[pilgrim0]

> I found it took a while to get to the point.

You're just a great understandeour

> I agree that adding an "array" style of directory to our file systems would be really cool.

I think this is sort of a low-hanging fruit people have slept on. We've proved list-based systems are extremely versatile for data structures (s-expr) and programming (lisps). What about for media in general? Everywhere I look I just see lists, with very minor stylistic distinctions between them. Of course there're abysmal infrastructural differences between chats, feeds and what not, but it does not invalidate a universal list-based frontend, similar to what you developed in your comment.

[andyferris]

True.

I didn't really think about (linked) lists much, only (flat) arrays, but maybe that's possible already with files laid out like:

/head /next/head /next/next/head ... etc

or whatever.

I'm not saying its _ergonomic_, mind you. I'd like my file viewer to lay out the list flat, for example. There might be technical limitains with the length of the file path. Etc.

[syberant]

I'd argue that this concept for media is most commonly known as "playlists" and unfortunately only used within data silos, e.g. a playlist of YouTube videos, a Spotify playlist, a TV season of episodes, a series of episodes, a trilogy, etc. (Yes, I'd argue that your music library of mp3 files is also kind of a silo, although portable) Heck, even a slideshow is arguably a playlist.

I agree that putting a playlist-like concept into, say, the filesystem would be an extremely interesting idea but I think a big danger is running into the same problem as hardlinks and symlinks. This problem is that if a file is "present" in multiple places (or playlists) deleting/modifying/moving it can have unforeseen consequences and it's hard to reason about (and if you copy the file now you get to invent a way to track different versions too!). I think this is also holding tagging filesystems back.

I'm currently writing a non-hierarchical FUSE filesystem and have been thinking about this list-directory concept but I'm still not completely sure how it would work, especially since I need to remain backwards compatible with the POSIX interfaces. Will probably have to just try it out (xattrs to the rescue?) and see what sticks I suppose...

A linkdump of interesting somewhat related stuff:

- https://newsletter.squishy.computer/p/knowledge-structures

- https://newsletter.squishy.computer/p/all-you-need-is-links

- https://thesephist.com/posts/search-vs-nav/

- https://karl-voit.at/2017/02/10/evolution-of-systems/ Especially "Information-Centric Systems"

- https://www.theatlantic.com/magazine/archive/1945/07/as-we-m... A classic 1945 article cited as inspiration by Ted Nelson, Doug Engelbart and Tim Berners-Lee.

- https://en.wikipedia.org/wiki/Content-addressable_storage

- https://www.nayuki.io/page/designing-better-file-organizatio...

[andyferris]

Good observation. If ordered data is only present in silos, maybe we should work towards the “narrow waste” for such data, to make it less siloed. :)

Surely we can discuss (computer programs can interoperate on) something ordered other than raw bytes (ie the contents of a file).

xattrs I suppose would be the best way to do something backward compatible yet descriptive enough to be useful? I agree it’s a steep hill to climb.