@innocentzero Back in the 1970s, “everything is a file” was a pretty revolutionary idea. It allowed you to configure hardware and services using ordinary text editing tools, and treat text as objects on a map (the filesystem tree), whereas before you would have had to write numbers directly to certain locations in memory.
It is still a somewhat useful idea nowadays, it is just that there are actually better ways to do things. For example, relational algebra like what SQL databases provide, or object oriented systems like what Smalltalk systems provide (I think IBM’s OS/2 was an attempt at making an “everything is an object” operating system), or everything is an environment, like what Lisp systems provide.
And it is a lie, actually, not everything is a file. For example, you can’t change the list of processes that are running by just editing a file, you can’t rearrange a network of interacting processes by editing a file — you can write software that does this (Guix Shepherd), but the operating system itself will not without a lot of extra software installed.
For better or worse (mostly worse) the Unix philosophy won, and the whole computer industry adopted the “everything is a file (mostly)” approach. There has been much dispute as to why, but few question it nowadays.
The Unix way is not the end-all-be-all of computing paradigms, anyone who thinks so is probably not all that well informed. https://tilde.town/~ramin_hal9001/articles/emacs-unix-04_lisp-does-fp-better-than-bash.html