We still have a whole lot of FORTRAN77 around at my work that I will do some work on. Some of it is very close to FORTRAN-IV. These are Gen-IV nuclear reactor simulation codes written by Argonne National Lab and friends mostly in the 1960s-90s. They were validated against experimental facilities back then and therefore have a good pedigree. Since most of the experiments are no longer operational and not all the results and test conditions were well documented, it's a challenge to update. We've upgraded some of the memory management stuff through the years at least and there's more FORTRAN90 every day, which is a nice modern language from this perspective.

Thankfully we have written gobs of Python to do most of the data management, multiphysics coupling, multiobjective optimization, etc. so it's really not too bad.