Skip to content

deathandmayhem/jolly-roger

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Jolly Roger

Jolly Roger

Jolly Roger is a coordination tool for collaboratively solving puzzlehunts like the MIT Mystery Hunt. The team Death and Mayhem created Jolly Roger in 2015 for the 2016 MIT Mystery Hunt, and it is still actively maintained and in production service.

At its core, it allows tracking puzzles and guesses in an ever-evolving hunt structure, and provides a chat room and a Google spreadsheet for each puzzle.

Want to help develop Jolly Roger? Join our Discord server

Join our Discord server

Features

Automated spreadsheet creation/sharing

Like many Hunt collaboration tools, Jolly Roger automatically creates a Google Sheet for each puzzle entered into a hunt. If a user has linked their Google account to Jolly Roger, Jolly Roger will additionally share the sheet with that user on first load, which will cause the user's cursor to show their name, rather than "Anonymous Animal".

Users who do not link a Google account (or don't have or use one) can still participate in the sheet anonymously.

Spreadsheet creation

Per-puzzle chat rooms

Jolly Roger provides its own persistent chat with each puzzle. No need to constantly jump between the spreadsheet and a third-party chat service. No losing relevant discussion to 10000-message retention limits in the middle of the Hunt. Supports basic formatting -- bulleted lists, _italics_, *bold*, `monospace`, and ```code blocks```.

We intentionally keep the chat pane always visible on each puzzle page, so remote hunters desperately trying to get your attention are hard to accidentally ignore.

Chat

Tag-based structure

In many hunts, it's not always clear from the beginning what puzzles will be related or may contribute to which metapuzzles. Jolly Roger solves this by eschewing a top-down hierarchical round structure and instead allowing multiple tags to be applied to each puzzle to construct dynamic groupings as you gain more information about the hunt structure over time.

Groups are automatically detected for any tag starting with group:. The metapuzzle for a group is automatically recognized by having an additional meta-for: tag matching the group tag.

We've found this approach able to be capable of modeling every hunt structure we've thrown at it.

Tag structure

Once a metapuzzle is solved, the puzzles in that group are usually less interesting from a hunt-progress perspective, so we automatically reorder things on the puzzle list page, sending the whole group to the bottom of the list when you solve the meta for the group.

Hover over any tag to see other puzzles that share that tag. This is particularly useful for metas where you want to see the answers that feed into the meta, or for backsolving to see what constraints all answers in that group appear to satisfy.

Tags on hover

Viewer tracking

Sometimes you want to find a puzzle that doesn't have much attention on it right now, or you want to find something where people are already looking at the puzzle. Viewer counts in the puzzle list make it easy to find what you're looking for.

Instant filter search

Find puzzles in the growing list by searching for any piece of the title, answer, or tag.

filter search

Operator guess queue

Avoid irritating Hunt HQ with your team's excessive ill-fated backsolving attempts by staffing an operator queue that can tell people "I'm not calling this in until you convince me you're not just submitting every possible combination of three letters."

Or just make sure your team is ready to receive callbacks before calling things in.

Guess submission

Guess queue

If you don't like the overhead of tracking and requiring operator submission, you can turn it off and let people mark their own answers as correct or not.

Multiple answer support

Some hunts have puzzles with multiple answers, which present a challenge for many systems: how do you track a correct answer while still indicating that there's more to be figured out? If you make a separate puzzle for each answer, you lose the context of the discussion/spreadsheet; if you just comma-separate answers, it's hard to tell which puzzles are finished vs. ones that are only partially completed.

Jolly Roger has first-class support for puzzles with more than one answer, and it Just Works™ with all the smart sorting too.

Announcements

Make sure everyone on the team sees teamwide announcements, whether they've been actively hunting the whole time or are dropping in and out and just need to get up to speed.

Announcements

Realtime updates

You never need to refresh the page. Everything tracked in Jolly Roger updates in realtime. It's delightful.

Setting up a Jolly Roger instance

Interested in developing or testing Jolly Roger locally, or running your own production instance? See DEVELOPMENT.md.