Had a friend in college who was redoing nuclear decay calculations written in Fortran so they’d work as C++/C# libraries. The calculations had been a historical standby for decades, and people were coming up with increasingly elaborate daisy-chains of dependencies to get them to work properly in modern environments.
There’s definitely a point at which the physical hardware and modern network/interfaces need you to catch up your code with the current technology. But there’s also this terror around trying to touch code that’s got an archaic datestamp on it, particularly if you’re working in a language or dealing with a particularly baroque procedure where the guy who wrote it retired 20 years ago.
Old doesnt mean bad.
Unreviewed Code is bad code unless proven otherwise. Maybe that latest iteration really is time tested and bulletproof. Or maybe Microsoft Execs simply won’t allocate time/money to the kind of routine review and maintenance a codebase needs from time to time.
Had a friend in college who was redoing nuclear decay calculations written in Fortran so they’d work as C++/C# libraries. The calculations had been a historical standby for decades, and people were coming up with increasingly elaborate daisy-chains of dependencies to get them to work properly in modern environments.
There’s definitely a point at which the physical hardware and modern network/interfaces need you to catch up your code with the current technology. But there’s also this terror around trying to touch code that’s got an archaic datestamp on it, particularly if you’re working in a language or dealing with a particularly baroque procedure where the guy who wrote it retired 20 years ago.
Unreviewed Code is bad code unless proven otherwise. Maybe that latest iteration really is time tested and bulletproof. Or maybe Microsoft Execs simply won’t allocate time/money to the kind of routine review and maintenance a codebase needs from time to time.