TheCaconym [any] @ TheCaconym @hexbear.net

Posts

0

Comments

147

Joined

5 yr. ago

How do people not think of that, putting a quarter of your income away monthly, so obvious, I wonder why they don't do it

Also a "retirement" implies a functioning biosphere in which to retire, fat chance.

I really wanted to avoid a debate (doubly so in a thread where some dude just wanted some help), which is why I'm trying not to engage the various answers I got; though just one thing since I apparently can't help myself: Qubes, which you cite, is indeed an example of such improved security done correctly, through an hypervisor and a solid implementation; not cgroups, some duct-tape and the same kernel, and thinking your security has improved. Thanks again, at any rate.

I disagree with most of the benefits you list (chief among them "increased security") - not to mention half of them are already supported by traditional package managers - but I was genuinely curious so thanks for the rationale.

Can I ask why you choose to use one of those weird "immutable" distributions in the first place, out of curiosity ?

I want an openbox/fluxbox look and UI. About the only one I know of is labwc, and it's shit (despite being proudly on your list). I'm fairly sure that a lot of these, in fact, aren't close to usable.

Again, it's not that relevant, for now I can still use Xorg. For now.

Previously, on Linux, your desktop environment is made out of:

• The display server (xorg), in charge of dealing with the video card (by talking with drivers in the kernel through a unified interface, DRI), and handling how to display stuff properly on your particular combination of hardware, including your physical screen and its peculiarities.
• A window manager, in charge of asking software for what they want to draw, then drawing windows, decorating them, etc. and more generally organizing what will be displayed on the screen and how it will be displayed.
• A protocol allowing both to communicate between each other.

That protocol is old, shitty, and insecure. Those are rightful criticisms of it, and it could be argued there is a need for an alternative. This is the often touted justification for wayland.

Note that the way windows and the general desktop environment is handled in the model above is completely distinct from the actual display server; this has a nice advantage: one can write a WM relatively easily, and as such there are hundreds available for linux users to choose from - including some that traditional Windows and Mac users would consider visually exotic and different, such as tiling WMs. This has long been considered a distinct superiority of Linux over, for example, Windows, where all of this is a monolithic block.

Now the dudes that introduced wayland didn't just decide to secure the protocol; they decided to do away with that separation. Now a "compositor" handles all the stuff both xorg and the WM used to do. This means that almost none of the existing window managers work on this thing (actually the truth is none of them do, but Gnome and a few others for example created whole new compositors - today, you can run "gnome" either with that shit or with Xorg, for example), and that there will be far less of them to pick from in the future. The people implementing wayland didn't even consider this an issue at first (everyone uses gnome or KDE, right ? imbeciles), so IIRC third party devs eventually tried to implement a library to restore some degree of separation (wlroots). This still requires reimplementing a WM though, and ultimately is extremely limited anyway due to the very "security" concepts the wayland protocol introduces. Some stuff that was trivial on Xorg will not be possible at all.

You might be considering why we're talking about security in the context of a display server.

Well, the Wayland people noticed that more and more, people were installing software on Linux not through the official repositories of their distributions (which are high quality, somewhat audited, etc.) but from a galaxy of alternatives proposed by a variety of actors: flatpak, AppImage, snap, etc. The reason for this is the quality of software in general has taken a dive, and so has the quality of developers in the open source community; the usual process for someone wanting to be published on, say, debian, would normally have been to follow a few simple rules and to publish your package, accepting it'll be audited and you may have a few points to work on before it'll get up on the repos. Many devs these days are not interested, and deploy their software through the alternatives I mentioned above (which are basically all container or chroot based approaches to produce a "minisystem" with a set of defined libraries, meaning only your kernel will differ from the person having published that package).

As a result, a lot of clueless people are now installing shady software like monkeys on their system, coming from anywhere, just like on Windows. As such, the Wayland creators consider stuff such as an application discreetly capable of capturing the screen, or copying the clipboard from another app, to be potential "security issues". You may be interested to now such "security measures" do not exist on, for example, Windows (but the "security issue" do).

