Zed A. Shaw, Painter

I train Rust since 4 years (Meetup, Hack & Learn and actually as a paid trainer for 3 years now) and I _hate_ it when people bring up "learnability" in language RFCs. Most of the time, it's a bit of anecdata, scrapped together.

If we want to make learnability of Rust better, the first thing is a structured assessment of what problems are actually learning problems. It's not like I want to say that I know all about it, but I _do_ believe that we, as a project, don't know enough about it.

Given that I never learned teaching, though, I feel ill equipped to do those structured assessments.

What we end up with, is "learnability" being the argument that everyone considers valid for _their case_, because there's not much meat to it and it's hard to argue away from.

I'm pretty sure @zedshaw gave a talk about how most "learnability" arguments are null and void, because they are basically just individual appointments, but currently can't find it.

I used to speak about it a lot but nothing formal, mostly tweets. However, I think this article hits a lot of the points I frequently make: https://fredrikdeboer.com/2017/03/29/why-selection-bias-is-the-most-powerful-force-in-education/ … Most education "studies" are after the fact analyses of selections of students who didn't need the education.