Distributed information retrieval is a well-known approach for accessing heterogeneous, highly autonomous sources of unstructured information. Selecting and querying only a number of relevant sources can help improve its performance, but most resource selection algorithms are limited to syntactic comparisons. We present a framework for applying resource selection in the context of a semantic federated product information system, and evaluate the performance of the well-known CORI resource selection algorithm in this context. Key words: resource selection, distributed information retrieval, federated search, enterprise search, enterprise information systems