Jeff Morrison

> I've always just imaged that anything more complicated (including dereferencing this properties) should just go in the constructor, where all of the other complex initialization logic resides. My only...

cc @allenwb who was in favor of initializer bodies having the semantic intuition of a method body

@ljharb: It's worth mentioning that your example does behave the same. It's only the [relatively rare] case where the user uses `arguments` (and not the less-deprecated rest-arg) that it would...

@ljharb: If we were to go back to early error, what would you say these expressions should do within direct eval?

Unfortunately detecting direct eval statically is undecidable, so there's not much we can do there. We could make it a runtime error only for direct eval, but we do have...

I guess it's also worth pointing out that the current spec text (i.e. not updated since the latest committee discussion) bans `arguments` in a not-so-precise way -- which is that...

When the issue of define vs set first came up we explored it quite a bit. The conclusion we ultimately came to was that neither is 100% obviously better than...

> If you want const foo = 42 in the future, why the syntax is not let foo = 42 now? The reason `let ` was dismissed as a prefixing...

Yes I agree, I think a future const fields proposal should probably dismiss the `const foo = ` syntax in favor of some other syntax. Let’s try to keep this...

> Are we JS or are we XML? I am human. I am dancer.