A radical new look at science

Kantian question: what scientific and mathematical theories are strictly required for the design, engineering and effective implementation and production of modern technological and industrial artifacts (a computer, a drone, a car, a MRI machine, a satellite, a vaccine, a nuclear reactor). This is philosophically similar to the simpler question Aristotle posed about products of art (technê). It is seems plausible that there is a huge amount of published theoretical material which is neither necessary nor useful, strictly speaking.

A disturbing question: there are presently technological artifacts which are not fully understood (or even capable of being practically engineered) by a single human being. For instance current versions of the most common operating systems. Could technological artifacts come eventually to automate their own design, production and maintenance outside the possibility of human design and control ?

This approach favors the view that the differential mathematical apparatus in science is never meant to be nor should be read as a model strictly speaking but as a computational device synthesizing modelling information for concrete more realistic and complex models. In other words, the Navier Stokes equations do not model fluids, we are in the presence of a mathematical device for compactly representing the potential computational information of actual approximate models of reality. Abstract devices which are used to extract finitary computational content (like the convenient fictions of Leibniz but without taking any philosophical position on mathematical platonism).

There are interesting arguments for the Lebesgue integral being useless and meaningless in the context of actual concrete engineering practice (limited by measurement and scale).