One thing that jumped out while working on the JavaScript chapter of the Web Almanac was the incredibly high amount of time spent processing JavaScript on the median mobile page where Ember.js was detected. (The story was the same when I wrote The Cost of JS Frameworks.)
