Conversation

So tbh the best thing you can do to counter this on a DEX is to make it fast. Uncertainty over 400ms of ordering is soooo much less bad than uncertainty over 5 minutes. This is just so much less of a problem on Solana than slower chains.

Quote Tweet

Matteo Leibowitz

@teo_leibowitz

Aug 29, 2020

‘if the DEX isn’t being front run by the validator set, is it really decentralized?’ twitter.com/teo_leibowitz/

Replying to

and

Still, the entity creating the block can frontrun any transaction in that block. Don't suppose there is any way around that without a trusted third party.

Replying to

yeah alas it's pretty hard. You could try to have timestamps on them, there are proof of time things in Solana that might help, but it's messy. In the end you do the best you can, and at least this is way better than the status quo.

Replying to

and

It also comes down to: why do you want to use a DEX? Because you don't trust centralized custody or don't trust centralized execution? For most it's probably the former. So maybe a DEX where custody is decentralized but validation is done by a trusted federation might work.

Replying to

and

I think Serum or something similar is the best you can do for "fully-decentralized", but there are a lot (majority?) of people who like you said only care about non-custodial. For those people check out our new version launching soon.

idex.io

IDEX High-Performance Decentralized Exchange

Never compromise performance or security. Buy, sell, and trade cryptocurrency on the world’s most advanced cryptocurrency exchange.

Replying to

and

how will it compare to e.g. loopring?

Replying to

I’ve tried both Idex 2 (demo) and LRC. imo Idex will be more fluid and faster due to optimistic vs ZK Rollup in LRC. Found this when checking out reddit scaling entries. This highlights the differences medium.com/matter-labs/op

Replying to

and

wait is loopring actually not optimistic? isn't that insane?

3:21 PM · Aug 29, 2020Twitter Web App

Replying to

ZK rollup. It's all about how you "prove" to the network that the contents in the merkle root are legitimate (fraud proofs in OR, zk proof in zk). ZK proofs are pretty solid, though cryptography is relatively new/untested and there are some goofy UX things (additional keys, etc.)