On the one hand it is wonderful that students can use commercial software and freeware to quickly plot graphs, do statistical analysis, do Monte Carlo simulations, perform quantum chemistry calculations, ....
In principle, this should lead to deeper and more sophisticated analysis and more time for reflection.
However, it seems to me sometimes the opposite is happening.
I increasingly encounter research students who seem to treat software as a magical "black box" and lack a basic understanding of the
- physical principles and equations the software is based on
- limitations and reliability of the software and the underlying equations
- need to perform systematic checks and comparisons with analytical results in well understood limits
I once heard of a respected condensed matter theorist who made all his students write their own software from scratch. Perhaps, this is a bit extreme but I can see why he did it.
So should there be limitations on how much software undergraduate students use in research? Should they have to pass a test before they get a "drivers license"? What might that test look like?
I agree. I think that care needs to be taken with the formulation of the project. For a PhD, some significant insight must be gained, and today it is not sufficient to just do a basket of calcs and write up results. This may be adequate for undergraduate research or maybe honours (for a broader target question).
ReplyDeleteIt is extremely important for the advisor to know when enough calculations have been done and when the time has come to think and analyze. Often, the student will get swept up with what they can calculate, and won't recognize this. I have seen this firsthand, and you have to be direct about it. I have also seen advisors fail to realize when this point has passed, with bad consequences.