# Keep (what's working well) - let challengers provide their own endpoints - instead of request docker images to run in a controlled environment for example (@anbo-de) - providing the CLI esp. for self assessment (@anbo-de) - ... # Change (what needs improvement or should be stopped) - to test against "success by accident" - also validate the SPARQL query for correct answers (@mommi84) - ... # Do More (what should be amplified) - new category: "ontology known, instance data unknown" (@anbo-de) - provide a better detail view on each questions and how it was performed for each endpoint (@seebi) - include questions with intentional empty result set (@mommi84) - ...
