Example: Download a PDF of a book. Read it. Notes derive from that. Those notes can concern a diary, quote for some other work, inspiration for art, language learning, etc. The book itself can contain text, data entries, tabular, photo, etc.
There is a multitude of possibilities to find categories for every aspect of this 'idea'. The goal is to find a (not 'the') abstraction to handle data so it is possible to work with it. As someone already mentioned, even if such a system can be found, it is unclear if this would make it easier to work with all it contains, than how we do it conventionally.
Addit: DevonThink comes closest to what I described.
How does this differ from other note taking apps like Obsidian?
What this leads to ('what I imagine'), is a modular system. Very bare bone in its core form. Data is processed with editors for each data type. Editors come in the form of addons to the system. Data (and meta, params, etc) would always have to remain human readable - comparable to MarkDown or something. As you must've figured by now, I don't have a definitive picture in my head. All of this must sound very delusional or dreamy ... :)