Which programming language is best? (part 2)

There are a few languages better suited for scientists, each with upsides and downsides

Seven years ago I gave a presentation at a meeting of Python programmers. I investigated some of the questions I’ve been dealing with in this newsletter. Among the things I said was that in five or ten years, “Python + numpy + pandas may be looking as obsolete as Delphi does now and as Fortran did in the 1990s”.

This Fortran-looked-obsolete-in-the-1990s statement is my younger self speaking from the past, when I lived in my own microcosm. I no longer think Fortran looks obsolete, and I don’t believe it looked obsolete in the 1990s either.

I wanted to clarify that my previous email isn’t a recommendation to use Fortran. It’s just a fun fact: I used to laugh at it but now I respect it.

For most scientists and engineers, R, MATLAB, Python+numpy(+pandas), and Fortran all seem to be good choices. Each one has upsides and downsides. What a new scientist will learn will usually be a matter of circumstances and/or curiosity.

Just make sure you don’t use COBOL 😀.