See sources below...
Fortran is one of the earliest "high-level languages" ever created. As such it has an immense amount of baggage, folklore, and old, old source code associated with it.
Fortran is an acronym for FORmula TRANslator. It was created in the late 1950's at IBM by a team led by John Backus. The idea was to create an efficient high-level (for the time) language in which it was easy to express mathematics. At the time numerical programming was dominated by machine-specific assembly languages.
Fortran was the first computer language to use common math syntax for arithmetic operations, including:
• "a + b" for addition
• "a - b" for subtraction
• "a / b" for division
• "a = expr" for variable assignment
Lacking a handy multiplication symbol, Fortran was the first organized system of any kind to use "a * b" to denote multiplication.
Fortran rapidly became popular with scientists and engineers, due to its intuitive (for the time), compact (for the time) notation. As scientific computers proliferated throughout the 60's and 70's Fortran came to be loved for its (for the time) portabliilty. One could take a metallurgy, or automobile crash testing code written for an IBM 360, and, with only a few weeks or months labor, convert it to work efficiently on a VAX or a CYBER-205. This may not sound that great today but there would never have been any hope of porting such an app written in assembly language, it would have required a complete rewrite from the ground up, years of effort on each new system.
Fortran grew over the years to capitalize on these strengths, becoming the first language to add:
• strong (for the time) numeric typing
• multiple floating point precision support (coining the term "double precision"
• a native complex data type
Modern Fortran features an number of other specialized features for scientific computing:
• true multidimensional arrays
• compact array/vector syntax for large scale initialization, reading, and writing.
• input and output formats for scientific and engineering notation
• built-in functions for floating-point precision selection, vector dot product, matrix multiply, etc.
Notice that conspicuously absent from this discussion are text processing capabilities, graphic capabilities, OS system call support, object technology, etc. For years Fortran standardization concentrated on the language's perceived strengths for scientific programming and left "all that other stuff" to languages like C, C++, and Java. As such, the "first ever general purpose programming language" has now been relegated to "usually useful for numeric stuff".
The most recent Fortran standard (2003) has corrected that somewhat, adding POSIX/C compatibility and object technology, but it still remains the fact that most new scientific/engineering code is written in C, C++ and (to a lesser extent) Java. There is still an huge volume of Fortran legacy code in use out there. It still needs to be tweeked and maintained and occasionally updated to add a new feature or so. And there are people writing new code in new Fortran.
But most people who do that sort of computing would agree that Fortran's time of dominance has long passed. It is now one of many suitable choices for programming in the scientific/engineering niche market.
Yahoo! Answers does not like these URLs. Take out the spaces to get them to work:
http://e n.wiki pedia.org/wiki/Fortran