Great read. Found this particularly interesting wrt this community:

In software construction, we often start by reinventing the form and, consequently, we have to face a very wide range of design problems. Are there cases of software construction that are more akin to the vernacular or unselfconscious design? One possible area of interest might be how people solve problems in spreadsheet systems like Excel. Spreadsheets define a relatively fixed form and allow the user to focus on skilful solutions to specific problems.

Make me wonder if Excel is special, not so much because of its spreadsheet properties, but just because it’s constrained. Perhaps we can “discover” other excels simply by exploring other constrained forms of programming, rather than assuming Excel “gets it right” in some way.