•  

Sessions at GothamJS about Web Performance

Your current filters are…

Saturday 9th July 2011

  • The Once And Future Script Loader

    by Kyle

    Like the tales of King Arthur, the history, mythology, and lore of script loaders is teeming with both fact and fiction. If we're going to achieve the destiny of true web performance optimization in our page loads, we're going to have to separate the myths from the truths.

    We're going to take a brief look at the history of script loading techniques, looking at what survived and what didn't, and why. We'll also look at some of the techniques that are still around but shouldn't be!

    Then we're going to jump into a broad review of current script loading techniques. We'll cover from the very basic single-script snippets (like the "async snippet" from Google Analytics), to general script loaders like LABjs and HeadJS (and several others). We'll also look at task-specific script loaders such as those found in frameworks (Dojo, YUI, etc). Our goal will be to identify the characteristics of each loader and loading technique, to give you solid information you can use when choosing the right script loader for your sites.

    We'll also survey the current state of browser technology and the HTML specification standards for these loading techniques, and also talk briefly about where those things are headed in the future.

    Finally, we'll gaze into the crystal ball to look at the future of advanced use-cases for script loading, such as module dependency (RequireJS) and deferred script execution (ControlJS), as well as upcoming ES-Harmony Modules. We'll talk about why these things are so valuable, and give you practical do's and don'ts for implementing these techniques right now.

    More than just a technical talk (although there's plenty of code!), this is a talk for advocacy and awareness among the developer community on the topic of efficient and performant script loading, and what members of the community can do to become involved in shaping the next generation of script loading technology.

    The rich history, diverse current state, and bright future for script loaders is both technically challenging and highly rewarding in performance gains, if you accept the quest.

    At 10:00am to 10:40am, Saturday 9th July

  • The Fallacy of Microlibs

    by Yehuda Katz

    These days, the idea that large JavaScript applications should be built on top of tiny, independently developed libraries is all the rage. The argument goes that with the latency requirements of mobile devices, we can't afford to ship code that we're not using. Like Unix, the argument goes, small libraries designed to do one thing and one thing well are the way to go.

    There's another way. In this talk, Yehuda will provide an alternative approach to thinking about developing large JavaScript applications, the approach the SproutCore team is using for developing SproutCore 2.0. He will explain why 5k vs. 1MB is a false dilemma, and why framing the discussion that way is crowding out good solutions in the JavaScript community's marketplace of ideas.

    We're still in the early days of building rich applications in JavaScript. Now's the time to learn from those who came before us, not the time to prematurely declare that small incremental changes to the existing browser-based programming models are all we need.

    At 4:50pm to 5:30pm, Saturday 9th July