Our first conversation is about the quantitative measurement of open data. That is one particular good example, because as a proxy, everybody who see, "Oh, 500,000 datasets — it means something." Even though it means nothing really. It correlates with nothing. [laughs]