Conversation

Notices

1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 20:05:36 EDT Internet Turtle Ⓐ
   !scheme I have finally decided to begin writing a SRFI for optional type annotations. I am writing a type checker and it's going well.
   1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 20:06:46 EDT Internet Turtle Ⓐ

      in reply to
      I am thinking the best way is to introduce the new forms typed-define, typed-lambda, and typed-let. Let me know if you have better ideas.
      1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 20:07:38 EDT Internet Turtle Ⓐ

         in reply to
         Alternatively I could base it directly off of typed racket. That would also extend the syntax while keeping compatibility.
         1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 20:08:29 EDT Internet Turtle Ⓐ

            in reply to
            Note that these types are only for static analysis, and will be completely ignored during runtime.
            1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 21:47:07 EDT Internet Turtle Ⓐ

               in reply to
               Now that I am digging into it, I am not sure whether I am aiming for "Typed Racket" or "mypy" level of complexity... I am thinking mypy
               1. Internet Turtle Ⓐ (taknamay@quitter.se)'s status on Monday, 30-Oct-2017 21:47:47 EDT Internet Turtle Ⓐ

                  in reply to
                  It will result in a rather unexpressive type system but the implementation will be simpler that way.
                  1. Koz Ross (koz@quitter.se)'s status on Monday, 30-Oct-2017 22:39:47 EDT Koz Ross

                     in reply to
                     @taknamay You can always build up - additionally, SRFIs go through a lot of feedback and changes, so simpler is better at this stage.