I started programming computers in February of 1967, when I was a junior in high school. After dropping out of grad school I began a 41 year career in information technology in January of 1977. I have seen a lot of computer languages come and go. So I read 5 Programming Languages You Won’t Likely Be Using by 2030 with some interest. The only one on the list I had ever used was Perl. It was kind of fun, but I did not get very attached to it 🙂

Meanwhile, some much older languages live on. Last year COVID-19 demonstrated how much the financial world still depends on COBOL:

• An Ancient Computer Language Is Slowing America’s Giant Stimulus
• What Is COBOL, and Why Do So Many Institutions Rely on It?
• Why COBOL will never die

I never seriously used COBOL, but starting in high school I used FORTRAN II and FORTRAN IV. During much of the 1980s I worked with FORTRAN 77 in a VAX/VMS environment. In 2021 FORTRAN is also still going strong:

• Old Is Gold: This Very Old Programming Language Is Getting Bold Again
• This old programming language is suddenly getting more popular again

All of this suggests that the Lindy effect applies to programming languages.

Of course, FORTRAN has changed over the years. Arthur O’Dwyer’s Making Castlequest compilable details the work involved in making a FORTRAN IV program work in a modern environment. It reminded me of something I heard c. 1980: “Nobody knows what the scientific programming language of the year 2000 will look like, but everybody knows it will be called FORTRAN.”

FORTRAN IV did not have any kind of character data type, but it was possible to read character strings into numeric variables and manipulate them from there. O’Dwyer describes some of these methods, which I remember from back in the day, notably the EQUIVALENCE Statement. IIRC, some company (IBM?) actually wrote a FORTRAN compiler in FORTRAN. FORTRAN 77 and subsequent versions do have a character data type, and the tools to manipulate character variables directly. With these changes some of the old work-arounds are no longer allowed and O’Dwyer had to revise the code for Castlequest. However, it gladdened my heart to see that Hollerith constants are still supported in GNU FORTRAN. I also enjoyed reading about how he coped with FORTRAN carriage control characters, another part of “normal” for me in the old days.

Despite the nostalgia, in programming as in so much else in life, I do not want to go back to the “good old days.” If I ever write FORTRAN code again (stranger things have happened!), I will not use Hollerith constants, FORTRAN carriage control, or any of those other tricks.