Describing web resources using formal knowledge (i.e., creating metadata according to a formal representation of a domain of discourse) is the essence of the next evolution step of the Web, termed the Semantic Web. The W3C’s RDF/S (Resource Description Framework/Schema Language) enables the creation and exchange of resource metadata as normal web data. In this paper, we investigate the use of RDFS schemas as a means of knowledge representation and exchange in diverse application domains. In order to reason about the quality of existing RDF schemas, a benchmark serves as the basis of a statistical analysis performed with the aid of VRP, the Validating RDF Parser. The statistical data extracted lead to corollaries about the size and the morphology of RDF/S schemas. Furthermore, the study of the collected schemas draws useful conclusions about the actual use of RDF modeling constructs and frequent misuses of RDF/S syntax and/or semantics.