1: Randomized selection with tripartitioning 2: Initial Experiences Re-Exporting Duplicate and Similarity Computation with an OAI-PMH aggregator