What do you mean exactly by "What is your opinion about the effect of simulation software on reducing the programming skills of researchers?"

Are you suggesting that the other way, i.e. not using simulation software, i.e. having each researcher code their own fluids mechanics software be better?

I think you must consider aspects like:

With all these points, you might get the idea that I don't think programming is important in your field. Quite the opposite!  It's just that programming skills development is quite a different issue than should simulation software be use or should researchers "roll their own"!

The best digital artists use Photoshop but they don't code their own image processing software... That is not to say that they won't build plugins or scripts to extend their main tool capabilities, but they are not reinventing the wheels when a really good one already exists.

I hope I understood your question correctly?

Regards,

Bruno

P.S.: What kind of simulation software are you referring to exactly?  Do you have some specific software in mind?

http://eigen.tuxfamily.org/index.php?title=Main_Page

http://www.cgal.org/

http://web.stanford.edu/~vcs/AAAS2011/1102_aaas_reproducibility_fperez.pdf

A book you might be interested in:

https://www.amazon.com/Software-Engineering-Science-Chapman-Computational/dp/1498743854

Dear Khaled, I was doing some programming many years ago when it was necessary to write code in FORTRAN myself. We studied also numerical methods and knew what amendments to program are possible if it does not work. People that time also typically knew equations and even their derivation.There was not a big problem to model the solution of system of equations in some rectangular domain. Problem was to model real geophysical shapes, like using a grid with ocean bathymetry for the study of propagation of tsunami waves. But it took time to catch and correct errors in FORTRAN.

Today I am a bit away from this field, but I know that people from geophysics still are able to write code for complex surfaces, simulating for example the dynamics of flood.

Modern software (like Mathematica) uses some standard routines and often gives good graphical output. However, some math problems can still remain for complex tasks when function involves discontinuity or singularity. Sometimes they are solved by hand, but it is bad that we cannot modify code when necessary.

I think that general mathematical and physical education of modern programmers became worse. I can track it from some questions on RG. There is too much specialization and people doing programming often do not understand the context of equation and how it was derived.

Based on my experience as applied mathematician teaching to civil and environmental engineers, my feeling is that the main problem with simulation software is that it develops in students and other users without sufficient background or critical attitude the false belief that the software results are error free. In some areas of research and model application, doing your own coding is no longer the best option, as suggested by some of the previous comments. But understanding correctly what the model does and to which extent it can be considered reliable or not should still be a top priority with students and researchers in their training phase. Often, for a number of reasons, I am afraid this is not the case.

I agree with the concern of M. Bonaventura but that's not the whole story. 

One thing we must not forget though is that knowing the limits and uncertainty of your model is also true if you code your own stuff.  In that respect, it can be seen as an "orthogonal" concern as to wether or not you should code your fluids mechanics models from scratch instead of using an existing, maintain and documented framework.

I am doing the simulation using own house code. But still, I am not reluctant to use a fancy software. I believe that people who were using the basic code in their study for years, can easlity accelerate their work using "simulation software".

Best

Nisat

It is a difficult question because it is very problem dependent. But in general what is the point in reinventing the wheel? 

I did a Masters in scientific software engineering, a PhD and post doc work that involved a lot of coding (Fortran), and now I am a full time CFD engineer. I do like coding, but ultimately the only things I really miss is being able to jump to exact routines in the code to write out what is going on, and to have better control over output. I work with commercial software and the only coding that comes up is writing user functions.

This means I have a lot more time to think about the physics rather than the implementation

Dear Colleagues,

    Thanks for you valuable opinions and experience that appear in the discussion. My question doesn't mean that I am against using simulation software, but seeks about suitable answers to many questions like when should we use these software? and who should use this software? Is it suitable for student level study? should we infinitely trust on their results with knowing the validity of using its packages? in brief, I thick a researcher should have enough programming skills before using these software.

Regards

Khaled. 

It had destructive effect on imagination of any researcher. Our mind be week as calculators made our engineering ideas weak and weaker. Anywhere it has specific advantages and so dangerous!

I do not agree with M. Fanayi.  Many researchers can be restrained because they lack the advanced programming skills or time to implement their ideas from the ground up.  Using advanced simulation software like Matlab, Maya, Mathematica, etc. can save HUGE amount of implementation time for results not 50% as efficient and let the researcher concentrate in its own field of expertise... which is most frequently not High Performance Computing...

The problem is not in the CFD software but in the users that use that... Too often I read questions (for example, have a look to the question in the cfd-online forum) about diverging solutions, wrong setting of the BC.s, wrong type of formulations, all of them depending on the fact that simply the users have no background in fluid dynamics, numerical analysis, physics, mathematics and CFD. The problem is the a CFD software is not a TV you turn on and see...

And, let me say, I have to admit that the academic education has his own responsibility in this lack ...