After dark Book

I wish I learned Standard ML later. I also wish there was a newer edition of this book.

It's not a bad book to learn Standard ML from. You could understand most of it without having to run smlnj, and for the most part, it was a pleasant read. However:

* It's an old book and it's typeset pretty horribly. I guess it's OK for 1997, but it's annoying to read it it 22 years later.
* It tries a bit too hard to bridge in C/C++ developers, instead of just jumping in the functional waters head first.
* Occasionally there are some really lengthy explanations for some really obvious code that just get in the way.
* It ends up with a definition of the syntax of ML that's just a bit awkward to read. You may infer the language from it if you think hardly enough, but examples would have been better.

Apart from all that, it's a pretty decent book. One thing that I particularly enjoyed was the exercises - occasionally there was something cool to learn in them, like 2-3 trees.