A survey to identify factors for vocabulary reuse and requirements for vocabulary recommendation tools

Abstract

The choice of appropriate vocabularies is one essential aspect in the guidelines that guide a data engineer when modeling Linked Open Data (LOD). In general, this leads to an easier consumption of the data by LOD applications and users. However, making decisions considering the adequacy of various vocabularies is not straightforward and a well known challenge; the same applies to the engineer's decisionmaking regarding the total number of vocabularies used in one dataset. Therefore, it is not surprising that according to some LOD data provider studies, there is still an insufficient compliance towards this particular best practice. In this paper, we examine the current importance of the best practice “vocabulary reuse”, as well as the factors that influence the engineer's decision whether to reuse a specific vocabulary or not. We provide results of an online survey comprising an aggregation of knowledge, practices, and design motivations of several LOD publishers and practitioners with respect to the reuse of vocabularies. These results show that the insufficient compliance considering vocabulary reuse is not because of its lack of importance, but most likely because of deficient tool support for deciding which and how many vocabularies to reuse. We address the increased need for such tool support, and based on the results of the study, we derive several requirements for future vocabulary recommendation tools.