Introduction

Controlled vocabularies are a widely-used tool (e.g., in libraries or as taxonomies in scientific research) to structure and organize information. Thesauri, for example, define standard terms and the relations among them in a specific domain. Their value lies in providing a model of an area of knowledge that can be used, e.g., to index documents for information retrieval. Several standards exist for expressing controlled vocabularies but with the wide adoption of the Linked Data concept, the Web-based SKOS data model has become the choice of many contributors who want to integrate their controlled vocabularies into an interconnected Web of Data.

We are investigating whether or not it is possible and useful to measure the quality of SKOS vocabularies and, most importantly, how this could be achieved. We propose a set of quality issues that can lead to the development of computable metrics that indicate the quality of controlled vocabularies on the Web. Given the growing number of SKOS vocabularies, we believe that such metrics could complement the intellectual quality assessment process domain experts must pass through when choosing vocabularies from the Web.