Initial JS for CS showcase can be slightly shorter http://ryanflorence.com/2012/javascript-coffeescript-rewrite/

cs-showcase.js

// No function wrapper as it's CommonJS/NodeJS module

var slice = [].slice

this.events = {

  events: {},

  on: function (topic, handler, context) {
    (this.events[topic] || (this.events[topic] = []))
      .push({ handler: handler, context: context || this })
  },

  trigger: function (topic) {
    if (this.events[topic] == null) return
    var args = slice.call(arguments, 1)
    for (var event, i = 0; event = this.events[topic][i]; ++i)
      event.handler.apply(event.context, args)
  }
}

@rpflorence

1. Not really, JavaScript is not only in browsers, but also server-side. This is the way you write code for Node.js, and you can write same way for browsers with help of tools like webmake - That's how I work for over a year right now ...and other thing, in module: this === exports, so it indeed exports what you expect ;)
2. I just shown, that in many cases you may write similar constructs as in CoffeeScript. If you don't like it because it's doing too much at once you should in first place refrain from using CoffeScript, as it's how it works just one magnitude larger.
3. Experienced JavaScript programmers know and use that construct.There's nothing unusual about that.
4. You are exaggerating, and I guess you know that. Many great JavaScript projects don't use semicolons, and to the point - it's definitely not more dangerous than using CoffeeScript.

I just wanted to show that this example was a little biased, that's it. Whether it's good to use CoffeeScript or not, it's totally different story. I'm actually not into it, because I don't see it as replacement but addition (you can't write CoffeeScript without knowing JavaScript), so no matter how simpler code looks like, it's an additional baggage for mind. However if some of the CoffeeScript stuff were to be proposed for EcmaScript 6, I guess I would be very happy with that.