There's an interesting discussion going on at the Research Software
Engineers mailing list.
Sadly you have to sign yourself up (for free) to see the archive, but
the thread has many interesting insights and experiences:
Sign up: https://www.ease.ed.ac.uk/friend
Read thread (after signing up): https://mlist.is.ed.ac.uk/lists/arc/researchsoftwareengineers/2016-09/msg00007.html
I'm afraid I'm quoting myself here, but I think the point is good:
"If you take your work seriously, think collaboration is important,
and
have some exposure to practices that encourage correct and
readable
code, there's a decent chance that you will make yourself
unemployable
as a scientist pretty quickly."
The Google+ URL for this post was
https://plus.google.com/+MatthewBrett/posts/A6BSbJDn4we