Note that the clients being FOSS is of little relevance because all they do is forward a recording to a blackbox proprietary service run by a for-profit company.

The code that has access to your audio and does the actual task at hand is not FOSS in the slightest.

If this is for a user programs rather than system components that must be managed by apt, you could use Nix.

By its nature, it keeps track of all dependencies in a queryable format and Nix stores are actually quite portable; you can just

nix copy /nix/store/6gd9yardd6qk9dkgdbmh1vnac0vmkh7d-ripgrep-14.1.1/ --to /mnt/USB-drive/

and that will copy that store path aswell as any dependency (including transitive deps) to e.g. a USB drive.

You’d then do the inverse in the target environment to do the opposite:

nix copy /nix/store/6gd9yardd6qk9dkgdbmh1vnac0vmkh7d-ripgrep-14.1.1/ --from /mnt/USB-drive/

And then /nix/store/6gd9yardd6qk9dkgdbmh1vnac0vmkh7d-ripgrep-14.1.1/ aswell as its entire runtime dependency tree would exist in the air-gapped system.

Because Nix store paths are hermetic, that’s all you need to execute e.g. /nix/store/6gd9yardd6qk9dkgdbmh1vnac0vmkh7d-ripgrep-14.1.1/bin/rg.

You’d obviously just adjust your $PATH accordingly rather than typing all of that out and typically would install this into what Nix refers to as a profile so that you have one path to add to your $PATH rather than one for each package.

I used a single package here but you could build an entire environment of many packages to your liking and it’d be the exact same as far as Nix is concerned; it’s all store paths.

You do need /nix/ to exist and be writeable in the target environment for this to work though.

Sure :)

I knew about bit rot but thought the only solution was something like a zfs pool.

Right. There are other ways of doing this but a checksumming filesystem such as ZFS, btrfs (or bcachefs if you’re feeling adventurous) are the best way to do that generically and can also be used in combination with other methods.

What you generally need in order to detect corruption on ab abstract level is some sort of “integrity record” which can determine whether some set of data is in an expected state or an unexpected state. The difficulty here is to keep that record up to date with the actually expected changes to the data.
The filesystem sits at a very good place to implement this because it handles all such “expected changes” as executing those on behalf of the running processes is its purpose.

Filesystems like ZFS and btrfs implement this integrity record in the form of hashes of smaller portions of each file’s data (“extents”). The hash for each extent is stored in the filesystem metadata. When any part of a file is read, the extents that make up that part of the file are each hashed and the results are compared with the hashes stored in the metadata. If the hash is the same, all is good and the read succeeds but if it doesn’t match, the read fails and the application reading that portion of the file gets an IO error that it needs to handle.

Note how there was never any second disk involved in this. You can do all of this on a single disk.

Now to your next question:

How do I go about manually detecting bit rot?

In order to detect whether any given file is corrupted, you simply read back that file’s content. If you get an error due to a hash mismatch, it’s bad, if you don’t, it’s good. It’s quite simple really.

You can then simply expand that process to all the files in your filesystem to see whether any of them have gotten corrupted. You could do this manually by just reading every file in your filesystem once and reporting errors but those filesystems usually provide a ready-made tool for that with tighter integrations in the filesystem code. The conventional name for this process is to “scrub”.

How do I go about manually detecting bit rot? Assuming I had perfect backups to replace the rotted files.

You let the filesystem-specific scrub run and it will report every file that contains corrupted data.

Now that you know which files are corrupted, you simply replace those files from your backup.

Done; no more corrupted files.

Is a zfs pool really that inefficient space wise?

Not a ZFS pool per-se but redundant RAID in general. And by “incredibly costly” I mean costly for the purpose of immediately restoring data rather than doing it manually.

There actually are use-cases for automatic immediate repair but, in a home lab setting, it’s usually totally acceptable for e.g. a service to be down for a few hours until you e.g. get back from work to restore some file from backup.

It should also be noted that corruption is exceedingly rare. You will encounter it at some point which is why you should protect yourself against it but it’s not like this will happen every few months; this will happen closer to on the order of every few decades.

To answer your original question directly: No, ZFS pools themselves are not inefficient as they can also be used on a single disk or in a non-redundant striping manner (similar to RAID0). They’re just the abstraction layer at which you have the choice of whether to make use of redundancy or not and it’s redundancy that can be wasteful depending on your purpose.

if it’s a 1:1 full disk image, then there’s almost no difference with the costs of raid1

The problem with that statement is that you’re likening a redundant but dependant copy to a backup which is a redundant independent copy. RAID is not a backup.

As an easy example to illustrate this point: if you delete all of your files, they will still be present in a backup while RAID will happily delete the data on all drives at the same time.

Additionally, backup tools such as restic offer compression and deduplication which saves quite a bit of space; allowing you to store multiple revisions of your data while requiring less space than the original data in most cases.

In this case he’s talking about restic, which can restore data but very hard to do a full bootable linux system - stuff needs to be reinstalled

It’s totally possible to make a backup of the root filesystem tree and restore a full system from that if you know what you’re doing. It’s not even that hard: Format disks, extract backup, adjust fstab, reinstall bootloader, kernels and initrd into the boot/ESP partition(s).

