I have played around with a language dubbed WWT (for "World Wide Types") https://jtree.treenotation.org/designer/#standard%20wwt). Instead of starting with the type system and working my way toward the user have gone the other way starting with the "what would the simplest language look like for building types" and worked my way toward the type system (just compiles to TypeScript for now). My thought is it should be editable in Excel.
What I think needs to happen for TypeDefs (or similar) to nail this, is to create a type language that solves the "last mile" problem. Every organization in the world has their own perspective on things, and will need to extend/modify some global CommonTypes repository. So a great type language would be something that worldwide CommonTypes could be defined in, but that is easy for a dev at the local real estate agent's office to write some code and extend. An "HTML for Types" in that sense. Just a guess.
I have a prediction for this space.
I think we will see a DSL for type defs like typedefs.com (but not quite typedefs.com—too complex), arise that becomes really popular and becomes the "English" of the Standards Organizations worlds.
Sort of like DefinitelyTyped or Schema.org but 1,000x+ bigger.
It won't be because everyone will finally come together and agree on a language for types. That would suffer from the https://xkcd.com/927/ problem.
Rather someone will discover some simple and beautiful way to define types that has significant technical advantages to existing standards and will spread like wildfire.