/analogy_farm

A Web-based puzzle from MIT Mystery Hunt 2013.

Primary LanguagePython

Analogy Farm was a popular puzzle from the 2013 MIT Mystery Hunt. The puzzle is a web application; your goal is to interact with the web application until the answer becomes clear. The answer is a single word.

solution.txt and solved.svg contain spoilers, if you couldn't tell from the filenames.

Interesting features of the code

  • The app maintains state in a SQLite database, and updates it live. If you and someone else are solving the same instance of the puzzle, and they fill in a box, you see the box update in your browser.
  • The boxes you fill in on the page are auto-generated by finding the rectangles in solved.svg, in the script parse_svg.py.
  • It actually survived the load of the Mystery Hunt.

Interesting misfeatures of the code

  • The parse_svg script generates .json, and you need to add an assignment to it to make it a valid .js file. Oops.
  • Static file paths are hardcoded.

To do

  • Get it hosted again, probably in a gunicorn + WSGI setup (There's a server available for this purpose, if someone figures out the configuration)
  • Make it so that a single copy of the software can host multiple "teams" of progress
  • Add a reasonable way to reset progress