Re: language design tradeoffs

jlg@cochiti.lanl.gov (Jim Giles) writes:
>3) The EOL character should be the statement terminator. [...]


rob@guinness.eng.ohio-state.edu (Rob Carriere) writes:
> Hm. Did whoever did that research take large mathematical expressions
> (which naturally translate to multi-line statements) into account? My
> single most common syntax error in FORTRAN and Matlab is forgetting the
> continuation marker on a multi-line statement.


It may well have taken that into account. Age-old research by Barry Boehm
and others concluded that most FORTRAN assignment statements were of the
form A = B, most of the rest were assignments to binary expressions -- A =
B + C. I don't trust my memory of actual numbers, but am fairly sure that
the first group accounted for >50% of all assignment statements, the 2nd
class was perhaps around 30%. Statements complex enough to need
continuation lines were a VERY small percentage of FORTRAN code, something
less than 5% I believe.


Based on Robert Firth's comments I'm inclined to think BCPL had the best
approach, treating EOL as a statement terminator if it makes sense
syntactically.
--
Paul Raveling
Raveling@Unify.com
--

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