Number of mapped ontologies – more is
better – target 3 during experiments
Complexity of mapped ontologies  -- more complex is
better – target 5+ level hierarchy, 100+ classes, 20+
properties
Difficulty of mapping phenomena – more difficult is
better – target 5+ requiring code support
Fidelity of translation – Precision (% correct)
– Recall (% correct coverage)  -- Higher for both
precision and recall is better but where is the optimal
tradeoff?