I'm not even trying to argue whether or not they're wrong here. I think mostly they are - the amount of issues and use cases they didn't consider is incredibly large, and it's been biting them in the ass ever since - but it's irrelevant; in theory this would not be much of a problem because, you can just keep using Xorg and your WM, right ? the fear is that maintainers and support for these will dry up (I doubt that, personally), but also and more cruciallly that as Wayland becomes more and more omnipresent for many users, various features from various critical software - such as the browser - will eventually become problematic for Xorg users.

I cannot try it, as my window manager (and in fact almost the entirety of small lightweight window managers) is not compatible with it, and never will be given the insanely higher requirements to implement a compositor compared to a WM. Wayland supporters say that'll change; I don't see how.

I was about to say "no it doesn't" (having installed bookworm a few weeks ago, and most definitely not having wayland), but actually it seems you're right, and "by default" just means "if you choose one of the compatible desktop environments", one of which appears to be the default selection.

If that's all they plan on doing: awesome, actually, this way anyone can pick what they prefer. I was afraid they were going to pull something like systemd (though ultimately it makes sense, as maintaining sysvinit stuff for all services would have been unfeasible; not so, at least for now, with X11/Wayland).

Thanks !

I pray Debian will never include that wayland shit as the default but sadly, I suspect they will.

Yes it is. Much much faster.

There are several reasons, among which (the list is not exhaustive):

• Many positive (and negative, but the overwhelming effect is definitely positive) feedback processes are known to exist, but cannot be quantified. If they cannot be quantified, they are not taken into account into the models, only as vague "climate tripwires" with no certainty of when the trigger will be pushed or in some cases if it has or hasn't already. Some of those have the potential to rival human emissions in scale if we do trigger them.
• Many other similar feedback processes likely exist but we are not even aware of them.
• The source of governmental decisions is the IPCC; when an IPCC report is published, the research it uses is already years, sometime more than a decade, out of date (this makes sense but it does mean there is a lag between current reality and decision-making).
• The IPCC report itself, once actual scientists have finished writing it, is then provided to political actors (the US, the UAE, etc.) in order to reword or rewrite parts of it they deem incompatible with their strategic objectives. There was a leak of the politically-unedited IPCC report about two years back, and the wording was very different.
• Many scientists self censor, consciously or not, because they'll usually end up being called "alarmists". This, among other things, results in the models the IPCC designs and uses being highly optimistic.

As a sidenote, and only somewhat related to your initial question: for a few years now, all IPCC trajectories that do not end up in widespread societal collapse and potentially human extinction rely on imaginary technologies.

Interesting, I assumed they would all shed like my two little bastards

Those cakes will end up with cat hair in them.

The point is that electric cars are shit, have never been a solution to anything, and that they shouldn't be presented as one, doubly so when as a technology, public transport exists.

I Have No Mouth & I Must Scream, right ? still need to play it.

Ah yeah, drivers are another thing entirely. Especially for what I imagine is very proprietary undocumented hardware. The only thing that can help there is a reverse engineer / kernel module dev.

I use fluxbox but that doesn't prevent using gnome apps; my main issue with them is that god-awful look they all assumed overnight a few years ago, without the title bar and the like (I think to match the Ubuntu tendency at the time / trying to emulate the fucked-up universal touch interface thing Microsoft tried to introduce at the time ?), from gedit to, indeed, evolution. I really loathe it. And thunderbird kept a classic look (firefox didn't, which means regular css tweaking to achieve the same result).

Also thunderbird supports calendars and webdav/webcal sync with plugins (though perhaps evolution does as well now, I haven't checked).

Everybody does, comrade

I still use thunderbird. It works well.

On that Windows 95 anecdote, by the way, beyond gaming that's also one of the advantages of wine. Pretty sure their software would run perfectly on Linux with wine.