A RET tag modifies a word or a passage with a feature. Editors use tags mostly to make explicit a feature of the text that we can readily or certainly interpret but that computer software cannot. In practice, such tags enable software to make correlations between strings (words) and their attributes as we understand them to be. A COCOA tag, once inserted, characterizes all text after it with the tag's variable-value pair until another tag with the same variable appears, at which point its value (which could be a null one) replaces the value in the previous tag. For example, all text after <speaker Hamlet> occurs is spoken by Hamlet until another speaker tag occurs, e.g., <speaker Cordelia> or <speaker ->. An SGML tag holds until its second delimiting tag occurs or until another tag occurs that, by definition, closes the first one.

There are four kinds of tags used within RET:

  1. GLOBAL TAGS: these set characteristics of the entire book;

  2. FEATURE TAGS: these identify units in the book as being of a certain kind and often as having some feature or attribute;

  3. STRUCTURAL TAGS: these classify passages of text according to whether they repeat themselves, or re-cycle, within the work, especially nested inside some other unit of text;

  4. WORD-LEVEL TAGS: these mark a string with one of its attributes.