binary search tree

bst

class BinarySearchTree

  constructor: () ->
    @val = null
    @size = 0
    @right = null
    @left = null

  insert: (val) ->
    if (@val == null)
      @val = val
      @size++
      return true
    return false if @val == val

    if val < @val
      @left = new BinarySearchTree unless @left
      @size++ if @left.insert(val)
    else
      @right = new BinarySearchTree unless @right
      @size++ if @right.insert(val)

  contains: (val) ->
    return true if @val == val
    if val < @val
      return false unless @left
      @left.contains(val)
    else
      return false unless @right
      @right.contains(val)

exports.BinarySearchTree = BinarySearchTree

The simplicity achieved here are excellent.
I wish I could comment on individual lines...

One edge case might be avoided if you put the early exit for the equality check right at the start of the insert function. This will prevent a case where a tree can end up with a size of 3, but no children after this code is run:

bst = new BinarySearchTree();
bst.insert(null);
bst.insert(null);
bst.insert(null);

That will also nicely match the code of the contains function.

Actually, with that done, you might as well then go ahead and set up the entire insert method as an if/else if block with 4 conditions (equal, null, lesser, greater). Then you can drop the explicit return true statement when saving the value. Implicit return will return the new size, which will always be truthy (since it must be greater than 0). The whole function will then always return the child tree's size (or false), instead of sometimes the size and sometimes true. See fork.

But that's all just minor nit-picks.

Really really good work boiling this down to it's essence!