anarcat

right, those are the two ways.

and how does niri do it? a workspace is shared among all monitor, or it's one workspace per monitor?

can you expand on that? how does it compare to sway, for example? what's unacceptable and acceptable for you?

I think they built it. :)

I had a short moment where i felt "omg i can fit a framework board in there", but alas the 1300 is only 9" wide. Still, I really like this idea, and I wonder if i might not just make exactly such a thing but with a larger pelican case. Then you can cram a lot of data in there with a bunch of modern 3.5" spinning disks (which, hopefully, the pelican will keep from rusting too much).

Beautiful idea, would love to see open specs, but really nice to be sharing!

(Update: the framework board is so close. The first daughtboard in the schematics <https://github.com/FrameworkComputer/Framework-Laptop-13/blo...> is actually 9.071 inches (115 + 115.4mm, so 230.4mm if i read it right) wide! But alas the mainboard is 274mm (10.8 inches), and that is without the expansion cards added. With the cards, we're talking about 11.7 inches (297mm) and then you still need to plug in things in there! Still there's something there, i think.)

I just tried the free trial (1GB, 1 year, nice and thanks!) and this is pretty impressive software. I like the mobile app: it's snappy, and Just Works. Obviously, I couldn't fully test it because I have many gigabytes of photos on my phone (which is a problem I'm trying to fix!), at least not without going out of the free trial range. Also, uploading all that stuff takes time, and I'm not sure I want to share all that personal stuff with you fine folks, even though you seem to have your house in order.

Which brings us to self-hosting, of course: really nice that the server is open source! I found ente first through f-droid when the app landed there but put it aside because the server was closed-source then. But wow, really nice design! I like the docker-compose just shoves a minio in there, really neat and probably how I would build something like that myself if i would start from scratch.

Compared to photoprism and immich, for sure server-side machine-learning is missing, but then that's obviously a tradeoff you must live with if you want E2EE.

As I mentioned in the goodbye note, I won't be using this short term because I do need to have something on my desktop I sync photos with that's not a bulging pile of Chrome (AKA "electron"). I really appreciate you spent all that energy writing those apps, but I really need something more lightweight on the desktop.

Right now I'm syncing photos with git-annex, and I wonder if ente could be a "special remote" there, even, but for now this is not really compatible with my workflow.

But congrats on this tool, it looks really nice and I'm likely going to recommend this to friends and family as a hosted solution.

The post says:

    Edit: Case in point: this post was quietly removed by Hacker News moderators within 40 minutes of its submission.

Is there any transparency with regards to moderation here, now that I think of it?

but see that's the thing, i don't even need that silly thing. i just run syncthing everywhere I need to (and it pretty much runs everywhere, including the kobo), and that's it. no middleman required (other than the syncthing meet-me infra of course, which is not as much as i expected: https://relays.syncthing.net/).

This is actually what I want Calibre the most for. It sorts things in folders quite nicely.

I am one of those weirdos that really appreciated iTunes setting up my music by Artist/Album/Track.mp3 format back in the days... until I realized how much badly-tagged garbage I had in there at which point, yeah, I was upset. :p

But nowadays I just have books setup by Author/Book/book.epub and that ... works okay. I mean it's one way to sort things around. I am actually considering a fully flat structure now so that koreader can show me previews, and I think Calibre could allow that, as I can customize its naming patterns...

But yeah, it can be kind of annoying for software to rename files for you especially if you're not starting from scratch. But Calibre is the first thing I used to sync books on my ebook reader (a sony PRS-T2 i think!) so i did start from scratch and just kept going, so i didn't get many bad surprises, and it in fact forced me to keep my metadata clean to a certain extent (otherwise books end up in the wrong folder).

wow. that is courageous!

looks like an interesting project... but I must admit it's not really something I use or need anymore. I used to do some book conversions when I would only find (say) a `azw` or a `mobi` file and would need to convert to `epub` but these days (a) I just find a `epub` but even if not (b) koreader can read `mobi` files and a lot more i can throw at it and (c) `azw` files are basically unconvertible, if I followed that correctly.

Speaking of which, do you do the DeDRM dance?

thanks!

interesting project! added, thanks!

Author here, fancy seeing this here (again: https://news.ycombinator.com/item?id=21190657)! I keep updating that article as I find things, the most progress there is is in ebook readers. I mostly use Koreader on a Kobo now, but they also have an Android app which I used when I had a tablet (which has been unfortunately bricked recently). This meshes well with Syncthing as I have read progress propagate everywhere, and I use git-annex for archival / backup / redundancy.

