cmlib VS ponyo

I am not the first one to come up with this idea, but I rediscovered it independently. Also, cmlib's implementation uses partial functions, whereas mine uses only total ones.

Check out cmlib (https://github.com/standardml/cmlib), or if you're using SML/NJ, ORD_MAP implementations (https://www.smlnj.org/doc/smlnj-lib/Util/sig-ORD_MAP.html).

Here's a hash table implementation in SML for example: https://github.com/eatonphil/ponyo/blob/master/src/Container/Dict.sml.

mldoc (from Ponyo)

Chris Cannam's blog has some nice writing on the differences in getting compilers working and some general formatting thoughts [1,2]. Alex Alegre's A Tour of Standard ML [3] is pretty nice, too, and I think it starts to get through your 3rd bullet. Jake Zimmerman's blog has some nice setup information for developing Standard ML with Vim [4]. Naturally there's your writing at ponyo.org, too.

I'll throw out my neglected standard library project, Ponyo, as one that is not advanced at least as the Standard ML compilers or other languages implemented in Standard ML.

They are basically handwritten parsers. Take a look at this parser in Standard ML for parsing Standard ML. The entire parser is using combinators to parse. It doesn't use a third-party library like parsec but it's basically the same thing as far as I understand.

Example of use [here](https://github.com/eatonphil/ponyo/blob/master/src/Net/Http/Response.sml#L8)