Hacker News Reader

Comment by drKarl | Hacker News Reader

IPFS is pretty awesome, and the ideas it builds upon like a Merkle tree DAG are really cool. I love the idea of DApps or Decentralized apps. I've played with some cool projects like MetaMask etc. But I really don't like the idea that all (most?) DApps are based on Ethereum and require huge amounts of Gas to run. Can't we have DApps without Blockchain/Ethereum?

I know IPFS doesn't use Ethereum, and neither do things like SSB (Secure Scuttlebutt, based on the gossip protocol) but the biggest ecosystem of DApps is based on Ethereum and requires gas...

View on HN

You absolutely don't need some crypto-based financial scheme to run decentralized web applications. Just some kind of open standard & a set of motivations and incentives to keep the system running and improving.

American ideas of decentralization tend to make the root assumption that financial incentives are the only thing that drive human motivation, so almost every decentralization project gets "blockchained" eventually, at which point progress on the core standards or applications gets halted because the attention, focus, and energy shifts to figuring out the myriad, predictable problems with adding "operate a global financial system" to your JIRA tickets.

I don't really know any way out of this bottleneck for Americans; they're simply too obsessed with complicated financial schemes to make clear progress on usable dApps.

I have slightly more hope for Europeans; if you notice the Matrix project has been chugging along for years, with millions of users, and is poised to actually compete with big tech products via traditional funding and company operation, like normal people.

What you call "american ideas" is the only thing that works in the anonymous environment. Communal solidarity works fine in small social groups where "having something at stake" that keeps people cooperate honestly is measured in many things beyond money, which includes the multi-faceted hard-to-quantify reputation. But if you want random people to cooperate on the internet, you can't have any of those properties but stupid cold cash.

And, if you try to recreate reputation mechanisms, consider how exposed people would be with persistent perfectly precise digital identification of their cooperation efforts whenever such efforts cross interests of some other group. In non-internet life a lot of things are understood and agreed upon without making those decisions public. Even in private contracts, parties prefer that the details of the contract to remain known only to themselves unless (in rare and unfortunate circumstances) their dispute warrants participation of an arbiter.

> What you call "american ideas" is the only thing that works in the anonymous environment.

What about BitTorrent or its various file-sharing predecessors? It has no cash, they had no cash. Or Tor? Exit nodes don't demand money as compensation from attracting the attention of people in authority.

Someone has to pay for the resources. If you're going to compete with Facebook, then someone has to pay for billions of dollars worth of infrastructure. Voluntarist P2P can work to an extent when everyone donates their computer to the task, but a better way is to allow people to pay and let experts run the nodes. People get better quality service, and experts can make a living by providing the service.

Voluntarist networks can't guarantee quality of service. You just get what you get. If you want to have 1 Gbps downloads and 100% content availability, you have to pay someone to provide that.

I 100% get that, and fwiw I'm fully supportive of any scheme that eventually works out to fund traditional app development.

I admit I was more skeptical in ~2017 when these grand claims were made, and it seemed like you needed to be deeply up-to-date with latest crypto happenings to be able to navigate the 'blockchain dApp' world.

I'm very supportive and admiring of things like Protocol Labs, which have put their money where their mouth is to translate the windfall from the Filecoin IPO into projects that make more sense to my traditional webapp brain, like the latest Matrix. So genuine kudos to everyone involved, I'm happy that this model exists and the ecosystem does eventually deliver in some cases.

I have to say that I don't know any blockchain P2P-networks that have succesfully done this (Sia/Skynet maybe). I don't advocate for 'token for every app' or ICOs, which end up being ponzi schemes. Currently I'm more interested in what's being built on Bitcoin and Lightning Network, such as https://www.impervious.ai and the ecosystem on https://getumbrel.com

It is hard to get away from a gas based system because of spam and resource limitations.

Regarding the expensiveness there are a lot of new technologies rolling out today such as optimistic rollups(arbitrum/optimism) and zkrollups(dydx,zksync) that allow for cheaper dapps to run. Then there is also things like Avalanche or Solana that are completely separate chains with different decentralization trade offs that allow for cheaper transactions. The space is still going through rapid advancement and experimentation.

> It is hard to get away from a gas based system because of spam and resource limitations.

I went to a Hyperledger workshop just to get a more objective opinion on it.

All the demo functions were showing us how it was possible to threads and put the state on an internal blockchain, but that all nodes had to run the threaded functions we wrote.

hmmmm, why is this taking so long, oh because everyone in the class was spamming the nodes with arbitrary executions! with no anti-spam resource it was as actually dumb as the theory suggested.

that said, I've come to accept that the market can bear many things in the distributed ledger space, so even if other theoretical solutions can make participants cooperate, an out of the box solution is here that the world is selling so hard that it is easy to sell itself.

