Is There a Role for Standards in the Future of Fortran?
Miles Ellis · ACM SIGPLAN Fortran Forum · 1998
The first Fortran compiler was delivered to customers by IBM in April 1957. By the time the author wrote his first Fortran program in 1965, there were more than 50 Fortran compilers in existence - - and almost as many different versions of the Fortran language! The first (American) Fortran Standard came the next year; the first Interna- tional Fortran Standard was published in 1980 following the publication of the American Fortran 77 Standard in 1978. Further Fortran Standards followed in 1991 and 1997. However, the innovation shown by the early Fortran compiler writers has now been replaced by a reluc- tance to add any significant new features to compilers until they have been standardised internationally N and the international standardization process can be very slow! It seems ironic that in a world in which a new lan- guage such as Java can be conceived, developed, launched, and become dominant in its field in less than five years, it takes at least twice that long for a major language such as Fortran to add a few simple features. Standards have undoubtedly replaced the chaotic, competitive, potentially incompatible language develop- ment of the 1960s with a regular, safe, well-coordinated language development of today. But at what cost? This paper will argue that the cost may be too great, and that Fortran must break out of tyranny of the International Standards process if it is to survive into the 21st century as the major language that its history deserves.