Celebrating 70 Years of Fortran

Original Article ↗ Hacker News Discussion ↗

Marking Fortran’s 70th anniversary, programmers reflect on how a language born in the punch-card era evolved from bare‑bones formula translation with GOTOs and COMMON blocks into a modern, structured language still central to high‑performance scientific computing. Commenters recount its historical impact (from BLAS/LAPACK to nuclear engineering and climate models), contrast archaic FORTRAN 77 with post‑1990 standards, and debate whether Fortran remains the “right” tool versus C, Python+NumPy, Julia, or Rust. Many conclude that despite its age and quirks, Fortran’s simple semantics, strong array support, and decades of hardened numerical libraries keep it relevant for heavy numerical work.

Historical significance & longevity

  • Fortran is celebrated as likely the oldest programming language still in active use, predating Lisp and COBOL, with Guinness cited.
  • Several commenters recount programming in FORTRAN II/IV in the 1960s and 70s, emphasizing how early it was relative to most modern computing.
  • Fortran is credited with being revolutionary for translating mathematical formulas directly into code and with heavily influencing later languages.

Evolution: FORTRAN → modern Fortran

  • Modern Fortran (90/95/2003/2008/2018/2023) is described as almost a different language from early FORTRAN.
  • Legacy constructs like COMMON, ENTRY, EQUIVALENCE, Hollerith constants, fixed columns, and non-recursive semantics are widely described as archaic but retained for backward compatibility.
  • Some lament that discussions still fixate on FORTRAN 77 despite decades of newer standards.

Current usage & domains

  • Fortran remains common in scientific and engineering computing: nuclear engineering, geophysics, mantle and flow modeling, chemical and civil/structural engineering, climate models, power grid simulation, and oil-related ocean modeling.
  • Core numerical libraries (BLAS, LAPACK, classic linpack/fftpack suites) underpin much of the Python data science ecosystem, even where now wrapped or partially translated.
  • Some institutions and research groups still mandate FORTRAN 77, often due to advisor familiarity and compatibility concerns.

Technical strengths

  • Strong support for arrays and numerical computing; compilers exploit non-aliasing and restricted semantics for aggressive optimization, especially on vector machines.
  • Modern Fortran is seen as simpler than C for domain scientists while still compiling to highly optimized native code.
  • Good interoperability: common pattern is Fortran for heavy numerics with other languages (e.g., C, Python) as frontends.

Critiques, pain points & nostalgia

  • Many recount painful experiences: punch cards, 72-column limits, COMMON block bugs, uninitialized arrays, giant subroutines, cryptic six-character identifiers.
  • Some feel early exposure to FORTRAN 77 turned them off programming; others credit it with foundational understanding of performance, optimization, and multi-user/mainframe environments.
  • Opinions diverge on Fortran’s relevance today: some consider C or other modern languages preferable; others argue Fortran remains “best in class” for numeric simulation and trustworthy legacy code.