I was listening to some of the old episodes of the podcast and have been thinking about projectional editing in general. I'm not sure if this is a novel insight or not, but there seems to be some form of equivalence between reading/writing code textually vs with blocks/shapes and the equivalence of algebra/geometry. There are many proofs and problems that seem to lend themselves to geometry over algebra but there are also situations that being able to reason about something algebraically makes it substantially simpler. I think this is the situation where you need both and it's possible at certain levels of abstraction you would want to be able to easily shift back and forth. For example I could imagine a system that had components built in a textual representation but then the component linking all done utilizing shapes.

Makes me think of this scene Then Will says "Here, for you code block / linear proof people, I wrote it down"

I completely forgot about that scene but that's exactly what I'm thinking about.

Really it's what MVC was originally meant for (one of Trygve's intentions) of having multiple possible views for a model, letting the user select the view that most directly mapped to how they think and work - https://folk.universitetetioslo.no/trygver/themes/mvc/mvc-index.html I love how tools like ZenKit allow the same data table to be represented as a list, a table, a kanban board, even a mindmap and calendar

my favorite part of the mother of all demos is when the shopping list becomes the map of your commute home via all the stores for the shopping list

Blocks/shapes are just one projection that a projectional editor could choose. The good stuff is having multiple possible projections, as per what @Chris G mentions. See also Glamorous Toolkit.