Thomas Pellissier Tanon
| Tags | Analytics, API, Collaboration, JavaScript, Lua, Mobile, Multimedia, Structured Data, User Experience, Wikibase, Wikidata |
|---|---|
| Primary Session | Knowledge as a Service |
| Secondary Sessions |
Title: Content structuration and metadata are keys to fulfil our strategy
Content: The Wikimedia mouvement strategy is making a focus on serving more different kinds of knowledge and sharing them with allies and partners [1]. I believe that the most important ground work for reaching these goals is to focus on the outgoing project of moving MediaWiki from a "wikitext plus media file" collaboration system to a platform allowing people to be able to collaborate on many kind of contents and to organise them in a cohesive way.
Two axes seem important to me to pursue this goal: 1. Support a broader set of different contents, not just wikitext, Wikibase items and Lua/JavaScript/CSS contents but also images, sounds, movies, books, etc., that should bee editable just like wiki text pages in order to allow people to improve them in a collaborative way. 2. Build platforms and tools allowing contributors to create and clean metadata about these contents in order to build together the broadest cohesive set of knowledge ever available and increase its reusability.
Going in these directions would allow us to:
- Allow sister projects (and possible new ones) to use relevant content structure for their projects instead of a designed-for-everything wiki markup. It should lead to an increase of their reusability and their user-friendliness, just like what the Structured Metadata on Commons project is aiming for.
- Build powerful APIs to retrieve and edit content just like Wikidata has and so, make working with partners easier.
- Increase the connections between our contents and their discovery using their metadata
- Build better tasked-focused mobile viewing and edit interfaces
- Be more ready for the possible new environment changes like voice-powered interfaces
Some examples of projects we could work on in order to move in this direction:
- Use the multiple content revision facility to migrate progressively the data that could be structured out of Wikitext on all our projects (like the structured metadata on Commons project is aiming for files)
- Federate all our structured content into a "Wikimedia Query Service" that would allow to do unified and powerful analytics and to ease the discoverability of all our contents
- The logical granularity of Wikisource, Wikibooks and Wikiversity contents (and maybe other projects?) is not the wiki page but the set of wiki pages storing a book or a course. MediaWiki should be able to support such use cases by providing a "collection" system allowing us to add metadata and to do operations (renaming, add to watchlist) on sets of wiki pages.
- Switch projects that stores fairly structured data in wiki text templates (like Wiktionary or Wikispecies) from a Wikitext storage to a structured one. Build on top of them user interfaces to edit their local contents (and maybe also the relevant data from Wikidata) and provide nice displays and APIs to make humans and machine both able to retrieve these contents.
- ...
