Topic: What if all transactions go 'Off Chain'? (Read 1710 times)

Maybe we agree. To clarify, there are a couple of different questions about block sizes.

Q1) How big would it be helpful for blocks to get? Opinions being:
 a) 1MB is right for a long time, for reasons to do with mining on Tor or something.
 b) Ideally smaller than the amount of non-spam transactions people want to make, to raise transaction fees by creating artificial scarcity.
 c) Ideally bigger than the amount of non-spam transactions people want to make, to make Bitcoin as attractive as possible and create lots of transactions to get fees from, which should in turn raise the value of the block reward.
 
I take the OP to be talking about (a), when it becomes a case of (b). In other words, say loads of people want to make transactions, but the block size is stuck at 1MB. Long-term does that mean more fees for miners (higher per-transaction fees!) or less fees for miners (fewer transactions, less on-chain activity, possibility that everyone switches to another coin!). What's happened in the last two years doesn't really tell us anything about that, because we're at (a) and (c), not yet (a) and (b).

What hasn't really been discussed on this thread (which is fine, because it's been discussed on plenty of other threads...) is:

Q2) Who should decide how big blocks should go?
 a) The community, but in a way that's weighted to Gavin / Developers / Miners, by the current model where the core developers make a proposal, stick new defaults in the client, and it gets adopted if the miners will run it.
 b) The community, by some kind of voting mechanism - for example, some people were proposing a plan involving proof-of-stake voting with transactions, to reduce the freedom of the developers and miners to make changes. (The plan as stated also seemed to be designed to prevent any changes actually happening in practice, because it counted the "didn't vote" of people who don't follow Bitcoin politics as a vote for "no change"...)
 c) Miners, by a dynamic consensus process: You can mine whatever you like and broadcast it, but if it's too big to handle the other miners will ignore it.

[Minor edits for clarity]

Not if you mix it

But anyway, I definitely agree with this point:

From a regulatory point of view it's completely different to trusting a third-party service that actually handles the transaction. It's also much easier to distribute (other people can run mirrors of the directory) and validate (since you can check what addresses I'm giving out, and blow the whistle on me if I get cracked or start pulling shenanigans).


I also do that if I stick a Bitcoin address on my website or in my signature. (OK, I don't necessarily put my email address on my website or my Bitcoin Talk signature, but anyone who can use Google can connect the two.)

If you don't want that you shouldn't use my public directory, but you could do it with a non-public directory that fed a new Bitcoin address to each user, while keeping the transactions on-chain. I guess one decentralized way to do that would be that you'd stick an "address server" URL in your DNS, which you could run yourself, and have that feed a new address to each person who shows up.

PS. I'm not saying there isn't value to off-chain systems for a lot of use-cases, just that off-chain isn't the only way to do it, and has a lot of downsides.

It is also Satoshi who set the block size of 1MB, as well as a soft cap of 250 KB.


Clearly that hasn't happened for the past 2 years.

If you don't cap it through, a rogue miner can publish a 1 TB block and kill the network.

If you say 'oh then have a limit', then you're back to capping it.

Transaction fees are needed to keep the network secure and provide an incentive against 51% attacking the network. The block size cap protects Bitcoin.

Is 1MB is the right limit? That's another question.

The people who want to cap block sizes tend to see it the other way around, because if there's no competition for space in the block, transaction fees stay very low. So the thought is that by arbitrarily limiting the block size, you create artificial scarcity and push up transaction fees. The way they see it the off-chain processors still end up clearing transactions in the block-chain, but the transactions they clear are much bigger, so they don't mind the $1 or $10 or $100 transaction fee.

I'm not convinced that's really what would happen. There are a lot of problems with using off-chain solutions for some of the use-cases that Bitcoin currently covers, not least regulatory ones if you try to do this globally on any scale (even the massively-lawyered PayPal can't operate in a lot of countries, and they need very tedious identity validation procedures in others). Many of these problems are the ones Bitcoin was specifically designed to avoid.

