Yes...teaching regexes is one of my goals...since the beginning I've forced them to jump through the hoops of setting up Github and submitting Markdown files, to get used to the idea of dealing with plain text (which of course, is key for understanding CSVs)...so regexes are a natural thing to learn for finding patterns and, on a day to day basis, cleaning data...I think in terms of technical skills, it's probably the most useful thing I could teach given how much time is ultimately spent on data munging.
I haven't looked close enough at what you are doing or what they are doing to pretend to make a comparison, but I guess it isn't too risky to say that they are at least thinking about some of the same things you are.
Yes I have...they're hitting a lot of what I want to do in a class for next quarter, which will be heavily command-line focused. Besides using some of their lessons, hopefully I'll come up with a few of my own to contribute to the project. But I do agree that both scientists and journalists could benefit a lot from learning how to work with their computers at a lower-level.
http://regex101.com/ is (I think) a really nice way to learn/use regex. It explains each piece, displays capture groups from the regex applied to sample data, and has built-in documentation. Simple example: http://regex101.com/r/nB1bR1/1
