Re: Algol 60 Syntax

> (actually, I think BNF was developed specifically in order to
> describe Algol-60)


Minor quibble. Actually BNF was essentially developed by Backus to
describe, post-priori, Algol 58's syntax at a UNESCO conference in
1959. It was adapted (and adopted) by Naur for the development of
Algol 60 when he discovered that the language Backus described was not
what he thought they had designed and that the textual description was
ambiguous enough to allow both his and Backus's interpretations.


Major point: The poster seems to be unaware of all the semantic
subleties that can complicate the translation from one language to
another. Before he invests too much time in his project I suggest he
think about the implications, for example, of the large number of
different arguments passing mechanisms: call-by-value, call-by-name,
call-by-reference, no modifiable aliasing (Fortran), call by need,
copy-in/copy-out, call by descriptor, etc. That alone makes writng a
general purpose C to Fortran translator impractical, for example.


William B. Clodius
[It's true. But ironically, the other direction is easy because Fortran
is deliberately vague about all that stuff so valid Fortran programs
can't depend on one or another argument passing scheme. -John]

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