Hacker News Reader

Comment by sddfd | Hacker News Reader

What do you think about Standard ML as a language for teaching?

I like it because the syntax is clean, it is functional and higher-order, but also has an explicit concept of references (i.e. mutable memory) and using functions with side effects (i.e. printing) is no trouble at all.

View on HN

I have been programming for over 20 years, most of it in FP land. ML is tractable, a beautiful prism through which one can peek at Computer Science as a subject, but it has a frustrating lack of _practical_ tooling and beginner-friendly community. For anything outside basic I/O, you will need to dust off compiler-specific extensions, most of them last used on "Un*x Workstations".

Just look at the Basis library. Safe to say it's nowhere near Racket, Scala, Clojure, Kotlin, ...

http://sml-family.org/Basis/manpages.html

Using ML trough ReasonML + BuckleScript and the related tooling should make ML palatable.

Alternatively, F#. That fixes the tooling issue pronto.

F# or OCaml then.

I never understood why a lack of tooling would be a serious problem for a language used for teaching FP.

It's a tough sell if the language has no hope of industry adoption. Also tough to use a language, especially one with type inference, without editor support.

Most people in my SML class seemed to have no problem with either of those.

Maybe some classes are just that good. A lot aren't, I'm guessing, and you would want to at least take away something you could use in the real world from them.

While I was already familiar with Common Lisp, SML gave me an appreciation for functional programming, as I worked through a set of lecture notes and examples that I found on a CMU website (apparently no longer available.) I am in no position to say that it is the best choice, but it worked for me (and the quality of the notes may have been an important factor.)

SML was my first exposure to FP, and while I didn't _really_ get it then, it was responsible for my initial interest in Haskell after graduation. I think SML is a great tool for teaching FP.

I second that. I followed an 8 week university course teaching introductory programming using SML and picked up Haskell on my own afterwards. I think SML helped me get enthusiastic enough about FP to motivate me to study the more advanced concepts

Standard ML would be a great beginner's language except for the weak beginner's guides and lack of support. Even basic printing to console can be agonizing in Standard ML.

Not really:

    $ poly
    Poly/ML 5.7 Release
    > print "hello\n";
    hello
    val it = (): unit

What about Ullman's wonderful book?

Reply on news.ycombinator.com