I've worked with numerous people with PhDs in Physics (and Math and other heavy science fields) and they tend to be pretty awful at production software while at the same time having very strong opinions that the field is "broken". I mean, think about it, 99.99% (at least) of physics and math is learning what has already been solved. It is heavy, complicated learning, but is far from the type of learning you do in a new field like software development.
Non-CS people tend to be completely pragmatic in my experience. They aren't the ones having arguments over the 'art' side of software engineering like everyone else.
