Originally Posted by mdatmo
Yup, the scientific community (including myself) uses a lot of FORTRAN. As it is array orientated it is easy to quickly write fast code for numerically solving equations. Fortran 90/95/2003 (I don't recall which feature came when) also feature many new capabilities such as derived types, dynamic memory allocation, and object-oriented programming support.
A very basic FORTRAN resource from NCI.Another resource,
however it doesn't cover the new additions in 2003.Another online
resource I use when I need to look something up.
I can attest to this but it's very situational. For example, for small algorithms or less than 100 lines of code, you will see a very distinguishable difference in the execution speed between Fortran, C and Python.
However, when you get into very large algorithms with a desire for memory efficiency and parallelism, you will be hard pressed using FORTRAN or even C. At that point, a language like Python becomes a lot more efficient if done correctly and that's often the difference between a good lab and a lab of novices or outdated coders. Python done correctly will see a small difference in shorter algorithms but will be much much better for large algorithms with a lot of efficiency and parallelism being implemented.
I think every engineer should learn the basics for FORTRAN though, just for my aforementioned industry advantage.