Project:Scatterbase

From Librarybase

Scatterbase is a from-scratch reimplementation of Wikibase. It was first conceived in 2019 and the development of a reference implementation in the Rust programming language began in 2025.

Rather than fight data ingest limits or build a system parallel to MediaWiki, the new goal is to develop Scatterbase as a replacement backend for Librarybase.

Foreign entity space

Librarybase-native items and properties are prefixed Q and P as before.

A new class of entities with three-letter prefixes will allow foreign entities to be natively incorporated into the Librarybase graph. The first two letters are a provider and the third is a namespace. These should be assigned sparingly, when the alternative would be the reification of millions of alias entities. Rather than reflect a particular database, they reflect a directly incorporated source. For example, because OpenAlex already includes what is in PubMed, there is no need for a separate PubMed namespace.

  • WD: Wikidata
    • WDQ: Wikidata items
    • WDP: Wikidata properties
  • OA: OpenAlex
    • OAW: OpenAlex Work
    • OAA: OpenAlex Author
    • OAS: OpenAlex Source
    • OAI: OpenAlex Institution
    • OAT: OpenAlex Topic
    • OAK: OpenAlex Keyword
    • OAP: OpenAlex Publisher
    • OAF: OpenAlex Funder
    • OAG: OpenAlex Award/Grant

Identity collapse

There will be cases where a record exists in more than one place, such as a journal article being in both Wikidata and OpenAlex. The same way Q numbers can redirect to other Q numbers on Wikidata, a foreign entity identifier can redirect to something else.

If two or more entities are found to be exact matches of each other, the canonical identifier is chosen deterministically based on a precedence order: Librarybase-native identifier, then Wikidata, then OpenAlex. The other identifiers automatically redirect to the first preferred one.