left and right by these malicious peers? If it's free money and as easy as you make it out to be, such attacks should be very popular.
>You might find that acceptable as The System is a sufficient safeguard, but even the Lightning team admits that there are unsolved problems (i.e. monitoring of the BC is required to mitigate these flaws, which brings us back to initial point).
The "unsolved problem" of having to monitor the blockchain every two weeks means... what? The whole project is DOA? That we shouldn't use it? That we should use sink the whole thing and go back to "regulated finance industry"? Is the internet DOA because of the "unsolved problem" that you need to keep your system up to date and patched to not get hacked?
"constantly" in this case means "once every 2 weeks". You can also have a third party do it for you without having to hand over your keys.