Adrian Sampson

Looks like Zola doesn't like the most missing a name—it should be `YYYY-MM-DD-name.md`. Can you please add that?

Because you are doing this entire project after the deadline, you will need to keep it small. I recommend not attempting to do these extensions: > inheritance, constructors, object-specific methods,...

All sounds good! TBH I think you should probably not bother comparing performance against current existing fancy/limited/experimental compilers… I'm not sure what you would learn from that, really. Comparing against...

Sounds good to me! Curious to see how it goes.

All sounds good (including sticking with the dynamically-typed language). Since you mentioned that built-in operations will rely on your hand-written C/C++ functions, I want to suggest an approach that you...

I think this makes sense! Is the point that IR generation via the OCaml bindings is harder than it looks, whereas there are plenty of resources (Kaleidoscope among them) showing...

Got it. I guess my suggestion then boils down to this: don't bother finding a way to serialize the AST itself as JSON. Instead, serialize an intermediate format that is...

OK, sure! Let's set your new deadline on Monday. Thanks for the progress report. I have thoughts on both of these challenges: > Figuring out how to get LLVM to...

Sounds fairly substantial! You didn't say it explicitly, but I think you will be following the recipe from [Morrisett et al. 1999](https://dl.acm.org/doi/10.1145/319301.319345), right? Or is the CPS route you're imagining...

Yeah, I think your idea to basically invent your own typed language will work fine. Starting with typed Racket and doing a very (very) restricted subset of it will work...