Re: Why is Ada considered "too specialized" for scientific use

[Georg Bauhaus <rm-host.bauhaus@maps.futureapps.de>]

On 4/4/10 6:46 AM, Nasser M. Abbasi wrote:
> I was browsing the net for scientific software written in Ada, and came
> across this strange statement:
>
> http://farside.ph.utexas.edu/teaching/329/lectures/node7.html
>
> "Scientific programming languages
> What is the best high-level language to use for scientific programming?
> This, unfortunately, is a highly contentious question. Over the years,
> literally hundreds of high-level languages have been developed. However, few
> have stood the test of time. Many languages (e.g., Algol, Pascal, Haskell)
> can be dismissed as ephemeral computer science fads. Others (e.g., Cobol,
> Lisp, Ada) are too specialized to adapt for scientific use.
>
> ......
>
> The remaining options are FORTRAN 77 and C. I have chosen to use C "
>
> I find this strange, because I think Ada can be the best programming
> language for numerical work.  So, I do not know why the author above thinks
> Ada is "too specialized to adapt for scientific use".  Is there something in
> Ada which makes it hard to use for  scientific  programming?

The text, by Professor Fitzpatrick, Physics, 2006, appears to be an
fine example of quite useful justification rhetoric.
(Of the kind well explained by, e.g. Leon Festinger or Irving Goffman.)
It is full of false facts that in general would offer
to an author the comforts of blissful ignorance and a convincing
appearance at the same time. (Who has not been in this situation?
Assuming that a quick proactive defence of your standing is
more tempting than the alternative: is to shut up?)


C99 (note the year) has complex types,  says C hasn't.  Well, it hadn't, 
as some point in the last century.

Inexpensive compilers for Fortran 90 were available that year (2006),
AFAIR, from Intel or the FSF.  NAG has academic pricing for its
Fortran compilers, at least now, possibly earlier.  There are more.

The info about C++ is rather dated in 2006---templates and
interfaces would be closer to its focus, I'd think. The remark
seems to draw its persuasive power from a brevity that only
repeating hearsay can offer, as someone noted.


During the last few months I had a chance of hearing about Fortran
versus Ada in scientific computing where the subject is concurrent
execution on many processors, with some communication.  It turns
out that there is no reason to dismiss either Ada or Fortran,
judging by the results: tasking can be as good as MPI.
There is, however, reason to believe that OpenMP does not scale
well.  (From a superficial glance at OpenMP 3.0, I see so many
words sounding familiar in an Ada context (task, barrier, shared,
parallel regions, ...). Are they performing mostly the same experiments
that, I think, were done in the 1970s? I'd speculate that the
parallel constructions aren't novelties in an HPF world, either?)

As Dmitry Kazakov has recently said, when Ada run-time systems
starts addressing the properties of multicore hardware
there is hope that it could really shine: Not just because concurrent
sequential processes are so simple to express using Ada tasks
---and you'd be using only language, not a mix of libraries,
preprocessors, specialized compilers, programming conventions,
etc.  But also in case the fine grained complexity of OpenMP 3.0
can be bridled by simple language and a good run-time system.
At little cost.