Files and directories are just one of many possible abstractions for storing data. You have files and directories. Directories contain files and directories. Your whole device is one big directory. Files are identified by name. There's absolutely no reason to think this is the best possible model.
Here's another: Your device contains a bunch of photos and videos. Photos are identified by looking at them. Videos are identified by a still image, or by playing the video within the video chooser.
Here's another: Your device contains a bunch of apps. Apps can organize their own data however they see fit.
... Microsoft's OLE really was the most well-integrated document-centric desktop we ever got, wasn't it?
... OLE is one of those things that have been forgotten under POSIX hegemony.
What people remember of the Office '95 file formats was that they were 'proprietary' but it was worse than that in the sense that these were really meta file-formats that would let you embed a 'file' inside a file which would be handled by some arbitrary EXE, so to process any Office '95 file you really need a Windows installation with all the EXEs and DLLs required by that file, thus when Microsoft came out with ActiveX [1] I was really terrified that you'd need a Windows computer to browse the web in a year or two.
I find the 'app' model to be user-disempowering thing since the app chooses what you can do with the data, but 'photos' and 'videos' are something really different because there is a standard format for those so I can take photos with my DSLR or my tablet or make images with a Python program or edit them with Photoshop or MS Paint or the GIMP and view them with different apps, print them out, upload to a web server, etc. A system could provide a different API for accessing this kind of functionality but I'd say it is fundamentally a step back to have to read data with the same program you wrote it with.
Here's another: Your device contains a bunch of photos and videos. Photos are identified by looking at them. Videos are identified by a still image, or by playing the video within the video chooser.
Here's another: Your device contains a bunch of apps. Apps can organize their own data however they see fit.
... Microsoft's OLE really was the most well-integrated document-centric desktop we ever got, wasn't it?