Open and Closed World Assumption

The Open World Assumption (OWA) is the assumption that a statement may be true irrespective whether it is known to be true. This in contrast to a Closed World Assumption (CWA) where failure to infer a fact implies it to be false and its negation to be true.

Consider the constraint that every animal must have at least one friend and assume there exists an animal without friends. Under OWA the constraint does not cause a violation as there may be an unknown friend for this animal. This is often undesirable for validation. So under OWA a constraint is only violated if there exists a contradiction.

Unique and Non-Unique Name Assumption

The Unique Name Assumption (UNA) is the assumption that different names always refer to different entities, meaning that two different names (or URIs) may refer to the same entity.

The absence of UNA, called the Non-Unique Name Assumption (nUNA), may lead to non-intuitive inference and validation. Take the example by Heflin et al. 2017 where an edge cardinality is constrained to one. If a resource has two of these edges to two different resource identifiers it may look like a violation, but in fact it is valid as they could refer to the same entity. As result it will infer that the the two resources are the same while this may not necessarily be the case.

Conclusion

Validation and inference typically assume different semantics that may lead to different results depending on the type of constraint. Reasoning requires presence of OWA and absence of UNA, where as validation typically depends on CWA and presence of UNA. (Hartmann et al. 2015, Labra Gayo et al. 2017)


Hartmann, Thomas and Acar, Erman and Nolle, Andreas and Eckert, Kai. Sept 2015. The Role of Reasoning for RDF Validation. doi: 10.1145/2814864.2814867.

Heflin, Jeff, and others. 2017. An Introduction to the Owl Web Ontology Language. http://www.cse.lehigh.edu/~heflin/IntroToOWL.pdf.

Labra Gayo, Jose Emilio, Eric Prud'hommeaux, Iovka Boneva, and Dimitris Kontokostas. 2017. Validating RDF Data. Vol. 7. Synthesis Lectures on the Semantic Web: Theory and Technology 1. Morgan & Claypool Publishers LLC. https://doi.org/10.2200/s00786ed1v01y201707wbe016.