> It would be interesting to know if anything older than Fortran 77 is still in active use.
In active use, that's actually easy, especially in the industrial space (that's why I mentioned distributed as well as active, which gives another metric entirely)
Just about a year ago I was tasked with recovering software from a HP1000/RTE machine (driving an industrial oven in a forge), written in a Fortran dating back from before it was properly standardized.
We recovered the code through the machine's serial port, with the built in print functionality, which conveniently prepended a header along with the file's content. Together with pyserial at the other end, a few hours of hacking had a laptop listen via a PL2302 serial/USB adapter, parsed the printed header to get filename and a few metadata, and wrote content to the proper location.
The harder part was manually moving through directories, and printing the files at 1200 baud, through a terribly laggy and completely burned out 80x23 passive monitor.
Headers in the code mentioned it was actually older than me.
With a few massaging, the code actually compiled on a recent gfortran (F95!), but hte linking part was entirely different, and many proprietary symbols were missing. We went through the 30 years old documentation and found what we needed to stub them and plan how to act forward. The thing was actually quite close already to Unix and what we find on Linux these days, only much much more primitive. The code contained critical functions whose role was to compute thermal data for automation, according to various (thermodynamic and other) complex and undocumented rules, so I built a detailed plan about possible actions to take, including linking to C wrappers binding the stuff into Python, complete with a prototypal HTML5 visual status feedback of the automation system. That's when I left the project (and the job).
In active use, that's actually easy, especially in the industrial space (that's why I mentioned distributed as well as active, which gives another metric entirely)
Just about a year ago I was tasked with recovering software from a HP1000/RTE machine (driving an industrial oven in a forge), written in a Fortran dating back from before it was properly standardized.
We recovered the code through the machine's serial port, with the built in print functionality, which conveniently prepended a header along with the file's content. Together with pyserial at the other end, a few hours of hacking had a laptop listen via a PL2302 serial/USB adapter, parsed the printed header to get filename and a few metadata, and wrote content to the proper location.
The harder part was manually moving through directories, and printing the files at 1200 baud, through a terribly laggy and completely burned out 80x23 passive monitor.
Headers in the code mentioned it was actually older than me.
With a few massaging, the code actually compiled on a recent gfortran (F95!), but hte linking part was entirely different, and many proprietary symbols were missing. We went through the 30 years old documentation and found what we needed to stub them and plan how to act forward. The thing was actually quite close already to Unix and what we find on Linux these days, only much much more primitive. The code contained critical functions whose role was to compute thermal data for automation, according to various (thermodynamic and other) complex and undocumented rules, so I built a detailed plan about possible actions to take, including linking to C wrappers binding the stuff into Python, complete with a prototypal HTML5 visual status feedback of the automation system. That's when I left the project (and the job).