C++ programming

[Cristian Tibirna]

Le Dimanche, 20 Octobre 2002 7:24, Eray Ozkural a écrit:
>
> From experience, the next stage after that is tackling a really hard
> problem with C++, hating C++ together with its stupid bag of template
> tricks and then finding out that there are well designed languages like
> ocaml or haskell. But it doesn't take 9 years to figure it out, only 7. ;)
>
> To prevent a long rebuttal, here is what C++ is excellent for: projects
> like KDE.
>

And this brings about the most important argument. Perhaps the language itself 
is a secondary decision factor (at least as long as language knowledge is a 
secured asset). Much more important (just like in the case of KDE) are the 
libraries one has available. (A long discussion is possible about how a 
particular language eases the creation of high quality libraries through 
features that encourage development of generic programming interfaces). I 
grew more aware of this through the last year experience in our research 
group, when we evolved from pure MPI usage (*shudder*) to writing our own 
nice C++ interface to MPI that _completely hides the details of parallel 
programming from the end-user/researcher that finally writes the parallelized 
application code. It _is_ possible to make parallel programming easy 
(probably independent of the programming language).

The large problem remaining at hand is a fulfilling the need for larger 
collaboration in the "milieu". Things like PETSc 
(http://www-fp.mcs.anl.gov/petsc/) are obviously possible. Still, generally 
accepted equivalents for other domains are lacking (to my best knowledge as 
of today).

-- 
Cristian Tibirna				(1-418-) 656-2131 / 4340
  Laval University - Quebec, CAN ... http://www.giref.ulaval.ca/~ctibirna
  Research profesional at GIREF ... ctibirna at giref.ulaval.ca
  PhD Student - Chemical Engng ... tibirna at gch.ulaval.ca