FORTRAN (was Re: indentation)

[skaller]

Mark Jackson wrote:

> We have legacy code written, and "legacy" xerographic physicists most
> comfortable, in the language.  

The physicists are right, because they care about efficiency.
Fortran is still MUCH faster than C. I recently implemented a
Fortran back end for a high performance array processing compiler,
there was a C back end too. Initial tests indicated about
the same speed -- until the boss pointed out I forgot to use
the optimisation switch on the fortran compiler. With that
switched on, f90 executed the SAME highly optimised program
something like 30% faster than C. [Using Solaris on a large sparc
server]
[now consider parallel computing ...]

C9X has a few new features that might narrow the gap --
variable length arrays on the stack, a 'restricted'
qualifier for function arguments which emulates fortran's
built in rules about aliasing, and inlining.

[Forget OO, it was designed to support abstraction dynamically,
and that usually involves overhead]

-- 
John Skaller, mailto:skaller at maxtal.com.au
10/1 Toxteth Rd Glebe NSW 2037 Australia
homepage: http://www.maxtal.com.au/~skaller
voice: 61-2-9660-0850