I've been using the pandas library for a long time for financial applications. The data alignment and missing data handling features in pandas are far above and beyond anything else I've used in similar applications.
I think from a data structure point of view it's already better than R (and FAR better than Matlab). R/Matlab should fear the day that the pandas statistical packages gains roughly equivalent features.
I was just trying to get him to clarify his point. There was nothing in his original comment that indicated he believed there was an optimum point between 0-100%. I'm at a loss as to why my comment was so much more inflammatory than his comment about the US being "arguably less civilized".