I thought it was a pretty good article. Having taught math for a mere 30 years, I know I don't have all the answers, but I always found that just teaching procedures and "covering material" to be just about utterly ineffective; engaging kids inn the topic at hand, and getting them to understand what they were doing was more effective. Some memorization is needed; but people, like other creatures, construct their own memories and understandings in rather idiosyncratic ways. Getting them to play with the math helps.
I was taught some of the New Math aka SMSG and some of the integrated chemistry and physics during the 1960s, and actually enjoyed it. The old-fashioned math of the day bored me to tears. I was one that also thought that the methods described in the NCTM documents of around 1989 were pretty darned good, and tyried to put them into practice, occasionally with success. It seemed to me that some of my colleagues tended to think otherwise, and much preferred the old-fashioned Stein, or Saxon Math approaches. Whenever I tried those approaches, they weren't successful.
College physics departments ARE doing scientific studies of what works, and they find the lecture-practice-lab (similar to the "I, you") model does NOT lead to actual understanding. Richard Hake has written on this ad nauseam, buty accurately afaict.