Mods and friends
I am myself a PhD candidate, and I am now preparing and studying for my thesis which is about the 4D space-time finite element method for hyperbolic conservation law. Actually I am a bit confused and a little worried about my future. Here is a list of questions for becoming a computational scientist/numerical analyst I am most concerned with:
What projects/topics we should choose for a career path we would like to have? Should we take a more utilitarian approach to choose relatively simple projects that could lead to some average quality publications? or should we tackle an important yet difficult question that has lots of application in physics/engineering but you would undergo a several-year studying process.
What kinds of skills should we learn for future career? For the example of programming, the academia mostly adopts MATLAB/Mathematica, Python/C(FEnICS,METIS), Fortran(LAPACK), however, learning from the discussion with my friends in various industries, VBA/C++/C#/Perl etc are more popular due to their graphical UI compatibility. Which path should we choose?
If we would like to pursue a position in the academia, like PostDocs, should we focus on more of a theoretical analysis aspect of scientific computing? or more of the computational results? How about the positions in some research facilities like national labs?
If the mods think only part of these questions can be asked in Scicomp.SE, I will modify and repost in the main forum.