Hacker News Reader

Comment by BoysenberryPi | Hacker News Reader

I used SML as part of the Programming Languages course taught by University of Washington. It's easily one of my favorite languages and I wish there was more library support for it.

View on HN

I feel the same way. Maybe we can be the change in the world we want to see. There's also OCaml which seems to have more libraries, a single compiler (yay). The downside is OCaml has a ton of features whereas I find SML to be small enough to hold in my head.

On the other hand, you don't need to know all of OCaml to be effective and unknown features are not going to come and bite you while you don't use them.

Most people don't know the object system. Most advanced features are rarely used and kept well hidden in libraries internals (first class modules, extensible types, ...)

You will need to understand functors for some libraries, though, but I would guess that it's true for SML as well. :)

The UoW course did emphasize the Module system quite a bit.

Random question, how hard would it be to port OCaml libraries to [s]ml ?

Frankly, this question doesn't make sense. OCaml's type system is much richer than SML's.

The opposite is pretty trivial, though.

I got my answer somehow :)

While that's somewhat true, OCaml is still a somewhat "small" language when compared to lots of other functional languages (Haskell, Scala, and even F# to an extent). At the very least, it has an agreed upon subset of the language to use in everyday code (which most languages of similar size or larger lack), which makes almost any code you find on GitHub or wherever perfectly readable. On the other hand, the standard libraries (or lack thereof, considering there's no wholly agreed upon standard library to use) can feel at times vast and complex since every standard library alternative has its own subtleties that can make it difficult to go back and forth between all these alternatives.

IIUC they didn't have re-runs, a bit sad about that too.

Reply on news.ycombinator.com