Nick Smith

(Below is just a brain dump. It's by no means a final verdict.) I'm starting to suspect that storing types in runtime variables, e.g. `var Type: AnyType`, doesn't really make...

As part of this, please consider disabling the `__` ligature. (It certainly isn't obvious, because it obscures how many underscores are used.) And if you're going to offer a minimalist...

Also, it might be worth considering a "gapped arrow" ligature, rather than no ligature at all. Something like:  This makes it clear that...

It's also worth considering the ligature style that the 0xProto font [uses](https://github.com/0xType/0xProto?tab=readme-ov-file#4-ligatures-that-dont-deform-shape).

Oh, I see. Clearly I haven't noticed that subtle distinction. Consider my confusion evidence that the term "generic" needs to be explicitly defined, then 😉.

Just FYI, Mojo is [likely going to offer](https://github.com/modularml/mojo/issues/1458) _explicit_ return value optimization, which will be useful in cases where you really, really want to make sure that a return value...

> This is an issue that has proven VERY divisive. So far, everyone who I've seen comment on this topic seems to agree that the casing rules should be uniform...

I think what you mean is that some FP languages have a notion of "type constructor". Applying the terminology to Mojo, `SIMD` would be a type constructor, and `SIMD[DType.float32, 4]`...

@hardskulls as Richard points out on that thread, Roc's tags are basically the same as OCaml's polymorphic variants. So they're not necessarily anything novel. That said, polymorphic variants are a...

> Though it does unfortunately clash That's true: the specific means of introducing the colorization in VSCode would require careful consideration. Though, not everyone uses the (non-semantic) bracket colorization feature...