Watir project consists of several smaller projects. The most important ones are watir-classic, watir-webdriver and watirspec. Watir-classic Watir-classic makes use of the fact that Ruby has built in Object Linking and Embedding (OLE) capabilities. As such it is possible to drive Internet Explorer programmatically.[6] Watir-classic operates differently than HTTP based test tools, which operate by simulating a browser. Instead Watir-classic directly drives the browser through the OLE protocol, which is implemented over the Component Object Model (COM) architecture. The COM permits interprocess communication (such as between Ruby and Internet Explorer) and dynamic object creation and manipulation (which is what the Ruby program does to the Internet Explorer). Microsoft calls this OLE automation, and calls the manipulating program an automation controller. Technically, the Internet Explorer process is the server and serves the automation objects, exposing their methods; while the Ruby program then becomes the client which manipulates the automation objects. Watir-webdriver Watir-webdriver is a modern version of the Watir API based on Selenium. Selenium 2.0 (selenium-webdriver) aims to be the reference implementation of the WebDriver specification. In Ruby, Jari Bakken has implemented the Watir API as a wrapper around the Selenium 2.0 API. Not only is Watir-webdriver derived from Selenium 2.0, it is also built from the HTML specification, so Watir-webdriver should always be compatible with existing W3C specifications. Watirspec Watirspec is executable specification of the Watir API, like RubySpec is for Ruby.