Whoever designs the first easy-to-use version of this trip will unlock a massive amount of creative potential.
The spreadsheet was a major move forward for unlocking creative potential. A part of that probably brought us the financial crysis of 2007-2008, but nobody can complain that it wasn't creative :p
Also Excel clouds the ergonomics of the lists and functions by duplicating common functionality in bespoke UI and littering it everywhere so it's not a very natural path from using a function through the UI to using it in another function.
Also Excel columns add weird semantics by putting every column in the same rowspace.
Also arrays have names. Columns are usually just numbered.
In other words: yes, spreadsheets are the most useful end-user programming tool we have, but it has so much extra junk layered on top of it it's nigh unlearnable. And it's not a starting point, it's a dead end. An EUP environment built on a spreadsheet is a car built from horse parts.
A well designed array UI would have the semantics of an actual modern array. Push, pop, each, etc. A well designed function UI the same. And both need to be presented clearly without distraction to count as a "good UI" in my book.
I came to the conclusion that we need a completely new language, a new communication paradigm. We can't continue to build single-purpose apps for every single thing. We need a single interface that's general-purpose enough to let people communicate what would today take 1000 different apps (or 1 natural language).
I believe that the future of UI will be a mix of task management concepts (to-dos, task dependencies, task delegation, timelines) and binary interaction (Akinator, Tinder, Mechanical Turk). Add a graph knowledge base, a blockchain, a network of trust, smart contracts and you're set.
I've been trying to start a discussion about this for years, without success. If anyone is interested in discussing this space, I encourage you to let me know.
I'm thinking about a completely new communication paradigm, one that's short of a brain interface.
p.s. Natural language is still in it's early stages. Think of the early (non UI) PCs. That is, over time voice as an interface will get better, and better. It will get there. Just not yet.
Wild experimentation always washes off after some standard gets any big adoption. So, it's very natural that our interfaces are evolving slowly and iteratively. It's expected as otherwise things would break a lot.
Yet, it's unimaginable how far we are from every concept in a interface being free to interact with every other concept on the ways that make sense. Current GUIs are incredibly brittle and underpowered. Why are we not getting a big revolution there?
You might counter that this is the domain of MIT Media Lab rather than Microsoft and Google, but at this stage new UI concepts have to be adopted by entities with significant influence to be adopted at all.