Ian Chivers wrote:
>
> I am interested in getting some information regarding the actual
> take up of f90.
>
> my two main interests are in
>
> 1. people actually teaching with f90 in preference
> to other languages (not just f77 though)
>
> 2. people who have actually moved production code
> over and now program in f90.
>
Within my Group we teach 4 MSc courses with programming content:
Numerical methods & Software Systems
Scientific Applications Software
Mathematical Modelling for the Physical Sciences
Defence Simulation & Modelling
to a mixture of full and part-time students. Our courses are modular
and in assessed work for the modules students are required to write
code, use packages, do analytic work and essay work.
Our full time students are taught Fortran 77, Fortran 90 and C to a
level at which they can understand and modify other people's code. They
are expected to reach a level of fluency in one of the languages so they
can write substantial code themselves.
Our recent experience is that students who arrive knowing C stick to C;
those with lots of (job wise) motivation learn C often on the coat-tails
of the experienced C programmers; and those who want to just learn how
to program use Fortran 90. It should be said that we do have more
experience within our Group in Fortran and if students are struggling
they can get more help using this language than C (although this is
changing).
Those students wanting to learn how to program find, I believe,
Fortran easier to read, automatic and allocatable arrays a blessing.
That said they tend to program as Fortran 77 programmers with a little
bit of F90 as opposed to using powerful modular features of F90.
Research wise I have written a small CFD code in Fortran 90. I am
currently taking advantage of operator overloading to easily implement
automatic differentiation so as to calculate Jacobians for a Newton
Solver. One of our PhD students Neil Stringfellow is writing Boundary
element software in F90 and using PVM/MPI for parallelisation. I think
he would be interested to hear from people with F90 interfaces to these
libraries,
[log in to unmask]
We also have other numerical work in mesh adaptation (mesh movement)
using Fortran 90.
Hope this of help
Shaun Forth
--
####################################################################
Dr Shaun Forth
Applied Mathematics & Operational Research
Cranfield University
Royal Military College of Science
Swindon SN6 8LA
England
tel: +44 (0)1793 785311
fax: +44 (0)1793 784196
email: [log in to unmask]
http://www.rmcs.cranfield.ac.uk/~amorg
####################################################################
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|