Looking at intermediate results of the Controlled Vocabularies Survey an interesting finding concerns the question which types of knowledge models are currently best fit for actual use in applications. So