No sooner was the Internet upon us than anxiety arose over the ease of accessing pornography and other controversial content. In response, entrepreneurs soon developed filtering products. By the end of the decade, a new industry had emerged to create and market Internet filters.... Yet filters were highly imprecise from the beginning. The sheer size of the Internet meant that identifying potentially offensive content had to be done mechanically, by matching "key" words and phrases; hence, the blocking of Web sites for "Middlesex County," or words such as "magna cum laude". Internet filters are crude and error-prone because they categorize expression without regard to its context, meaning, and value. Yet these sweeping censorship tools are now widely used in companies, homes, schools, and libraries. Internet filters remain a pressing public policy issue to all those concerned about free expression, education, culture, and democracy. This fully revised and updated report surveys tests and studies of Internet filtering products from the mid-1990s through 2006. It provides an essential resource for the ongoing debate.