The title is somewhat misleading: it is not the core Wikidata database which is splitting, it is just the WikiData Query Service (WDQS) which enables users to run SPARQL queries over its data.<p>WDQS is already very difficult to use – many queries run extremely slowly or time out completely, and working out how to write a query in such a way that it performs well requires a high level of expertise, both in SPARQL in general, and in the details of Wikidata's implementation of it in particular.