Carl Mäsak

> It would be _way_ cool if we could add type checking as basically an _optional_ module on top of 007 core. A good way to have this happen is...

Bracha's idea seems to me more about allowing one language to plug in many different type systems (or none at all). But yes, I would like to make the type...

> `Array` is a generic type, and just writing it like that would mean `Array`. Looking at TypeScript prior art, notably the spec section about [type references](https://github.com/Microsoft/TypeScript/blob/master/doc/spec.md#382-type-references), it seems `Array`...

> - Typed parameters in `if`, `for` and `while` pointy blocks are checked as early as possible against their expression. If something is obviously impossible, that's a failure at compile...

In TypeScript (which much of the described type system is based on) you can't do this: my x = 5; x = "foo"; That is, you can't suddenly change the...

I mean, I think I can live with that. It just means in practice that, if you _only_ add a pragma `use typechecker;` at the top of your program (thinking...

@vendethiel As an answer to both of your comments, I don't think I'm being unfair in this case. There are ways to code JavaScript that are perfectly justifiable, but that...

> My guess is that TypeScript went the stricter way because (a) it's a lot easier to write a typechecker if it can make the assumption that the type inferred...

Missed a biggie. * **Recursion**. Not OK in a normal NFA. In the grammar we have mutual recursion via the path `statementlist` → `statement:block` → `block` → `statementlist`. This means...

001 requires a semicolon after a non-last blockish statement. ``` my x = 5; if x { x = 0; }; x; ``` As the grammar is specified, that `;`...