I still use Calibre for importing books. I think I could get rid of it if I had Something that would:

1. take a book and search for missing metadata (e.g. publication date, author, title, series, publisher, cover)

2. record addition date (git-annex or filesystem m/ctime could be used for that)

3. ... well, that really is it

Really, I don't use Calibre for anything else anymore. In the article I talk about the book browser, but I don't really use that anyway. I'm actually considering "flattening" my archive so that all books are in the same directory. That would make Koreader possibly more usable as it would directly show book covers while browsing around. It might also make navigating through a file manager better for exactly the same reason.

More broadly, I wish we had better file managers on the Linux desktop. What I'm looking for, repeatedly, is something that would show me a preview of what's inside a folder, kind of a photo gallery / music browser / ebook browser all rolled into one. The generic idea is that if there's a `cover.jpg` or whatever it's named inside a folder, use that as a icon or at least some sort of overlay when displaying the folder. I'm not sure how that would work, I'm not a designer, but I can't help but think there Must Be A Better Way here.

I use syncthing to sync my library around now, as mentioned in the article (I believe). This works great for Koreader, which stores read stats in a plain text file next to the ebook file. That can lead to syncthing conflicts sometimes, unfortunately but I suspect that might be more because of my combined use of git-annex and syncthing to archive my collection than anything specific to syncthing.

It does mean I need to use koreader everywhere though, I don't have a way to sync read progress back into Calibre, but then I don't really have read progress there anyways. I have a "read/not read" flag there, but it's some custom thing I added somehow, Calibre doesn't support that out of the box anyway.

their servers are overloaded, but you can follow the news here:

https://twitter.com/FrameworkPuter

or here:

https://fosstodon.org/@frameworkcomputer

... depending on your social tech.

the actual post is also on archive.is now:

A pretty bad snapshot resides at:

https://archive.ph/fcY1H

a somewhat working one is at IA:

https://web.archive.org/web/20230718144114/https://frame.wor...

dupe of https://news.ycombinator.com/item?id=36773107

dupe of https://news.ycombinator.com/item?id=36773107

dupe of https://news.ycombinator.com/item?id=36773107

I did, and Marty, you will be my slave now.

yeah the irony... i have just fixed this, i think.

i have also just scanned the URLs underneath the left-wise expansion slot and it's still a 404 on my phone.

as mentioned in the fine article, it looks like many distros patched this out already, for some reason. Debian has that default flipped since Debian squeeze, published in 2011:

https://sources.debian.org/src/xterm/261-1/debian/changelog/...

so yeah, kind of a big deal, except not really, looks like some folks were already careful with that possibly dangerous setting. :)

other thread has more comments, see https://news.ycombinator.com/item?id=32646836

Canada needs more competition, not a single Kafkaesque provider of services.

I might need to repeat this a few more times, but I am not proposing a single provider. I'm proposing a multitude, regionally-based, monopolies, owned by the relevant state authority (municipal, provincial, federal), based on the scope. Those could be state-owned companies, coop, non-profits, I don't care: the point is that internet access is fundamentally a monopoly, so it should be state-owned, just like roads and water (ways).

author here: I'm a computer engineer, not an economist. had I gotten into "more of an economic analysis", I am sure proper economists (here?) would have slammed me down much harder than engineers (?) did here.

i kind of like having the video. actually running the fuzzer uses a lot of CPU: it makes my fan startup.

it works fine here on Firefox + Debian GNU/Linux 11 stable/bullseye.

ah. right. that's the bit I was missing (pun intended).

thanks for the clarification.

in that sense, yes, of course, if you have bit rot and another disk failing, things go south with just two disk. ZFS is not magic.

If you suffer a total disk failure of one of those disks in the array, you have likely lost some data. [...] The reason is, with a total loss of a single disk, any read error on any of the remaining disks is a lost/corrupted file.

Wait, what? If a RAID-(z)1 ZFS array loses one disk, there's data loss? I've ran so many RAID-1 and RAID-10 arrays with mdadm that I can't even being to count them, and I had many drive failures. If any of those arrays would have corrupted data, I would have been mad as hell.

What I am missing here? How is this even remotely acceptable?

That's pretty accurate, thanks for the stellar review.

That you need to use du to get per subvolume disk usage is the problem here, I think you missed the point I was trying to me. But that's on me, I couldn't imagine someone would seriously consider that I didn't know the difference between df and du...

I was fully expecting this, but you guys really delivered, thanks. A quote from the fine article:

I probably did something Very Wrong and should be punished by Hacker News.

i don't think the security model of package-install is good enough for my thread model. (to be fair, i still package-install some stuff, and, yes, with use-package, but i'm trying to eradicate that practice.)

I mention eglot in the article: the reason I'm not using it is it's not packaged in Debian (yet?). Apparently, it should land in Emacs core eventually (as mentioned in the article too) so I'm basically waiting for that. In the meantime, lsp-mode is fine for the light programming I do on a daily basis.