Comment by keypusher | Hacker News Reader

Comment by keypusher | Hacker News Reader

Parent Full thread

I used to believe this too, but not anymore. High transactions fees and slow times are a problem, but increasing the blocksize is not the answer. BTC can currently process 7 tx/second with 1mb blocks. If you double the block size, you get to 14 tx/s, it scales linearly. PayPal processes ~150 tx/s, Visa processes ~5000 tx/s, and can burst up to 50k tx/s. This is orders of magnitude off, and would require scaling blocks up to 1GB. By setting the precedent that increasing blocksize is the answer to slow transactions, you open the door to a future where only a very limited number of powerful entities can afford to run a full node, which brings you right back to a centralized currency. In order to scale BTC, it basically needs to move to off-chain transactions or some other new tech. The devs know this, they want fast transactions too, but they don't want to risk centralization to get there.

> they can always lower it later

I don't think this is true. Once you have blocks on the chain greater than 1mb, those blocks must be valid forever. That means protocol must support new blocksize forever.

One problem is that the proposed offchain solutions are centralized.

Also, it's misleading to say that large blocks make bitcoin centralized. Even with large blocks, there will still be enough players that the ledger is still protected through proof of work. In any case the hashing power remains centralized regardless of block size.

You are misunderstanding the meaning of the word "decentralized" when it comes to blockchains. It doesn't mean "available to be run by the masses". It refers to the solving of the Byzantine General problem through proof of work to prevent tampering and double spend, which would still function fine even if only a few thousand nodes existed around the world.

>One problem is that the proposed offchain solutions are centralized.

Centralized off-chain solutions are much less of a problem. Consider paypal. Everyone hates it but they're the only game in town because its network is self-contained; network effects prevent any legitimate competition. With centralized transactions through bitcoin, there are no network effects to prevent an upstart from competing in this space.

Sure but there is no threat of centralization of bitcoin itself.

If Bitcoin fees are so high that you have to use Lightning, then effectively Lightning is Bitcoin and you're potentially trapped in a centralized system.

Neither cost or ability of the everyman to use bitcoin define whether it's centralized or not. It is a decentralized algorithm by design, and will remain so no matter what changes in block size or how many people use it or who uses it or how much it costs to use it.

You are talking about concentration and consolidation at the human and social level, which does not change anything about how the blockchain is designed or functions, and is orthogonal. And nothing is going to change that. The government could shut off all electricity and the blockchain is dead.

> Once you have blocks on the chain greater than 1mb, those blocks must be valid forever. That means protocol must support new blocksize forever.

You could have the consensus rules take the block height into consideration, so blocks after a certain # must be under X MB, but previous blocks can be up to Y MB.

Can't you just increase blocksize to 2-4 megabytes for now until a more scalable long term solution comes into play (segwit)? Sounds like a "best of both worlds" solution.

Increasing the blocksize for now will have the effect of postponing any more scalable solution.

How?

The same way that a temporary solution postpones any definitive solution of any problem. Every time.

And, as the GP pointed out, the sum of practical temporary solutions will have a very bad long-term effect. Of course, that doesn't mean that anybody will change course.

Reply on news.ycombinator.com