Hacker News new | past | comments | ask | show | jobs | submit login

Here's a thought:

This article is exploring the wrong problem.

The filesystem where my files reside should already be a fully-distributed filesystem covering all the devices I own, right from the moment I buy and enroll them.

With that precondition, my file is already where I need, and any physical bit shuffling is just a detail managed by the OS, using whatever connectivity is already available (or prompting me to "please connect cable type X between devices Y and Z for N minutes" if devices are out of sync).




Isn't that exactly the model of Gdrive/OneDrive/DropBox etc? Except they might not do any LAN transfer optimizations, but that is minor implementation detail

edit: apparently at least Dropbox has had LAN sync for nearly a decade: https://dropbox.tech/infrastructure/inside-lan-sync


I don't think those come even close. Some blocking considerations:

- They are a bolted-on "magical part" of your filesystem. If you manage your files anywhere outside of the magical field of synchronization, they're invisible (e.g if your file is in the "Downloads" folder, but that folder is not synchronized by Dropbox, then you won't find it from any other device.)

- They are not distributed filesystems. They are a central filesystem (on their premises, outside your reach) and a smart replication strategy. You don't get a say on data sovereignty, and your data lives on rented disk space.

- Trying to operate on your files is dependent on which device you're using. If you're on a computer, you get first-class local files in first-class directory trees, that you can manage with the usual semantics, provided you have enough free space in the physical disk of that device. If you're on a phone, you download copies of files (and pray they get uploaded if you change them in phone), and almost no way to manage the directory tree aside from the application itself.


>The filesystem where my files reside should already be a fully-distributed filesystem covering all the devices I own

I'd rather not set up the possibility for a single software bug to destroy all my data at once. Even if you have good backups, there could be subtle data corruption that you don't notice until it's too late.


That's a very hard problem.

Have you found a solution yet?


What you describe is called Google Drive/OneDrive/iCloud.


I don't think that's true.

Let's take a real life example: My Android GPS (OsmAnd) records GPX files on my phone storage when I'm out on a trek.

How should I configure Google Drive/OneDrive/iCloud to expose those files transparently in my computer (a linux box at home) as they get created?

And more importantly, why should I have to configure anything at all? The idea I'm advocating for is that those are my files, and so they should be in my (distributed) personal filesystem by default without me having to set anything up.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: