The code in this repository models a real escape room in Los Angeles called Perfumer. It contains spoilers about the puzzles in that room.

An API for designing puzzles, with a focus on escape rooms as a type of puzzle.

Python 3 is required for core API. An installation of graphviz is required to generate figures of puzzle DAGs. I use a conda env for dev (but that isn't a dependency yet).

• Install Graphviz via Homebrew: brew install graphviz
• git checkout and cd into the repo
• run $ python main.py at your command line

The idea is to codify some of the real esape rooms we've done.

In escape rooms, puzzles in various forms (e.g. find a hidden object, solve a riddle, decode a cipher) are used as keys to locked spaces, which in turn often contain a key (or part of a composite key) for other locked spaces. Sometimes locked spaces contain other smaller locked spaces (for some definition of smaller). It seems we should be able to capture these dependencies into an architecture or data structure -- e.g. a directed graph -- and reason about their properties.

The goal of this architectural API is to help us reason about what makes a puzzle a puzzle, and how various associated concepts (e.g., locks, keys, riddles, puzzles, solutions, ciphers, etc.) all relate to each other.

Can we ultimately create a data structure that captures sophisticated layered/nested DAGs of puzzles. Can we automate (or at least facilitate and ease) the design and creation of new puzzles and escape rooms that are enjoyable and challenging but balanced?

This includes understanding and capturing the design constraints of puzzles and escape rooms, including the puzzle difficulty and balance, potential parallelism for teams working together, hint systems, and how to associate story-telling components (characters, back-story, plot, etc.) with puzzles.

A longer term goal is to extend the tools beyond reasoning about systems of puzzles that need to be solved to problems which needs to be solved and tasks that need to be completed.