The alternative (at least in the extreme) is bad as well --- scientists who can't program anything and are rely on pre-canned point and click analyses and progr

The alternative (at least in the extreme) is bad as well --- scientists who can't program anything and are rely on pre-canned point and click analyses and programs. There isn't a button for X and so I can't do X. I pressed the Y button and so of course my Y analysis is valid. Etc. Edit: you think this is hyperbole but I've heard the above two statements more than you might guess, coming out of the mouths of prominent scientists.