Client-side: JavaScript

• interpreted, no compilation
  • interpreter is embedded in another app (browser)
• scripts operate in document/browser
• dynamically typed — untyped variables, typed values
• objects

  • an object is a set of properties

  • no classes!

  • a property is a unique-name key:value pair

  • created with a constructor:

    var ob = new Object();
    ob.testing = "test";
    

  • deleted with delete:

    delete ob.testing;
    

  • built in: window (browser window), document (page in window)

• DOM API (Document Object Model)
  • by W3C recommendations
  • works with document tree
  • functions from JS are called through HTML attributes
  • can add noes, remove nodes, change attributes
• Events
  • when should shit happen?
  • after page loads, while(forever) wait for event
  • predefined: click, mouseover, mouseon, mouseout, keypress, resize, scroll, etc.
  • write a function to respond to event, then subscribe/register/bind it
  • browser calls it when needed
  • event “bubbling”: events execute from lowest node upwards
• AJAX (more like AJAJ or whatever, not really XML anymore, but AJAX sounds cooler)

  • recipe (always the same)

    1. Make initial page

    2. Set up event handlers

    3. When called, handlers do magic (request new data from server over HTTP in XML/JSON/whatever)

       • asynchronous, execution continues all the time, there’s no wait for the server
       • shitty using XMLHttpRequest, better with jQuery

  • fast, easier for users

  • weak browser integration (e.g. back button), low search engine indexing, accessibility, response time, security

  • same origin policy: requests to other sites will be blocked, unless the target site opts in explicitly

  • some AJAX calls return HTML, others JSON (originally XML)

• libraries — jQuery