Add Wordlist Scoring to GNOME Crosswords Editor

Mentors

Project length

Long ~350 hours

Could be short (~175 hours) if necessary, with fewer milestones reached

Description

GNOME Crosswords Editor uses some lists of words in order to produce crosswords. These lists lack the metadata to make more interesting puzzles. We'd like to find, gather, and encode that metadata with the lists in order to choose appropriate words when working with the editor.

For more information on the problem, please read this design doc.

This project will involve finding ways to calculate, measure, and encode values for each of the five traits listed. We will then use it when creating a puzzle to try and create more interesting grids.

Requirements

The biggest requirement for this project is a love of words, and a certain amount of comfort with uncertainty. I will expect the intern to do some independent research and exploration
This project will mostly be in python, and will involve a fair amount of data analysis as well as coding
Some low-level C programming knowledge is a strong plus, as well as the ability to use a hex-editor. The current wordlist is stored in a custom data structure
This project will involve working with some large data sets — possibly as big as 20 GB. A relatively fast machine with sufficient disk space will be required, as well as the ability to download large files.

Communication

Primarily matrix. The main crosswords channel can be found at https://matrix.to/#/#crosswords:gnome.org

We also will use gitlab issues and email as appropriate.

Video conferencing occasionally when necessary.