From FRANCIS GLASSBOROW
Metcalf does not do justice to the strengths of Fortran, its reason
for surviving so long and the problems it now faces. Initially Fortran was
grabbed by those concerned with numerical computation as being just what
they needed. It was considerably simpler than Algol in syntax and semantics
yet supported some attempt at standard input/output.
During the two decades when Fortran was the only general-purpose language
to support numerical specialists, these specialists have become accustomed
to using Fortran as a tool for communicating algorithms to each other. Even
if no further programming were done in Fortran, it would still survive for
many years filling this vital role.
The problem that needs to be tackled in the 1990s is that of writing
the user interface for programs in the various flavours of graphical interface.
Fortran is completely unsuitable for such purposes. An increasing number
of Fortran users are recognising this and are looking at languages such
as C++ for such provision. In the meantime, the C and C++ communities are
recognising the value of the numerical resources locked up in the Fortran
libraries. This realisation is causing the implementers of C and C++ compilers
to include support for access to Fortran from within C and C++.
Some of your readers may be interested to learn that one reason for
the quick production of the Numerical Algorithms Group Fortran 90 compiler
is that it works by translating Fortran 90 into C, which is then compiled
with any of the existing standard C compilers.
Advertisement
Francis Glassborow C Users Group (UK) Oxford
