In the context of the European Digital Agenda governmental authorities as well as virtual communities (e.g. datahub.io) have published a large amount of open datasets. This is a fundamentally positive development, however one can observe that many different that both for the metadata and the data itself different vocabularies are in use. Furthermore, the established vocabularies are often augmented with portal specific metadata standards and published in different (local) languages. If Open Data are to be integrated and aggregated across portals, this entails a lot of effort. In this paper we present examples how the problems of interoperability and multilingualism could be addressed for key open data asset descriptors. We focus on the analysis of today's problems and ways to solve them.