So many really important concepts and techniques Haskell makes available aren’t even representable in Rust, and it makes me sad because I really want Rustaceans to be able to think those thoughts.

It’s definitely grown out of the Haskell/pure/dependently-typed FP world. I do wish more people would spend a little time writing Haskell, just so more of the good ideas can be imported to other languages; Rust has already benefited heavily from this dynamic but there is SO MUCH MORE.

Ah okay, sorry, I misunderstood. That makes sense and seems much more interesting.

That is: I'm not sure the kind of guarantees you get from gradual typing really get you that much closer to the goal in the end. When you can nigh arbitrarily decide what is "worth" proving, it makes it impossible to really give any stronger guarantees about the system as a whole.

I know people who have put a *lot* of work into gradual typing research. In the end, it didn't seem like they had much to show for it. I'm not convinced it's the way. This might be a strong opinion but I'm not sure there's really room for a stepping stone here.

Whereas if I update and deploy a new dep version in some JS or Python app without any kind of compile time indication that there's incompatibility until runtime, it's less fun to have the rug pulled out from under me if my users start seeing errors/wrong behavior unexpectedly.

Interesting question.

I imagine the tooling in question also dictates this somewhat? In a language that I can reasonably trust the type system to tell me if a change breaks me, I'd imagine it's more "okay" to purposely break in a minor release for a CVE fix.

This is what I think too. I think when I say it there's a difference, but for me it's very subtle.

I didn't get that sense at all, reading the thread. The author says they are amenable to putting out a version that fixes it, but that yes it would be a breaking change since it would change existing behaviour. It sounds to me like someone just needs to open a PR and the author would merge it.

I think there's still an implicit contrast in that usage, but it's more subtle. Like, it's hinting at "the artwork is impressive.. kind of... compared to what it could have been, at least."

Or "The art is impressive per se, independent of its historical significance" - contrasting the artwork itself against any historical value it may have.

I think where it gets tricky is that it's also used just on its own. "The art is impressive, per se." - and here to me it implies hesitation/doubt

Hm. I think it's at least always setting up a contrast of some kind, even if there's no explicit negation. "Gold is valuable per se, regardless of its form" - implying a contrast against other things where form might matter.

Ehh. The OP said the mac is a year old, so it's Apple Silicon - which auto-enforces FDE. That makes things a lot harder.