In hypermedia systems, the cost of traversing a link is low but not zero. The PREP database/editor [] illustrates how this cost can eliminated. The database stores chunks of textual/graphical data and allows users to group these chunks into drafts. Each draft corresponds to some shared document that the users are collaborating on. The editor allows the chunks in a draft to be displayed in a variable number of variable-sized, named columns. Typically, one column contains the contents of the document and the other columns contain comments input by different users about the columns to their left. Moreover, typically, the contents column is shared by the users but not the comments columns, each of which contains the comments made by a particular user. Thus, reading left to right, one sees how the discussion about a part of the document progressed; and top to bottom one sees the comments input by a particular user about the different parts of the document. PREP can be considered a special case of a hypertext system in that a comment on some chunk can be considered a node linked to the latter. Unlike traditional hypertext systems, PREP requires no user effort to traverse hyperlinks (since the comments are always visible), a feature hypertext users desire []
Several versions of PREP have been built, this discussion applies to the initial version. We shall study features of subsequent versions later, under the issues of diffing and coupling.