Re: Languages in use

[jcm@ORNL-MSR.ARPA (James A. Mullens)]

From Jim Mullens / Oak Ridge National Lab
Oak Ridge is a large government research lab.  I, and most people here, are
part-time programmers and full time engineering/scientific researchers so we
tend to stick with the common coin of the scientific programming world,
FORTRAN.  Most people have grown up with IBM mainframes and DEC minis, where
FORTRAN is well-supported.  On DEC minis at least, the language is highly
integrated into the operating system, so you can even do most systems programming
from FORTRAN...  I think the major reason we do not change languages is that
we have some very good reasearchers who only know FORTRAN, and could not
participate fully in a computer project using another language -- plus, we
simply cannot afford to learn a new language just because it is claimed to be
the optimum for this year's computer project.
Lately I think we are seeing the growth of easy-to-use speciality packages
which are almost languages, instead the learning of new languages.  CSMP is
an early version of this, spreadsheets another, advanced data base packages,
statistical packages like SAS, and so on.  This is, to me, an interesting way
to go...  In AI, expert systems are characterized by the incorporation of
"domain-specific knowledge" instead of general rules of reasoning.  These
language-like application packages seem like languages with domain knowledge
embedded in them.
I would like to hear the results of your survey.  Thanks.