Reasons CoffeeScript is still relevant for me

I'm often asked why I still use CoffeeScript rather than ES6/7. Well, in truth I use both, but I do still prefer CoffeeScript for the following reasons:

Correctly trimmed multiline strings (ES6 includes the indentation because it doesn't have significant whitespace)
Existential operator (a?.b; b?(); a?.b?() and of course a ? b)
- a ? b is more appropriate than a || b if "" or 0 are values of a you wish to persist
- (a && a.b && typeof a.b === 'function') ? a.b() : undefined doesn't quite have the same ring to it as a?.b?()
Significant whitespace
- (a personal preference, but since you already lay your code out sensibly why do you need the additional visual distraction of braces?)
Convenient shorthand for common things (-> (regular non-bound function), @ (this.), :: (.prototype.))
in to check if something in a list
- a in b is a lot clearer and harder to mess up than b.indexOf(a) >= 0
Chained comparisons a < b < c (minor)
Ranges [a..b] e.g. in loops for a in [1..10]
Shorthand slicing a[b..c] / a[..] (minor)
Block regexps (a very good way to allow documenting of complex regexps)
Safe switch statements (never forget a break again; minor since this can be linted against)

For more you might be interested in my Long Live CoffeeScript and Long Live ES6 post.

There's a lot to like about ES6/7 though; here's some things that I miss from Babel when writing CoffeeScript:

let / const; and to be honest var (CoffeeScript should support an initial assignment operator such as a := b)
The ternary operator a ? b : c (CoffeeScript doesn't have this because of the existential operator and implicit object syntax; however I think that could be solved by switching to a ?? b and giving the ternary operator higher precedence than implicit objects?)
spreading objects {...a, b: 'c'} (CoffeeScript has spread for arrays, but not for objects; easy to work around with Object.assign({}, a, b: 'c') but I like sugar)

benjie/CoffeeScriptReasons.md