Jason

It would also be helpful to document how to reference Lisp variables within a coalton clause. AFAIK it isn't documented anywhere that this doesn't work: ```Common Lisp (in-package :cl-user) (let...

> Also, in response to your intro, I think you can pretty easily avoid monads and `do` notation, but I don't think you should shy away from type classes, as...

This looks really promising to me. Here is some code that would integrate the IO monad. It also defines a `char-stream-pure` package that uses a macro to wrap all of...

> > > We definitely need a monadic version of open? to work with > > This is not obvious to me, as in my mind `open?` is a pure...

> > one that defines the stream types and typeclasses, one with unbound read/write/open/close/etc functions, and one with the IO-bound functions > > The defs file seems superfluous. If we're...

> > Yes, this seems good, except I'd like to keep the concrete stream types not user visible, so I'd call it %FileInputStream. > > I don't think this is...

I'm seeing similar behavior in VSCode using the Alive plugin for Common Lisp. I'm getting an error underline under `coalton-toplevel` on the most recent version of Coalton.  

Aah, that's true. I checked out the revision you listed as giving the correct behavior, and in VSCode it still looks exactly like I posted above. Nevermind!

> Very cool! Thank you! The PR should be ready for review now.

> Overall, this looks good to me! - I looked at everything except the tests. My comments are just about style. > > There are a few instances of functions...