Although the Semantic Web vision is gaining momentum and the underlying technologies are used in many different areas, there still seems to be no agreement on how they should be used in everyday documents, such as news, blogs or wiki pages. In this paper we argue that two aspects are crucial for the enrichment of this documents with semantic annotations: full support for RDF and close integration of the annotations with the continuous text. The former is necessary because many common relationships cannot be expressed by attribute-value-pairs, the latter reduces redundancy and enables Web browsers to help readers using the contained data. To gain further insights, we implemented an RDFa-capable extension for MediaWiki and report on improvements for wiki use cases and other applications on top of the contained data.