The problem of information overload has been addressed by several systems. However, many approaches are limited to informal artifacts and need to refer to the user for reference on the quality or usefulness of retrieved information. We make use of semantic technologies, which facilitate the reification and extraction of scientific practice. Based on semantic differences and similarities of semantically marked up artifacts, we identify clusters of users with shared practice, i.e. virtual communities of practice. The common context and preferences of these communities can help users to cope with the selection, structuring, and adaptation of information. 1 Motivation The rise of modern technology has resulted in numerous specialized tools that support various scientific activities, but none of these tools provides an all-surrounding functionality. An all-embracing implementation even seems impossible since the requirements of scientists are very diverse and even contrary. In particular, ...