Schema.org has become the standard vocabulary for publishing structured data on the World Wide Web. Unfortunately, the flexible data model of Schema.org and its pragmatic view of conformance result in uncertainty for the publishers and consumers of structured data, which has consequently led to the fact that Schema.org is broadly used, but largely in an incomplete and wrong way. In this thesis, an
approach is elaborated to solve this problem through the introduction of semantic templates, which the author names "domain specifications". Domain specifications are formalized, machine-readable, and structured documents that define syntactic and semantic, domain-specific constraints for Schema.org annotations. This thesis also specifies a validation algorithm that verifies if a Schema.org annotation is in compliance with the constraints defined by a domain specification. The concepts proposed in this thesis serve the maximal goal to enable the domain-specific validation of Schema.org annotations and consequently increase the quality of structured data on the Web.