>I do design research. My papers always provide principles with examples and evidence. From what I remember of studying Human Computer Interaction (HCI) as an

>I do design research. My papers always provide principles with examples and evidence. From what I remember of studying Human Computer Interaction (HCI) as an undergrad in the 90s, the evidence was derived from learnability studies. This meant that research favoured point and click interfaces (which are the equivalent of caveperson grunting, IMO) over command line interfaces which allow much deeper expressivity. There was basically no affordance to systems that people use day-in day-out and are happy with the trade offs in the learning curve (vi, emacs, blender, lighting desks, audio controllers, Bloomberg terminal). I like that your example paper is about trying out different interactions in a small part of a daily-driver piece of software. But it's still trying a 'new' thing for some of the experimental users which means learnability is core. How do you manage the balance?