Project title:
Gamifying constraint violation fixes on Wikidata
Description of project:
Develop a game to make edits to Wikidata to fix constraint violations. The requirements for being a game here is not only that the tool is easy to use but also that there are methods that keep users engaged, such as scores, leaderboards, collaborations, and challenges. There may be aspects of community collaboration in some games.
There is a long-open Phabricator ticket to make more Wikidata games
https://phabricator.wikimedia.org/T165167
Constraints in Wikidata:
Wikidata has several kinds of constraints. One kind is the property constraint, a simple soft or hard rule on how a property should be used, such as the values of subclass of (property 279) have to be Wikibase items, i.e., not data values like integers, or father (property 22) should have only one "best" value. For more information on Wikidata property constraints see https://www.wikidata.org/wiki/Help:Property_constraints_portal. Other kinds of constraints in Wikidata come from the intended meaning of some of the items in Wikidata. For example, Wikidata has disjoint unions, so an individual that is an instance of two disjoint classes in Wikidata is a violation of the intended meaning of disjoint union.
Violations of constraints are thus places that show evidence of a potential breakdown in the way data should be modeled in Wikidata, and often are the result of incorrect information being put into Wikidata. Reducing the number of constraint violations thus improves the quality of Wikidata.
Because there are so many constraint violations in Wikidata, it is hard for small groups of editors to tackle the problem. Enlisting the efforts of more people via a game has the possibility of making significant reductions in the number of constraint violations.
Expected outcomes:
By the end of this project, the contributor will have developed a game, either based on currently existing software like the Distributed Game (https://wikidata-game.toolforge.org/distributed/) or inspired by it. The game will be designed to fix particular kinds of constraint violations in Wikidata. It will be possible to slightly modify the game to fix other issues that exist in Wikidata. The game may be used later on in outreach campaigns.
Bonus outcomes:
A great tool for introducing people to Wikidata.
A tool to help communities not currently engaged with Wikidata involved in editing Wikidata.
Skills required:
Understanding of Wikidata
Game Design
SPARQL proficiency
Proficiency in Python or another scripting language
Skills preferred:
Database management
Conducting reliability tests on user inputs
Ranking users according to skill
Bonus skills:
Community management
Graphic design
Possible mentor(s):
Peter F. Patel Schneider
David Martin
Size of project:
350 hours to complete
Rating of difficulty for the project:
Medium - The project requires the contributor to understand existing gamifying attempts at Wikidata, and either emulate them or build upon them. It also requires many different skills, such as database management, game design, reliability analysis for proposed changes,
Microtasks:
- Understanding what the Distributed Game does, as evidenced by a short report on how it works, is a great introduction to what games can do for Wikidata.
- Updating an existing game for the Distributed Game is an excellent task to get up to speed in game design, for example the New Wikiquote article and category matches (https://wikidata-game.toolforge.org/distributed/#game=90) could be updated to remove distracting bracketing in text.
- Making a minor update to the Distributed Game is a good task to understand how games can work internally.
Why are you proposing this project?
There are so many violations of Wikidata constraints (including both from Wikidata property constraints and constraints inferrable from the intended meaning of core Wikidata constructs like disjointness) that it is difficult for small groups of editors to keep up, or indeed even survey the extent of the problem for some kinds of violations. Games are one mechanism for encouraging large numbers of participants to fix problems like constraint violations. A game to fix constraint violations would help to reduce the number of constraint violations in Wikidata.
What is the expected impact?
The immediate impact would be the creation of a new game addressed as fixing problems in Wikidata. The desired eventual impact is a reduction in the number of constraint violations in Wikidata of the kind addressed by the game. Another desired eventual impact is the continued creation of similar games, which would expand the constraint violations being addressed.
Any other additional information that the interns should know about:
Closed Phabricator ticket that could be used as an example of an update to the Distributed Game https://phabricator.wikimedia.org/T252935
Open Phabricator tickets related to the Distributed Game
https://phabricator.wikimedia.org/T258067
https://phabricator.wikimedia.org/T253956
https://phabricator.wikimedia.org/T253839
https://phabricator.wikimedia.org/T210635