Which EVM chain is the right home for Farcaster? Would love to get input here, especially from people working on scaling other chains. https://hackmd.io/@farcasterxyz/rkfR8q0k2
Too early to tell? I feel like everyone is kind of waiting for a consensus to develop
I like L1 for now to keep the sign up friction high as a way ensure quality stays high. By the time it’s time to scale past that then there should be enough progress with zkEVMs and other developments to make the decision more straightforward.
What are your thoughts on the leading L2s? Fuel Labs has a really interesting concept I’ve been learning about recently.
Short term: really depends on team’s GTM strategies, but to keep the entry cost at a certain level before FC is ready for mass adoption seems to align more with FC’s value. Longer term: I would vote for option 6 for more sovereignty and more value capture.
What's the security problem with Arbitrum and Optimism, besides censorship? As long as you have somebody checking fraud proofs, should have the same security as Ethereum, right? Not sure if one can selectively check just fraud proofs that interact with Farcaster contracts. If so, only check txs you care about
Will Farcaster ever have ads? Ad free option for gas fee payers on L1, and waitlist + ads for freebie plan to subsidize the L1 gas fee for registration 🥴
I suspect using an L2 might be the right tradeoff right now.
Stay Ethereum L1 for now and reevaluate later or make your own OP Stack rollup
not an engineer, no bags in op and not scaling nothing but OP Stack looks/feels _just right_ … no?
imho, I would assume an L1-native zkEVM. Ethereum makes sense longterm even with scaling challenges.
I would say 3. It is too early to tell which L2 will dominant. No L2 can guarantee long term stability as this moment.
Option 3 - L1 now and wait to see how other tech stacks mature. As it stands neither OP or Arbi has fraud proofs, and Polygon is still just a sidechain. Combination of L1/L2 seems unnecessarily complicated.
Option 3 - Use Eth L1 now , migrate to L2 later . In my opinion , this will be best for farcaster.
I'd say mainnet atm, but it would require bundling the txs to lower the costs (or find sponsor clients that do that for free to onboard). None of the L2 are fully decentralized yet, although Optimism announced it's their top priority for them. Also on mainnet Devs might struggle with the same issues ENS has.
Big Q you have to answer is how much interoperability you need with other protocols - if you don't need it I'd go custom rollup. Using storage proofs you can still prove stuff about the farcaster ID registry on other chains trustlessly anyway.
Build interchain native then you don’t have to choose an L1/L2. No reason for a protocol like farcaster to lock itself to a single environment when you can fairly simply build for state to migrate or be accessible across chains.
3 seems to give the most optionality. I’m very intrigued by 6, and choosing 3 now still leaves 6 (and other options) open down the road
Option 3 makes the most sense now as there isn’t enough volume on other chains which really can be a problem but obviously that volume needs to be bootstrapped somehow. One thing I’d throw to the mix - since I’m a big fan of Polkadot, their EVM compatible parachain Moonbeam is worth looking into
if it's an identifier, would multichain support be considered?
For option 3 Extension: Extension: instead of re-registering fids, we could just register future fids on the L2 and keep existing fids on the L2. Did you mean to say keep existing g fids on the L1?
How important is immediacy? Proposal: put everything into a Merkle tree and implement a P2P gossip protocol where clients achieve consensus on the Merkle root
Optimism. Not only we could build our own customized roll up, in long run though , while doing so we will be supporting public good.
Given the ease of portability of this particular state, I think you would be well off on Arbitrum as a starting point. I can think of some straightforward logic to quickly migrate in the event of an Arbitrum attack, and it could even be coded in advance into hubs. No need to start by rolling your own chain on OP stack.
I like #3 and I also like how civilized this whole thread is. What a blessing.
how important is EVM to farcaster? would an environment such as Starknet be considered? or impossible to build farcaster with Cairo
Is it not inevitable that eventually the vast majority of actions / data are stored on a cheap chain for content ownership purposes (which means #6 is most likely)? We are building a read-write-own internet.
Option 3 and have Merkle Manufactory subsidize registration with meta transactions (ERC-2771)?
I'm glad to see HackMD used instead Notion for this. No other comment.