1. 14

  2. 6

    And still running strong on 3% of our Octave codebase.

    edit: this is across core Octave and our “external” packages, which due to the tight integration people expect between Matlab and toolboxes, doesn’t end up being so external after all.

    1. 5

      27% of R, which sounds high, but then the implementation has been around a long time, before lots of numerical code got rewritten in later languages.

      1. 3

        Yeah, it’s not like we get a lot of new Fortran code in Octave either. We did get a bit of new Fortran code around 2006 from a single Fortran enthusiast, but it wasn’t a huge amount.

      2. 3

        One of IBM’s crowning achievements.

        1. 3

          I work on HPC clusters that do scientific simulations, and Fortran is alive and well in our world. Not just for libraries, but for plenty of new applications. Though C++ is gaining ground…

          1. [Comment removed by author]

            1. 3

              That happens a lot in the CAD industry, too. New differential equation solvers, triangulators, intersection calculations, etc. will be implemented in more modern languages, and may even be faster and more accurate, but the older code stays around too, because it’s important the old part files are exactly the same as they’ve always been when they’re opened in the modern versions of the software.