An Atlas of Hypertext: Gaps in the Maps

This paper reports on an the initial stages of compiling a comprehensive, historically deep "atlas" of the structures of interactive stories, with initial surveys in branching narrative genres including gamebooks, hypertext fictions, visual novels, and Twine games. In particular, it considers the "gap" between approaches to two highly related yet radically different archives of branching works: an archive of over 2500 interactive print gamebooks stretching from the 1920s to the present, and contemporary collections of the approximately 1500-2000 extant Twine games available in popular public repositories such as the Interactive Fiction Database (IFDB) and itch.io. What do we find when we consider these forms of electronic literature (and their crucial precurors) as one comprehensive atlas of a vast transmedia territory of interactive storytelling? Which methods may be adapted between print and digital works, and which demand new approaches?

In summer 2017 the Transverse Reading Project began surveying an archive of over ~2500 interactive gamebooks in the Katz Collection at UC Santa Barbara -- and began building a collection of visualized interactive plot structures that shape a reader's choices. Mapping interactive stories is a tradition in gamebook culture, with examples of mapping by authors and readers dating back to at least the 1930s. Writers created hand-drawn maps as an aid to writing -- and then readers re-created their own maps as an aid to tracking the explored and unexplored options of interactive reading. In this project, data visualized "story maps" use similar network graphs that simultaneously reflect the branching plot structures of each gamebook or digital game, the way scenes are ordered in the pages of the codex, and the order of individual choices on each page. In addition, the patterns of an "interactive periodic table of elements" are extracted for each work.

In print, data was collected by student researchers using a custom format for rapidly encoding gamebooks (~30 minutes on average), and data-mined / visualized using the open source and cross-platform software tool Edger -- which was custom written for this project. These techniques of data collection, visualization, and exploration will are of particular interest to scholars in related popular interactive genres such as visual novels, Twine games, or life sims. In the second case study, on Twine, automatic harvesting and mapping methods where used to extract network patterns from a large subset of publicly available Twine works -- with results bringing in to focus both the deep similarities and the surprising differences in interactive works from the 1930s to today.

(source: ELO 2018 website)