Motivation

Many different SKOS vocabularies have been published on the Web and it gets increasingly difficult to find the "right one" to adopt or reuse. Vocabulary development is a complex process that relies on expert knowledge from possibly various domains and takes into account specific use cases, user groups and different views of the world. Furthermore, different thesauri may be created and maintained by unequally skilled thesaurus managers, rely on consortium decisions, or may be created collaboratively. As a result, many vocabularies with varying "quality" are available on the Web today.

The goal of this study is to determine what "quality" means in the context of controlled vocabularies and how automated mechanisms can assist experts in determining quality issues. From a Linked Data point of view, it might, for instance, be important that vocabularies don't form isolated "islands" but are linked with other resources on the Web. From an organizational point of view, hindrance in formulating inter-domain queries constitutes a quality issue.

An analysis of existing Web vocabularies against a well-defined set of quality issues could be a valuable input for organizations that need to decide on the adoption of vocabularies but also an important feedback-mechanism for vocabulary developers. We believe that thesaurus quality assessment is important

• for the vocabulary creator(s) to detect mistakes/areas of disagreement
• for contributors to the LOD cloud to find suitable vocabularies
• to increase Linked Data cohesion
• to achieve better results in search/retrieval scenarios

The purpose of this wiki page in this context is to

• present the already identified quality issues and how they are implemented by the qSKOS tool
• collect ideas for other quality issues, complementing our current list
• get community feedback on the qualit issues collected so far, e.g., if and how they influence the interaction with vocabularies or how they lead to improvements