I'm unsure if I'd have the time, but a potential 1.0 should:
1) minimize unsoundness; #15 should be done away with
2) have inference friendly API variants (#20)
3) decide what to do with replace i.e. toEffWith (#5, #19). It's supposedly nice to have this semantics, but this is dubious:
1) we haven't come up with any kind of (half-)formal semantics for it
2) this problem isn't known to arise in any practical use cases
I'm unsure if I'd have the time, but a potential 1.0 should:
1) minimize unsoundness; #15 should be done away with 2) have inference friendly API variants (#20) 3) decide what to do with
replace
i.e.toEffWith
(#5, #19). It's supposedly nice to have this semantics, but this is dubious: 1) we haven't come up with any kind of (half-)formal semantics for it 2) this problem isn't known to arise in any practical use cases