I’m working on an interactive fiction engine and accompanying game set in the Rooftop Cop universe. Old text adventures, like Adventure, Zork and Planetfall, and new works of interactive fiction (IF), like Lost Pig and Galatea, fascinate me. I want to teach myself to program the software systems these games require and explore what innovations I can bring to the playing field. Text-based games generate mysterious narratives that drive the player to see with the imagination yet they often require masterful wit to divine the solutions to designer-made puzzles. I want to experiment with new mechanics for interactive fiction, new interaction paradigms, that allow the player to explore a continuous space over short moments. My aim is to reduce some of the frustration new players experience by opening up the game mechanics beyond puzzles.

This summer, Jonathan and I brainstormed some settings that would fit within the Rooftop Cop universe and would provide fertile ground for a continuous-space interactive fiction game without being over scope. We thought of five small settings: a hotel room, a bank, a public restroom, a prison cell and an elevator. I then sketched a few interactions within the elevator setting to create a short Twine prototype titled Elevator, available here. However, the final game will not use Twine since the tool produces hypertexts, a form strictly different from interactive fiction.

I am exploring which tools to use to develop the final game. I have three possibilities: reuse an existing tool such as Emily Short’s Versu (in closed beta) that already supports the continuous style of gameplay I would like in the game, modify an existing tool such as Graham Nelson’s Inform 7 or Nick Montfort’s Curveship that doesn’t yet support a continuous style or implement my own from scratch. As a programmer, I lean towards option three. I am reading Nick Montfort’s papers about interactive fiction, examining the source code to software that runs IF, and reading research about natural language processing and natural language generation techniques. We’ll find out which option I choose.