No, of course not; in fact, the examples of misuse I was referring to
were in neither field. I singled out these two for the remark about
automated data processing because they are both areas where the fitting
of and inference from models *in real time* is increasingly important.
It can be tempting to use robust estimators in such situations, since
human analysis is often not practical except maybe later as a
post-mortem - 'prowess' doesn't really enter into it.
- David R.