On the The Unix Heritage Society mailing list, someone started a
thread commemorating the death of John Backus, who, in addition to
inventing the Backus-Naur form, lead the team that developed the
original FORTRAN. Someone else made a disparaging remark about FORTRAN
and that precipitated a long discussion of the good, the bad, and the
ugly of FORTRAN. Most people who’ve never used FORTRAN—and many who
have—aren’t aware that except in strings and a couple of other places,
blanks don’t affect the syntax and can be omitted. I recall once
looking at a FORTRAN compiler where the first thing it did after
reading a card image was to eliminate all the blanks.
Steve Johnson, the inventor of Yacc, the Portable C Compiler, and
Lint, offered a funny story about FORTRAN and blanks at Bell Labs. His
post is worth reading as an interesting piece of Bell Labs history but
what struck me was how the choice of the right tool for a job enabled
Johnson to complete a task much more quickly than even the world-class
engineers and computer scientists at the lab expected.
The TL:DR is that Johnson, just out of college, had a summer job at
the labs. He was assigned to write a program implementing a state
minimization algorithm. His supervisors expected it to take him the
summer to complete the program. But Johnson had heard about this new
language, SNOBOL, that seemed perfect for the task. Once he got SNOBOL
running, it took him only 3 weeks to do the job.
All of us have go-to tools that we turn to first. In terms of
languages, it might be C, Lisp, Python, Ruby, JavaScript, R, or
whatever we’re most comfortable with. But stopping to consider what
tool will make the job easiest can yield large dividends.