As mentioned in the previous section, markup languages are used to add meta information to data. Data can be anything, it could be the settings of a computer program, it could be entries in a dictionary or a library catalogue, it could be prose texts or poetry. Markup languages provide a way to describe this data. Markup languages should be distinguished from programming languages. Markup does not, in itself, process or style text. It must be used in conjunction with either a programming or scripting language, such as XSLT or CSS.

Take HTML for example. HTML is the language in which most of the pages on the internet are delivered in. HTML tags (such as <p>...</p> or <h1>...</h1> are used to structure the text (paragraphs, titles, footnotes, etc), to add in other media, such as images, audio, or video files. Browsers can interpret these HTML tags to process the text so that it not only appears on your monitor, but links via hyperlinks to other pages, displays images in the appropriate place and the appropriate size, etc.

HTML and LaTeX are examples of presentational or procedural markup languages. They are used primarily to indicate how text and data is supposed to be processed by a software. For instance, HTML developers are primarily concerned with how something looks in the web browser. However, in recent years with the development of HTML5, HTML has becomes more similar to descriptive or semantic markup. This kind of markup language describes the type of data that is encoded, e.g. an address, a pizza menu, or a stanza of a poem. The benefit of this kind of markup is that it can be better understood by software such as search engines. A typical example of a descriptive markup language is the Text Encoding Initiative (TEI). Markup is typically not visible to the reader of electronic text. Rather, it serves as instructions for software to process text.