There’s also the wasteful but dead simple method to backing up your whole system with all its configuration which is full-disk backups. The only thing this will not back up are EFI vars but those are easy to simply set again or would just remain set as long as you don’t switch motherboards.

I’m used to Borgbackup which fulfils a very similar purpose to restic, so I didn’t know this but restic doesn’t appear to have first-class support for backing up whole block devices but it appears this can be made to work too: https://github.com/restic/restic/issues/949

I must admit that I also didn’t think of this as a huge issue because declarative system configuration is a thing. If you’re used to it, you have a very different view on the importance of system configuration state.
If my server died, it’d be a few minutes of setting up the disk format and then waiting for a ~3.5GiB download after which everything would work exactly as it did before modulo user data. (The disk format step could also be automatic but I didn’t bother implementing that yet because of https://xkcd.com/1205/.)

I was thinking whether I should elaborate on this when I wrote the previous reply.

At the scale of most home users (~dozens of TiBs), corruption is actually quite unlikely to happen. It’ll happen maybe a handful of times in your lifetime if you’re unlucky.

Disk failure is actually also not all that likely (maybe once every decade or so, maybe) but still quite a bit more likely than corruption.

Just because it’s rare doesn’t mean it never happens or that you shouldn’t protect yourself against it though. You don’t want to be caught with your pants down when it does actually happen.

My primary point is however that backups are sufficient to protect against this hazard and also protect you against quite a few other hazards. There are many other such hazards and a hard drive failing isn’t even the most likely among them (that’d be user error).
If you care about data security first and foremost, you should therefore prioritise more backups over downtime mitigation technologies such as RAID.

ZFS and BTRFS’ integrity checks are entirely independent of whether you have redundancy or not. You don’t need any sort of RAID to get that; it also works on a single disk.
The only thing that redundancy provides you here is immediate automatic repair if corruption is found. I’ve written about why that isn’t as great as it sounds in another reply already.

Most other software RAID can not and does not protect integrity. It couldn’t; there’s no hashing. Data verification is extremely annoying to implement on the block level and has massive performance gotchas, so you wouldn’t want that even if you could have it.

RAID 1 can absolutely be faster than a single disk for read perf, and on Linux it is tuned to be faster.

You’re missing the point entirely. I never said to use a single disk, I explicitly compared it to RAID0.

As far as data security is concerned, JBOD/linear combination and RAID0 are the same, so you’d obviously use RAID0 if you didn’t need redundancy.

staging rebuild cycles only happen every two weeks or so.

The reason is always that something changed and causes all dependent packages to change, requiring a rebuild of those too.

Nah, bergamont works mighty fine offline in Firefox on Android right now already. You don’t need stupid LLMs to do machine translations.

The real difficulty is OCR and combining the two.

It depends on your uptime requirements.

According to Backblaze stats on similarly modern drives, you can expect about a 9% probability that at least one of those drives has died after 6 years. Assuming 1 week recovery time if any one of them dies, that’d be a 99.997% uptime.

If that’s too high of a probability for needing to run a (in case of AWS potentially very costly) restore, you should invest in RAID. Otherwise, that money is better spent on more backups.

Note that you do not need any sort of redundancy to detect corruption.

Redundancy only gains you the ability to have that corruption immediately and automatically repaired.

While this sounds nice in theory, you have no use for such auto repair if you have backups handy because you can simply restore that data manually using your backups in the 2 times in your lifetime that such corruption actually occurs.
(If you do not have backups handy, you should fix that before even thinking about RAID.)

It’s incredibly costly to have such redundancy at a disk level and you’re almost always better off using those resources on more backups instead if data security is your primary concern.
Downtime mitigation is another story but IMHO it’s hardly relevant for most home users.

If you needed to spend any time “setting everything back as before”, you didn’t have a full backup.

RAID does not protect your data, it protects data uptime.

RAID cannot ensure integrity (i.e bitrot protection). Its one and only purpose it to mitigate downtime.

Read perf would be the same or better if you didn’t add redundancy as you’d obviously use RAID0.

RAID is never in any way something that can replace a backup. If the backup cannot be restored, you didn’t have a backup in the first place. Test your backups.
If you don’t trust 1 backup, you should make a second backup rather than using RAID.

The one and only thing RAID has going for it is minimising downtime. For most home use-cases though, the 3rd 9 which this would provide is hardly relevant IMHO.

Magnet seems to be a window management shortcut thingy like rectangle but probably worse, costs money and likely to enshittify.

It cannot influence how the macOS window manager works internally, it can only ask it to e.g. place a window in a certain location.

From Windows

Low-latency VRR that works correctly

It does not feel quite right in kwin and the rather new “proper” support in Hyprland doesn’t feel right either.

In hyprland you actually have to enable a special option and set a lower bound for VRR because it doesn’t handle LFC with cursors, so a game running at 1fps will make your cursor jump around once per second which is totally unusable. With LFC that would typically result in at least e.g. 90Hz.

