Re: A question on Newton's Method

kia wrote:
> Jon Harrop wrote:
>> archaic theoretical foundation, other languages aren't. Unless you are
>> forced to use old code, I see no reason to use Fortran.
>
> Here's your reason, several million times over.
>
> http://netlib.org
> http://gams.nist.gov
>
> Until another language comes along with a multi million *open* book
> repository there's every reason to join Fortran literacy club.

No, not at all. Firstly, the quality of most of the code on those sites is
very poor. The minority of codes that are well written are already bundled
into libraries with APIs, have bindings to decent languages and are put
into most Linux distributions. The vast majority of users are not going to
want to tinker with the innards (if they do, it's the sign of a badly
written library) so there isn't even any point in these being open-source.

As far as recommending Fortran to a new programmer, that voluminous quantity
of mostly-hideous code is an excellent counter-example. The "conv2b.f90"
program posted by James Van Buskirk is exactly the kind of abomination I'm
talking about. New programmers should never see that kind of code.

In contrast, the code in Mathematica is vastly more robust. The ability to
do a Fourier transform by simply using the "Fourier" function is hugely
beneficial, for example. If I wanted to do a Fourier transform from Fortran
I'd probably use FFTW which is written in OCaml and C and I wouldn't even
dream of fiddling with the source to FFTW, for fear of breaking it.

--
Dr Jon D Harrop, Flying Frog Consultancy
http://www.ffconsultancy.com
.

Relevant Pages

  • Re: OT: Different types of counting in loops
    ... In fortran you have: for 10 i=1,5 ... some practical reason for it, or purely a decision of the original ... Why have most of todays languages adopted counting/indexing from 0, ... Algol had more influence on C and Pascal ...
    (comp.lang.fortran)
  • Re: variable names for different fortrans
    ... That looks a bit too much like the way other languages identify constants. ... For that reason I tend to avoid it in Fortran. ... Having said that, when I write Delphi or VB I use type identifier prefixes, ...
    (comp.lang.fortran)
  • Re: What can be done in FORTRAN that cannot be done in C/C++?
    ... >programs in several languages, I cannot find a reason to keep ... >programming in FORTRAN. ... That's the only reason ...
    (comp.lang.fortran)
  • Re: A question on Newtons Method
    ... > archaic theoretical foundation, other languages aren't. ... Here's your reason, several million times over. ... repository there's every reason to join Fortran literacy club. ...
    (sci.math.num-analysis)
  • Re: About alternatives to Matlab
    ... Jon Harrop wrote: ... same performance in NumPy and Fortran 95 on this kind of task, ... array slicing in both languages. ...
    (comp.lang.python)