Book 14 · Programming LanguagesPage 12 of 33

Fortran

Formula Translation

Pronunciation: FOR-tran

Updated October 3, 20262 min read

In short

Fortran is the first widely used high-level language, created at IBM in 1957 for science and engineering and still used in high-performance computing.

What is Fortran?

Fortran was developed at IBM by a team led by John Backus and released in 1957. Its name comes from formula translation: scientists could write mathematical formulas almost as they appear on paper, and the compiler turned them into machine code that was nearly as fast as hand-written assembly, which convinced people that high-level languages were practical.

The language has evolved a long way since the punched-card era. Fortran 90 brought free-form source, modules and whole-array operations, and later standards, up to Fortran 2023, added object orientation, parallel programming with coarrays and interoperability with C. Compilers such as gfortran and Intel's ifx keep it fast on modern hardware.

Fortran remains strong in numerical work on supercomputers: weather forecasting, climate models, computational fluid dynamics, physics and chemistry simulations. Fundamental numerical libraries such as BLAS and LAPACK, used under the hood by NumPy, MATLAB and R, have their roots in Fortran.

A common misconception is that Fortran is only legacy code. New scientific code is still written in it, because its array syntax and compilers make number crunching both simple and fast. Its arrays are column-major and start at index 1 by default, which matters when sharing data with C or Python.

Key takeaways

Fortran was created at IBM in 1957 for scientific computing.
It was the first widely used high-level language.
Modern standards add modules, array operations, OOP and parallelism.
It still powers weather, climate and physics simulations on supercomputers.
Core numerical libraries such as BLAS and LAPACK come from Fortran.

Example

Whole-array operations in modern Fortranf90

program waves
  implicit none
  integer, parameter :: n = 1000
  real :: t(n), y(n)
  integer :: i

  t = [(real(i) * 0.01, i = 1, n)]   ! fill an array
  y = sin(t) * exp(-t / 3.0)          ! operate on whole arrays, no loop

  print '(A, F8.4)', 'Maximum: ', maxval(y)
  print '(A, F8.4)', 'Mean:    ', sum(y) / n
end program waves

Readers ask

Is Fortran still used?

Yes, in scientific and engineering computing. Many weather, climate and physics codes that run on supercomputers are written in Fortran, and new projects in those fields still use it.

Why is Fortran fast?

Its design makes it easy for compilers to optimize numerical loops and arrays, for example because it restricts how arrays may overlap in memory, and decades of compiler work have focused on exactly that kind of code.

What is the difference between FORTRAN and Fortran?

FORTRAN in capitals refers to the older versions, up to FORTRAN 77. Since Fortran 90 the name is written with only a capital F.