Interesting question / idea: "allowing a tx to only lock part of the state trie, and pay for how much each locks"

Posted by @mhluongo on a Twitter thread:

Pretty sure this was discarded years ago in the annals of Ethereum L1 design, but have we totally given up on allowing a tx to only lock part of the state trie, and pay for how much each locks?

It’s crazy to me that every tx gets a global lock by default. Being familiar with a UTXO model, I remember I was shocked. No wonder gas is high — you should have to pay out the nose for that privilege.

Paying for the scope of a tx’s state lock would open up more parallelizable verification, etc, but it’d also be more fair to folks who are just pushing tokens around and are only touching a single contract.

It’d also be satisfying to see our flash loan arb friends pay for the superpower.

@prestwich responded:

there’s some movement in this direction with eip 2930. But it’s not sufficient. Stateless eth had some shakeups lately and I wouldn’t count on it being a thing for several years