I think what really happens when you try to artificially cap Bitcoin transactions is that everyone sods off to an alternate coin without that limitation, and your Bitcoins become a historical curiosity.

This is no different than trusting a third party service. Do you seriously want to broadcast all of your transactions to the whole wide world and tie them to your email address?

The problem with this is that anonymous transactions (eg a bet) can be double spent without any adverse action on the sender, because he is anonymous.

Zero confirms work great if you're sending Bitcoins to your friend or buying a cup of coffee (this isn't taking in account tx / dust fees through), but if you are running a web facing app that allows anonymous usage.. you're going to go bankrupt if you accept 0 confirms on the blockchain.

Impossible for that to happen.

I don't think any of the things you mention to compete with PayPal actually require off-chain transactions, although there are a lot of cases where they have a role.


This can definitely be fixed with a trusted wallet service, and possibly even with some client design ingenuity.


This can be fixed with an email -> Bitcoin address directory with a public API. I made one called Address Machine:
https://www.addressmachine.com/


You have to wait 10+ minutes for transactions to be _irrevocably_ confirmed, but with PayPal the relevant number is more like 3 months. Most practical situations where you need instant transactions can probably get by with zero confirmations, although the issues around this aren't quite trivial. If zero-confirmation transactions _do_ turn out to be too risky in practice for widespread use, there are other ways of working around this without entrusting the whole payment process with a trusted off-chain third-party.

You are aware that we (Inputs) don't charge a 1% or whatever fee compared to BitPay, Coinbase, BIPS, whatever?

Trust is always optional. The blockchain isn't going away, and there always will be bitcoin addresses and public / private key cryptorgraphy

Then that would suck, wouldnt it

Not possible. As far as my understanding goes.

But it's then becoming more and more tempting to be the one who centralizes most.
And random users will be back having to trust inputs.io or any similar place.

And when those places change their homepage, explaining they are now taking a 1% fee, random users will pay up.
Just check what power your banks have over your life today, and you'll know the rest of the story.

Good try, but you're missing something in this analysis.

Off-chain transactions can work, because anybody can set up their own 'supernode'. Therefore it keeps the P2P values of bitcoin.

Sure, bitcoin becomes more centralized, but still a hell of a lot less centralized than our existing banking system.

good evening, Happy to clear it up for you. Offchain transactions require trust. Bitcoin works because it functions using an advanced systematic indexing system. The larger the network the more confirmations are possible. This number is correlated to the degree of reliability.  We are always looking for ways to expand and build up asicause nodes. Feedback is always welcome and helps to boost morale in DM communities who are still byte-dark...Ira

Off-chain solutions MUST complement on-chain transactions in order for Bitcoin to be mass adopted.  It seems like many in the bitcoin community suffer from severe tunnel-vision, and have lost complete sight of the big picture.

What is bitcoin competing against?  We are competing against government-issued currencies.  

What can people do with government-issued currences that they can't do with bitcoin?  They can exchange them INSTANTLY, and store them safely in a bank (3rd party).

In order to compete with the likes of PayPal, we need bitcoin to be as CONVENIENT as PayPal. Currently:

- Maintaining a secure bitcoin wallet has a higher level of difficulty than maintaining a PayPal account.  
- Sending coins to a public key is more confusing than sending money to an e-mail address via PayPal.
- Waiting 10+ minutes for transactions to confirm is 10 minutes longer than PayPal transactions.

While the blockchain offers decentralization and pseudonymity, it must be complemented with off-chain solutions if we want people to be able to use bitcoins with the same ease as, say, US Dollars / PayPal.

good day, What you are suggesting would require the user to trust the offchain operator. You need to reread Satoshis' white paper. Bitcoin was designed to not rely on trust, I works because I depends upon proof of work confirmations within the peer to peer CPU user pool. Any questions?..Ira

ultimately the transactions must be settled in the block chain.

if some technology is claiming to take transactions completely off the chain, then you must consider the weaknesses inherent in the technology in question.

That makes no sense at all. Seals does on the chain transactions for cashins and cashouts and off the chain transactions for bets and player transfers. Not using each when appropriate would be stupid.