VRR in other apps works quite well though. I’m not sure how intended it is but it allows for some nice power savings on my Framework 16; when it’s just a terminal refreshing a few times a second, the screen goes all the way down to 48Hz and when I actually scroll some content or move the cursor it’s still buttery smooth 120Hz.

Sway feels very good w.r.t. VRR but it cannot handle cursors at all (visible or invisible): whenever you move the mouse, VRR is deactivated and you’re at full refresh rate until you stop moving the cursor. It might also not be fine because I could only test a racing game due to the mouse issue and it’s so light that it always ran at a constant rate, so that’s not a great test as what differentiates good VRR from bad VRR is how varying refresh rate is handled of course.

Xorg VRR also never felt right; it felt super inconsistent. Xorg is also dead.

VRR is fundamental for a smooth gaming experience and power efficient laptops.

From macOS

Mouse pad scroll acceleration.

If you’ve ever used a modern macbook for a significant amount of time, you’ll know that its touchpad is excellent. I’d actually prefer a macbook touchpad over a mouse for web browsing purposes.
On Linux however, it’s a complete shitshow and the most significant difference is not hardware but software. You might think that, surely, it can’t be that bad. Let me tell you: it is.

Every single application is required to implement touch pad scrolling on its own; with its own custom rules on how to interpret finger movement across the touch pad. I can’t really convey how insane that is. There is no coordination whatsoever. Some applications scroll more per distance travelled, some less. Some support inertial scrolling, some don’t. Some have more inertial acceleration, some less.

Configuring scrolling speed (if your compositor even allows that, isn’t that right Mutter?) to work well in e.g. Firefox will result in speeds that are way too quick for the dozens of chromiums you have installed and cannot reasonably configure while making it right for chromiums will make it impossible to use forwards/backwards gestures in Firefox and applications that don’t implement inertial scrolling at all (of which there are many) will scroll unusably slowly.

It’s actually insane and completely fucked beyond repair. This entire system needs to be fundamentally re-done.

There needs to be exactly one place that controls touch pad (and mouse for that matter) scrolling speed and intertial acceleration, configurable by the user. Any given application should simply receive “scroll up by this much” signals by the compositor with no regard for how those signals come to be. My browser should never need to interpret the way my fingers move across the touch pad.

Accel key

Command/super is just a better accel key than control. Super is almost entirely unused in Linux (and Windows for that matter). Using it for most shortcuts makes it trivially possible to make the distinction between e.g. copy and sending SIGTERM via ^C in a terminal emulator. No macOS user has ever been confused about which shortcut to use to copy stuff out of a terminal because CMD-c works like it does in any other program.

It also makes it possible to have e.g. system-wide emacs-style shortcuts (commonly prefixed with control) and regular-ass CUA shortcuts without any conflicts. C-f is one char forwards and CMD-f is search; easy.

Unified Top bar/global menu

Almost every graphical application has some sort of menu where there’s a button for about, help, preferences or various other application-specific actions. In QT apps aswell as most fringe UI frameworks, it’s placed in a bar below the top of each window as is usual on Windows. In GTK apps, it’s wherever the fuck the developer decided to put it because who cares about consistency anyways.

For the uninitiated: On macOS there is one (1) standardised menu for applications to put and sort all of their general actions into. It is part of the system UI: almost the entire left side of the top bar is dedicated to this global menu; populated with the actions of the currently focussed application.

If you’re used to each application having this sort of menu in the top of its window, having this menu inside a system UI element that is not connected to the application instead will be confusing for all of 5 seconds and then it just makes sense. It’s always in that exact place and has all the general actions you can perform in this application available to you.

There is always a system-provided “Help” category that, along with showing macOS help and custom help items of the application, has a search function that allows you to search for an action in the application by name. No scouring 5 different categories with dozens of actions each to find the one you’re looking for, you just simply search for the action’s name and can directly execute it. It even shows you where it’s located; teaching you where to find it quickly and allowing for easy discovery of related functions.

When you press a shortcut to execute some action in the app, the system UI highlights the category into which the executed action is organised; allowing you to find its name and (usually) related actions.

Speaking of shortcuts: When you expand a category, it shows the shortcut of every action right next to the name. This allows for trivial discovery of shortcuts; it says it right there next to the name of the action every time you go and use it.

This is how you design a UI that is functional, efficient, consistent and, perhaps even more importantly, accessible. Linux should take note.

I’m not sure what you mean? It’s a basic feature of the macOS window manager. Pressing the fullscreen button on a window does all of this.

While that is true for the files that make up the programs themselves and their dependencies, it’s not true for any state files or caches that programs creates at runtime. You need to clean those up manually.

full screen takes over a whole desktop

and creates it. It’s a whole new workspace just for putting an app in fullscreen and none of the shortcuts to jump to workspace x work with it of course.

The rest of the WM can be made bearable but there’s no way around that stupid design choice.

Everything is behind GrapheneOS in terms of security. It does a whole lot more than just being up-to-date.

Whether you need that degree of security depends on your threat model though.