I can't really tell how do we apply make this idea compatible with the common workflow/model of computing. In my imaginary world everything down to the level of a statement of what is a code file in the real world is a separate record in a relational database, there are no file names and no folders (only tags, commentaries and other metadata).

In the real world folder hierarchy is probably inevitable so yes, it has to be stored in a way (in a separate table perhaps so we can have many-to-many relationships sort of what hardlinks do).

> In my imaginary world everything down to the level of a statement of what is a code file in the real world is a separate record in a relational database,

I think you'd like prolog https://swish.swi-prolog.org/example/examples.swinb

It's cool, indeed. Nevertheless, besides files which contain sets of logical statements/data, there are pictures, audio and video recordings which are binary by nature. Does Prolog support this kind of data types?

Yes, and you can see a few hundred libraries here https://www.swi-prolog.org/pack/list

I’m still with you on the complete reimagining of how a computer stores and accesses ‘code files’, but there’s still a challenge with ‘small files’, which is to say the hundreds or thousands of chunks of code that can have dozens of uses (Plex’s metadata cache, for example, or the hundreds of shared libraries that allow larger more complex programs to be built small by referencing them).

One benefit of folder hierarchies (and I don’t think there are many) is that it’s easy to delete/move hundreds of thousands of tiny bits of associated data just by deleting/moving the parent.

I mean using tags instead of folders. You can make a tag that may mean anything and select all files with a particular tag.