Thanks everyone for engaging so much on this thread. Will focus this reply on the DEX part. I think there are three main questions:
- Should Statemint have a DEX? What would the goals be?
- If yes, what properties should it have?
- Who should build it? (And, what is “it”? What are the objective functions?)
First, I see this brought up all the time. Some examples here, here, here, here. I think everyone wants the Treasury to award proposals in stablecoins (one discussion here). I’ve heard from investors that they think Polkadot is not a viable ecosystem because it doesn’t have a DEX in the core protocol, and this hurts teams trying to get funding to start their project on Polkadot. SR Labs has also raised a few issues (one related to fee payment mentioned here), all of which are solved by the ability of synchronous asset conversion. I’ve argued against a Statemint DEX for a long time, but I have to say now that I believe any time not having one is hurting the entire ecosystem.
Second, I think a DEX on Statemint, as part of the Polkadot protocol must be trustless (ideally no privileged functions, and any of them granted to Relay Chain governance tracks), credibly neutral (not prevent any asset from being used), have a third-party security audit, and be usable with only DOT (but possible with other assets too).
Third, who should build it? But first, what is “it”? As people have pointed out (I think @alice_und_bob said it best), a DEX pallet is the easiest part of the entire thing. It needs tooling, an XCM interface for interaction with other paras, compatibility with permissionless access means (light clients, wallets), UIs/access points that cannot censor users/transactions (or at least enough access points such that users have a credible expectation of always finding one avenue), liquidity participation, and probably more.
I never said that Parity would do all that. And that’s one of the reasons we’re having this discussion. If it has any hope of meeting its intended goals, it needs effort from a lot of different groups. Yes, Parity can make a pallet, but that’s only a fraction of the work. There’s really no argument or delusion from me on this point.
Finally, I understand DEXs compete for liquidity and users. And I know that the simple xyk style is not optimal. A DEX on Statemint should optimize for security (buying execution, absolute permissionlessness), neutrality, and a means for the Treasury to reliably grant proposals in stablecoins. But I don’t expect it to optimize for competition with other options in the ecosystem.