Abstract. Metadata processing is recognized as a central challenge for database research in the next decade. Already, novel desktop data management and search applications (cf. Apple's Spotlight and Microsoft's WinFS) are enabled by rich metadata. Efficient and effective access to such data becomes a crucial issue for more and more application scenarios. In this article, we focus on metadata represented in RDF. A number of query languages for RDF have been presented in recent years. This article argues that most of these approaches fail to address properly two core issues: the provision of rich operators and constructs to adequately support RDF's graph data model and the ability to intertwine access to metadata (in RDF format) and data (in XML format). To address this points, two XML views over RDF data are expressed in the query language Xcerpt and discussed. Furthermore, it is shown how these views together with Xcerpt's rich graph patterns allow the succinct expression of complex, but common queries against RDF graphs.