Descriptive. Under the descriptive system of markup, authors identify the element types of text tokens. In Figure 1 the tag identifies the following text as a long quotation, and the tag identifies the end of the quotation.

Authors who are accustomed to procedural markup often think of descriptive markup as if it were procedural and may even use tags procedurally. The primary difference is that procedural markup indicates what a particular text formatter should do; descriptive markup indicates what a text element is or, in different terms, declares that a portion of a text stream is a member of a particular class. When a text formatter generates the presentational copy of a descriptively marked-up document, it first reads in a set of rules, written in a procedural markup system, that establish what it should do for each occurrence of each element type. By adjusting this set of rules, the author (or support person) establishes a presentational markup design that will be executed automatically and consistently. Moreover, should there be reason to modify the design, only the rules will require editing: the document files remain intact. Not only will the author be relieved of painful and monotonous hours of mechanical editing, the text will not be exposed to corruption.