Yes, but IOU-based online wallets like Coinbase de facto are banks even today. They will continue to work the same way with LN.
I agree with that view, and I would think that if many people have an on-line account on something like Coinbase, then Coinbase simply can act as a bank, if you transmit coin IOU from your coinbase IOU account to an account of another customer of coinbase. That's pure banking, like today. Coinbase wouldn't even need any LN or block chain to do that, it would be an internal matter of their website, just like with normal banking. So, online IOU wallets are, I would think, not even concerned with scaling, LN, or whatever, as one is not dealing with the crypto token itself, but with a centralized IOU "backed" by it.
But also in the case of hybrid online wallets like Blockchain.info, they can choose to sponsor the commit transaction fee to get you to use their node as hub. That would be already a bit different from banking, I would call it "semi-centralized", because they in no situation can run away with your money if you have a copy of all your private keys.
This is an important point, yes.
Mm, could I require, for instance, to get an e-mail with personal information (KYC/AML), or with the purpose of the transaction, and decide upon that information whether I accept to transmit it or not ?
(I'm trying to find out how much power the nodes can have over the transactions)
As far as I understand, as a node you could do that, but it's very possible that you'll have a hard time against the competition. What is possible is that if e.g. government actors use LN, they could be forced to use such a "trusted" LN node, but for the general public it should be impossible to force that.
I see.
Well, that's what I'm trying to understand: whether this is just a smarter way of transmitting transactions, or whether this sounds awfully lot like banking. It is not clear to me, and I know too little about it.
The big difference is that centralized actors like banks, in theory, can close down and lock or run away with your funds*. With LN, this is not possible, as you - in theory - would be able to retire the complete funds of the channel after the HTLC timeout if your counterparty is trying to scam you.
Indeed, that is still an advantage of LN over banking in the classical way, even if it were totally centralized.
My problem with LN - and the reason why I personally would only use it for small amounts - is the danger that there could be attacks by nodes that act as hubs and after opening a channel and have a few transactions broadcast an old settlement transaction and then try to disconnect their counterparties from the network (e.g. by DDOS attacks) to avoid that you retire all the coins before the CSV timeout occurs. I don't know if there is already a solution for this problem or if I am understanding something wrong.
Well, as I already said, I'm trying to find out for myself whether to be positive or not towards LN, and with all the political disinformation and hype it is very hard to make up one's mind on objective information. I think there are a lot of positive elements with LN, but "deep in my bones" I feel some issues, and two of these are:
- it looks to me that economies of scale will make LN hubs evolve towards a wheel and spokes kind of structure, because the more BTC you can lock up in many channels with many customers, the more chances you have to be able to transmit a lot of transactions through them, which will essentially give you the ratio between the profitable LN fee and the on-chain fee. So I wonder if this sentiment is correct, and has been studied, or if it is not something to worry about ; and whether it is a problem or not.
- however, what worries me most, is that fact that there is not enough room on the block chain to settle large parts of the LN network, if it needs to settle. As such, your "right to settle" is a right on paper, but could very well be thwarted by the practical limit of the room on the chain. In other words, in as much as the LN allows to scale to a large multitude of the number of transactions on chain, it seems to me that this large factor is also at the same time a serious security risk, because it can never settle most of the links on time. However, this is maybe also a gut feeling that is not justified in practice.
So I'm not clear as to whether I should be in favor of the LN or not (not that it would make any difference
), because I can't wrap my mind enough around what it really is going to be.
But I doubt the money transmitter-bank distinction exists in all countries. LN hubs could simply settle in the countries with the most liberal regulation. 
. If you want to scale on-chain to mainstream you need atleast min 1GiG blocks...wablief 
Now you know why dev's are making l2 projects.
