The first is not, the second is.
They are somewhat symbiotic though, they have a point-to-point bonded (10Gbitx2) network connection so most of the files I work with on my desktop actually reside on the server.

1: For ZFS ARC, and why not. (ECC DDR3 was cheap at the time, and none of the components were purchased new).
2: I compile software on it, and why not. More RAM more better.

Those are both at home, so neither are "work" machines. The first is a storage(~64TB)/web/mail/camera/streaming/everything else server, the second is my main general-purpose desktop.

Because Gentoo is awesome. If you want your OS to be your OS, there's no substitute. Think of it less as a distro, and more a collection of tools to build your own distro.
Gentoo is also the second distro I ever seriously used (after Slackware), so it was an obvious choice when looking for a rolling-release that didn't push systemd.

Cuz it's built into the BMC. That board doesn't even have a suitable PCIE slot for a GPU, and Xeons generally don't have integrated graphics.

I could (it's analogue VGA only), but I've never plugged a monitor into it. On the rare occasion I need to get at the physical console (or BIOS), it's via RS232 or network KVM over IPMI.

Since you never miss an opportuity to hold sermons on the doom of the human race any time anyone suggests something new or shit on people for not doing work you don't do either, I'm only responding in kind.

You mean theming. Whoop. De. Do.
Past theming I might add. Full of "was". Full of "would have been" and "has been".

By all means, encourage others to make themes if that's your jam, and you actually have a concept of what work is involved. Don't passive-aggressively berate them to do things you can't or won't do yourself, particularly when you crow about how little you know on the subject.

https://dev1galaxy.org/viewtopic.php?pid=57536#p57536 sums it up quite nicely, so I'll not rehash the obvious regarding the difference between tinkering with CSS and themes, and testing and packaging software or writing code... Or the general quality of your "administration" of this board for that matter.

exponentialmatrix came in here with a suggestion, having already done much of the work to make it happen. Rather than engaging in rational discussion on the topic, you almost immediately started in on some pessimistic "things were better in my day" bullshit, then followed up with passive-agressive demands for people to "contribute" to tasks largely unrelated to the topic at hand and which you won't lift a finger on yourself.

"Hard" is relative to ones experience. If you never try, you won't gain any of that and it will remain "hard" indefinitely.

No X11/GUI is the worst you could achieve, and that's hardly "broken OS". The CLI will remain usable with only the generic VGA drivers in the kernel.

I don't do spoon-feeding, and I don't have any nvidia hardware to test with so any "100%" would be meaningless. Generic backporting instructions are here, and the source package you will want to build from is here.
You might need newer versions of some tooling (e.g. debhelper), but there looks to be little else in the way of dependencies.

Also, somebody has already done the "instructions" thing, and it took all of 10 seconds to find on the 'net. There are probably others.
You could have found those yourself, so I assume what you really want when you say "100%" is some kind warranty... Which you will not get anywhere, so you may as well stop asking.

So why not backport the package, rather than switching release?
I swear I will never understand why people switch releases, or even change distros entirely, all to avoid compiling something.

All of this ridiculous back-and-forth, tangents into who or what what provides $random_gui, blind assumptions about what commands do and what they are for, and posting vague screenshots of who-knows-what or wanting to "tick" something instead of just typing 'man' or 'apropos' [whatever one wanted to know about] or using the perfectly good cron job Altoid served up on a silver platter...

The thread should have started and ended with:

RTFM. NTP is as old as time, nothing of significance has changed.

Generally you want both 'autoremove' to remove unneeded dependencies, and 'purge' to remove preserved configuration files. e.g. 'apt autoremove --purge okular'. IIRC the inverse 'apt purge --autoremove foo' would also work.

Some packages that were installed as dependencies may not be (auto)removed if they are also suggests or recommends from other packages still installed, depending on how you have configured apt (see apt-get manual & debian documentation for details).

Because nobody has volunteered to fix it yet?

If it's the default GUI time configuration "tool" (which, AFAICT it is not), then it should work with the default ntp daemon.
If it's just a GUI tool among several options... The same applies as with any other software in the distro - there's a lot of stuff available, and it's up to the user to decide which they want to use and check compatibility of their selections.

NTP isn't broken, your chosen optional GUI frontend is. Fix it, report it to the relevant bugtracker(s), or use something else.
You could of course just configure your NTP daemon using it's own mechanisims (and have been up and running days ago), rather than insisting on this ancient unmaintained GNOME castoff GUI silliness.
NTP will keep your clock in sync, GUI or not. How often do you really need to change timezone anyway, and why do you need a control panel for a one-line config change?

No, this is an obvious wetware error. Specifically, the user is running the unstable branch, but lacks the wherewithal to deal with the expected packaging flux or report bugs through proper channels.

All the big-tech corporations are the same, this is just a new example of that which has been obvious to anyone with two brain cells for a long time.
It's the same old song: Company sells you convenience and "safety" (cloud backups, all your keys in your online account in case you loose them etc.), charges you in loss of freedom and autonomy.

You can:
A: Behave like a good little consumer sheep, use the big shiny product, accept the default configuration, and deal with the fact that you have effectively no freedom to do as you wish with your technology and your data is not even remotely yours any more.
B: Be scorned by the normies as a "hacker", and either use free software/alternative technologies or (often illegally, if US law is to be believed) modify the usual offerings to not screw you over.

Either way, this isn't really "news", and whinging about it is just shouting at clouds. Most people really don't care, they like being sheep.

If you don't show us what is in /etc/exports and where your filesystems are mounted server, and /etc/fstab on the client, all that remains is speculation.
Speculation: You are trying to export a directory that contains mountpoints for other filesystems, without specifying the 'crossmnt' flag. See the exports(5) manual page.

* Yes.
* Not Currently.
* Depends on who you ask.

Enrico does indeed "break stuff" in master from time to time, but as of right now the Xlibre project as a whole is maintaining pretty good release quality.
So yeah, "broke stuff" is valid... If your goal is to never break anything, even in an unreleased master branch. Run things that way and you don't make any progress though, omelettes eggs, yada yada.

As for politics, I for one do not care in the slightest. I simply want a sane development process and software that fits the needs of both app developers and end users. Right now Xlibre is both, while Wayland is neither (wayland is policy > code, and development by committee. It shows.)

Not really, though conversely I don't see any reason not to set it to >=2.
The most common reason for a dirty filesystem is a power failure, so IMO the most logical time to fix it is when the system starts up again.
If check is disabled, you might find yourself needing to run fsck manually before you can mount the filesystem, I prefer the system to handle that automatically, at the earliest opportunity.

Setting check to '0' in fstab is roughly analogous to always selecting skip for non-system filesystems in that that startup chkdisk screen Windows will sometimes throw at you if it crashes hard enough or you pull the plug while there's disk activity. No big deal, but unless you're in a mighty hurry there's no reason not to just let it do its thing. Ext4 is a journaled filesystem, so a fsck is usually measured in seconds.

Feel free to install your preference and change the default CLI/TUI editor with

update-alternatives --config editor

You'll want to be comfortable with some CLI editor, doing things as root through the GUI is both awkward and generally discouraged.
Personally I like mcedit, it (and mc itself of course) should be very familiar to anyone who has used classic Norton Commander or MSDOS edit.