Orcmid's Lair status 
privacy 
 
about 
contact 

2007-03-20

John Backus: December 3, 1924 – March 17, 2007

I just learned about the death of John Backus from a testimonial to Fortran on the blog of programming-writer Charles Petzold.

There is more about John and his career at MSNBC and CNet.  The slashdot coverage is predictably banal.  Wikipedia is on top of events, although I am not clear about the sketched relationship of FP (the language, not the field), FL, and Iverson’s J.  The New York Times account is more rounded yet skimpy-feeling. Λ the Ultimate has a touching blurb and comment threads.  His 1977 Turing Award Lecture (PDF file) is also on-line.

John’s influence did not end with Fortran of course.  He established a powerful form of description, now known as BNF, that led to far greater regularity and expressiveness in later programming languages.  The concrete syntax of XML is specified in a version of BNF, as are Java and similar modern languages. 

Following work on ALGOL, Backus moved on to consider how to liberate programming from what he termed the von Neumann style.  He devised the language FP (and a formalized concrete version, FFP that is more like a programming language in the way that LISP is).  FP illustrated higher levels of abstraction that harmonized concepts of the functional-programming community and the array-oriented APL language of Kenneth Iverson.


I only saw Backus in person once, although I spoke to him on the phone another time (about functional programming, not Fortran) and I became acquainted with Ed Wimmers, who worked with John, by coincidence.

Fortran was my first programming language, before there was a Roman Numeral on it.  Fortran II was in beta and we were still using the original IBM 704 compiler. My first program was an experiment with a particular mathematical formula that I discovered was unstable (lots of opportunity for division by zero) when transcribed directly into Fortran.

It only now struck me that I’ve had the remarkable experience of being around at that time (I was 19 in 1958) of Fortran’s birth, with no expectation that at the end of the 60's I would be the Chairman of the ANSI X3J3 Fortran subcommittee.  That was in the transition window between Fortran 66 (aka Fortran IV) — the first-ever ANSI Standard for a programming language — and initiation of work on Fortran 77. By then I was still involved with Fortran but my devotion was Algol 60, functional languages, and languages for platform software (compilers, tools and operating-systems) development.

[update 2007–03–20T21:43Z Add NYT, lambda-the-ultimate, and Turing Award links]

 
Comments: Post a Comment
 
Construction Zone (Hard Hat Area) You are navigating Orcmid's Lair.

template created 2002-10-28-07:25 -0800 (pst) by orcmid
$$Author: Orcmid $
$$Date: 07-02-17 11:08 $
$$Revision: 26 $