Integrating content and style in documents: a case study of patient information leaflets

Perhaps the most powerful idea in recent document processing is that of specifying the style of a document separately from its content. This idea is implemented to some extent in modern word processors, which allow you to vary the appearance of a document systematically by loading a new style sheet; it is implemented more fully in systems based on the Document Style Semantics and Specification Language (), which allow detailed control over the presentation of a text that has been marked up in  (Standard Generalized Markup Language). Why is this idea so powerful? Suppose that a pharmaceutical company has designed patient information leaflets for all its products (these are the slips of paper inside medicine packs which explain how to take the medicine and warn you of possible side-effects). Drafts of the leaflets are printed, and passed on to the market research department for testing. One of the things that could result from a test is that patients have difficulty navigating the leaflets because paragraphs are too close together: instead of placing a new paragraph on the next line, starting with a tab, it would be better to leave a double line with no tab. Moreover, it could be found that patients dislike the indented square bullets in unordered lists, finding them ugly and that they demand round bullets aligned with the text. Implementing these changes is obviously far easier if the style for realizing paragraphs or unordered lists has been specified just once, in a general rule. With no separation between content and style, the author will have to go through every paragraph or list in every leaflet, correcting each case individually. In existing word processors, control over style is limited by the categories available for describing the content of the text. You can