I guess http://gun.js.org/ goes into that direction.

I was investigating GUN for a project last year. The core developer, Mark Nadal, is really inviting and responsive. The project does a good job of publishing numbers and research to prove the technology is a viable option to classic centralised architecture that developers are currently paid to produce. Definitely worth a look.

I looked into their [0]codebase and I'm concerned about stability and maintenance.

0] https://github.com/amark/gun

https://github.com/amark/gun/blob/master/src/chain.js

There are a bunch of EVM-compatible platforms that can host dapps now, and many projects are running cross-chain or have equivalent projects on those. Personally I think Fantom is worth taking a look at (and as a disclaimer I hold some); it already is handling more transactions per day than Ethereum and has a massive dapp ecosystem, NFT marketplaces, etc. Gas fees are typically 1-2 cents (simple transactions) or 2-6 cents (contract interactions)

Others worth looking at would be Avalanche (also PoS and DAG-based) or Polygon/Matic (which is L2)

Further away from Ethereum, the launch of smart contracts on Cardano from a few days ago is the beginning of dapps there as well. The fees on Cardano are getting to be prohibitive for casual usage unfortunately (minimum $0.50 per transaction), but what's nice is that the fees are deterministic, and you can't have a failed transaction that consumes or runs out of 'gas' without finishing (though they don't actually use gas)

(further disclaimer, I actually hold some of all of the above + Ethereum... but I guess it's fair to say I hold these because of the potential of what they're doing)

Only operations that change the state of the Ethereum network require gas. You can use an application and as long as you only perform read operations, you don't require any gas.

For example, in a decentralized Youtube-like application, you can browse and watch videos without using any gas. You would only need gas if you wanted to add content to the system, like uploading a video or posting a comment.

If such an application required no cost to write data, what would protect the system against spam? Who would bear the costs of operating the application?

>But I really don't like the idea that all (most?) DApps are based on Ethereum and require huge amounts of Gas to run. Can't we have DApps without Blockchain/Ethereum?

Previous generation dapps were mostly on Ethereum because it was the only real solution available. But the ecosystem is maturing, there are chains that consume so little gas it's effectively a negligible cost, and more and more dapps are making use of them. Not to mention that Eth 2.0 is coming soon™.

>there are chains that consume so little gas it's effectively a negligible cost, and more and more dapps are making use of them

Can you give some examples? How far along is the development?

It’s important to note that these chains that are cheaper to use have to make trade offs to enable that. The trade off they choose is generally to be more centralized. A better approach over using a new “cheap” blockchain is to use a Rollup network on top of a decentralized blockchain.

Rollups inherit the security of the blockchain they are based on, and a user can always leave the rollup network if they become censored.

The downside of current Optimistic Rollup networks is that many still have centralized points of control in place (such as arbitrary contract upgrades), but those are expected to be removed in the next few months for at least one of them.

Zero-Knowledge rollups are also coming out and have key advantages over Optimistic Rollups, including much lower costs.

DAG based networks (rather than blockchains) such as Fantom and Avalanche seem to be emerging as a way to address scalability issues found on blockchain-based smart contract platforms. Currently, Fantom is doing more transactions per day than Ethereum, with gas costing 1-2 cents for simple transactions and 1-6 cents for smart contract interactions.

The tradeoff, based on my understanding, is that there's not as much resilience to node outages (on both of these, 33% of validators by stake concentration going down could lead to block production halting)

Issues like “33% of validators by stake concentration going down could lead to block production halting” is exactly why systems like DAG-based networks would make great Ethereum L2 networks.

They can have speed and low cost, which inheriting the security of Ethereum. At the same time, if block production ever stopped (like we just saw with Solana), all users can still retrieve their funds by forcing a withdraw back to the Ethereum mainnet.

Networks with low attack thresholds shouldn’t be operating as L1s.

I'll speak for myself, as I just finished developing a fully functional dapp on the Polygon blockchain. Transactions are usually confirmed <30s (not fantastic, but acceptable) and cost less than $0.0001 worth of gas each. There are also other solutions which may work better for certain dapps. For example there is Arbitrum, which has much higher gas costs but instantaneous confirmation times. Regardless, the ecosystem is moving fast, improvements are always coming soon™ and new solutions are popping up all the time. But what's right here, right now is already good enough to be usable.

Not looking to get into a crypto-war with anyone, but I suggest looking at Cardano which just launched smart contracts. They have much more reasonable gas requirements that are calculated upfront, and there’s a lot of space for innovation as they released this week. Fun reason to look into Haskell too.

There's RGB which is built on Bitcoin Lightning Network. [0] It's not ready for use yet.

[0] https://www.rgbfaq.com/what-is-rgb

Polygon

Reply on news.ycombinator.com