"The experts don't care and the beginners don't know." I cannot agree with this more. And while he seems to have come to the conclusion that you can't get experts to change because they haven't got a problem they don't already know how to solve (even if suboptimally), and you can give non experts a tool they didn't know they wanted, he still wants to generate grand unified theories? You can't introduce generalists to it, because they need to know the mainstream. You can't introduce non-experts to it, because they can't use it for anything.

And yet it moves

Just got around to listening to this conversation (rant: why do podcasts have be videos today?). Lots of very good points being addressed, and what I have seen of @jonathoda's work makes a lot more sense to me now. The problem of improving practices missing from the agenda of every institution that could care about it is not unique to programming. Research in physics, chemistry, and biology is affected by this as well today. The only improvement people care about is for doing more quicker. That's why have problems like the replication crisis. I tend to attribute this to endemic short-term thinking. I love the idea of (1) focusing on data first and in particular (2) version control for data. The expert vs. beginner problem exists in some form for all technologies. What's missing for programming is a feedback loop that motivates experts to care about beginner's issues. Education fills this role in other fields. That's one reason why universities have combined research with teaching for centuries. Why does mechanism fail for programming?