Noam Chomsky and John Backus

reinhardgantar@hotmail.com
13 Sep 1998 22:56:54 -0400

          From comp.compilers

Related articles
Noam Chomsky and John Backus reinhardgantar@hotmail.com (1998-09-13)
Re: Noam Chomsky and John Backus rweaver@ix.netcom.com (1998-09-18)
| List of all articles for this month |

From: reinhardgantar@hotmail.com
Newsgroups: comp.compilers
Date: 13 Sep 1998 22:56:54 -0400
Organization: Compilers Central
Keywords: history, question, comment

Dear comp.compilers-community,


Here is a folklore-question for you:


My various compiler books state that John Backus (of Backus/Naur-fame)
was in charge of developing the original FORTRAN-compiler. That
project initially was a big mess because the team used a
seat-of-pants- approach that was not adequate for the complex task.
The idea of transformational grammar was not available to them
yet. Looking at the FORTRAN-syntax and its constructs supports this
evidence, at least from todays vantage point and more than 20 years of
compiler construction tools.


In 1957 Noam Chomskys published his book "Syntactical Structures" and
at about the same time John Backus came up with his concept of syntax
trees (a prominent feature in Chomsky's book) and corresponding
technical concepts. Since then the implementation of compilers is a
pretty straight-forward affair - John Backus got his project on
track. My question here is: Was John Backus directly or indirectly
inspired by Chomsky? Chomsky was an MIT-teacher at that time, so
Backus *could* have gotten hold of his work fairly early and in that
light the time table looks plausible. (I *think* the first FORTRAN
came out 1960) Does anybody know in what order things happened back
then? Personal accounts, anyone?


Thanks for your attention
Reinhard Gantar
[The Fortran project started in 1954 and shipped in 1957. But the
critical idea that made it possible to write compilers in a reasonable
way was syntax directed translation, that is, associating chunks of
object code with syntax rules and building up the translated code.
Ned Irons published the original paper on that topic in 1962. -John]
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.