Feynman did some things that were obviously great. Feynman also did some things that were obviously terrible. Neither of these conclusions requires a "nuanced"

Feynman did some things that were obviously great. Feynman also did some things that were obviously terrible. Neither of these conclusions requires a "nuanced" portrayal. There's a sort of hidden assumption that we need to assign a univariate reputation to people (historical and otherwise), which brings up the difficulty of how to combine these two observations on incomparable axes into a single conclusion. That doesn't require "nuance" either; the fundamental problem here is trying to figure out if thing X outweighs thing Y. That combination isn't going to be particularly informative about the actual person under discussion; it'll mostly just tell you what coefficients you're using when weighting the two observations. Surrounding every negative observation by positive observations is neither scientifically honest nor informative about the observation itself. It just says that the writer thinks the negative observation shouldn't carry as much weight. In this case, the author is making no such judgment and presenting the facts directly, which I think is an entirely reasonable thing to do.