Creating Languages in Racket: Sometimes you just have to make a better mousetrap

queue.acm.org

74 points·CowboyRobot··8 comments

Great article! I have been using Lisp languages for much of my work since the early 1980s, and this article is inspiring because instead of just building a "language" up towards the application domain by adding functions and a few macros, Racket supports more general creation of new languages.

I just read the article quickly because today is a busy day, but I'm going to play with this material more this weekend.

This game example... I think Racket would have been a much better starting environment than Common Lisp for the book Land of Lisp. You can download the IDE and get started within a few clicks.

You're not the only one to think exactly this. Just wait for Realm of Racket[1,2], as the rest of us :-).

[1] http://www.realmofracket.com/ [2] http://twitter.com/realmofracket

Arc is currently living within Racket/MrScheme.

Has it moved to Racket, then? As I recall (but I could easily be misremembering) Arc required some old version of PLT Scheme that still came with old fashioned mutable pairs. (And I think they switched to immutable pairs before they switched to the name Racket.)

Last I heard it was still on an old version. Arc really should update. For example, Racket has just changed it's IO system to support epoll/kqueue out of the box. I expect Hacker News would greatly benefit from this.

Arc can run on the latest version of Racket. You have to use Arc 3.1 - that build is a bit buried on the site: http://arclanguage.org/item?id=10254

I think it would be more valid to say that Racket is doing what Arc tried to. Or Clojure, take your pick. Awesome, right?

Reply on news.ycombinator.com