I work in applied mathematics and specifically in CFD. This may make me not a "true mathematician" to some of the mathematical community. In fluid dynamics nearly everything is an approximate solution to an imperfect model. In fact, there are whole branches of mathematics dealing with trying to figure out exactly how imperfect models are. "Perfect math" has limited applicability, and only perfectly reflects reality in contrived situations. For example, linear algebra works amazingly well in the contrived environment of a computer, but if you want to perfectly model the electron flow that is involved in that computation one must necessarily use an imperfect approximation.