I would consider myself pretty "technical".
And yet, I am not willing to spend more effort than absolutely necessary for setting up my Linux OS.
Just b/c I am a programmer and even love to use zsh and Git from the CLI does not imply that I have any patience for fiddling with drivers and kernels.
In the "Linux Community" there only seems to exist the hacker (who loves to spend ages tinkering with config files and debugging hardware issues) and the technically incompetent user (who most likely uses Windows anyway).
There is a lot in between.
I really like the idea of Arch but I am intimidated to jump in. What was your experience like when you first booted up Arch?
If you want to quickly try something similar, I'd suggest Manjaro Linux, which is based on Arch. It really feels like Arch but with non-minimal defaults for those who don't want to set up everything themselves.
To make things worse, the general response I've heard from other folks that use Linux full time was, "oh yeah, why weren't you just using USB peripherals?" The general feeling I got was that I was a n00b for even trying to use a BT mouse and keyboard in the first place.
I genuinely wanted to make the switch work, and it's a shame that it just didn't (for me). Honestly, the last thing I need sprinkled into every few working days is 45 minutes trying to get some driver re-installed.
But I managed to solve almost all of them thanks to lots of very kind, to the point and fast responding pros on forum.linuxmint.com.
wow, that's a pretty elitist comment, even for HN. essentially what you're saying is that people who don't like spending time trying to make their computer work are technically incompetent?
I had personally given up on Fedora years ago, but recently was told I should give it a second look and I've not had time to try it out.
There's a move to give stuff generic names, rather than the obscure names they had in the past. For example, Nautilus has been renamed to Gnome Files, or just Files. When a non technical person needs to search for hep this new name makes it impossible for them to create a useful search term.
[files foo bar] is going to be different from [nautilus foo bar]. Frustratingly the old name works for searching, but it's not in any titlebars or about boxes or menu items, so the non-technical person has to just know that files is also sometimes called Nautilus.
Fedora 20 has an appstore. This has something like 4 different names - in the menu, in the title bar, in the about box, in the icon.
For what it is (a rapidly released testing distro) it's lovely - nice community (from what I could tell) and lots of activity.
That's not so much a Fedora thing but a GNOME move, as the RPM containing "Files" is still called Nautilus.
Also, I am a bit surprised since it goes contrary to your argument, because when a user looks for a file explorer he's much more likely to find that looking for "Files" rather than the (rather strange, really) name of "Nautilus".
I would see myself as a technical user and yet I have no idea how the apps on my Android device are called. One is called generically "Gallery" and another one even worse, "E-Mail".
Thing is, the move to generic names is basically an ego trip from Linux developers. Consistent products like Windows or OSX, which are monolithic and have bazillion of users with the exact same configuration, can get away with it; but the Linux world is a forest of different apps from random developers, haphazardly packaged by this or that distribution and continuously updated every few months. In this environment, thinking you can just refer to "Ubuntu files" or "Fedora files" is a pipe dream; often the solution to your problems will be on sources that are not specific to the distro you are actually using (see for example the Arch and Gentoo wikis).
Making the name basically irrelevant unless they need to ask a question about it in which case enter googlability.
It's a bad trend. It's really frustrating when you're trying to find out what the executable or package name is.
In Fedora 20 the name "Gnome Files" was hard to discover (that may have changed in later Fedoras) and non technical people just don't know how to search.
In your desktop environment, 'Open web browser' (as an action), can be associated with whichever browser you prefer. And perhaps a context menu on that to choose between many.
I prefer something like: Gnome file manager: 'Nautilus', to the browser: 'Web'. How ghastly.
To differentiate between many, like both of Windows' web browsers: Edge and Internet Explorer. You could say Windows web browser Edge. Or Edge; Windows' web browser. Windows itself is a confusing name, but that's another conversation.
I think the usability is miles ahead of Ubuntu atleast - I realize that is a personal opinion, but I love that Fedora is a tightly integrated Gnome (and soon Wayland distro).
On the technical I love OpenSUSE for zypper, build (Any software you need is probably on there and it will even build for other distros), rolling releases in Tumbleweed, and the best KDE default environment for over a decade..
That said, stay away from Tumbleweed (rolling release). In my experience it's been even less stable than Arch or Sid. In the past I've had kernel updates making the machine unbootable, though thankfully the distro does keep all previously installed kernels, easily selectable in GRUB. Just yesterday I banged my head against a year-old bug that pgadmin3 is broken because it hasn't been rebuilt against the distributed version of wx (come on).
I use Fedora @work and Ubuntu @home. The reason I use Ubuntu @home is that my roommates run Ubuntu too and we get all the same Versions on whatever software.
But @work I run Fedora. There aren't any particular reasons except that it just "feels" better to work on that on Ubuntu. I definitely recommend to give it a try.
Using something like Fedy or easyLife makes setting up Fedora fun and fast.
That said I've had many performance issues on GNOME - I like the way it looks and the "feel" and I got used to the UI over the last year or more but frankly it's constantly hogging my PC down, both desktop an laptop - when I switched to KDE/plasma 5 I saw my WebGL chrome app go from 40-50 FPS to consistent 60 FPS - both tests done after a clean start and simply starting the app letting it run for a while. The entire system feels more responsive - can't trace the issue but the results are measurable and noticeable
I've had flicker and stability issues with KDE 5 when I last tried it a year ago but now it appears quite stable. My only complaint is that themes/design community are nowhere near to Gnome.
Any distro with a release schedule is going to have caveats about not having prebuilt packages for the latest XYZ. The only things I can think of that might give you a faster update schedule than Ubuntu (in terms of newer versions, not just point releases) would be Fedora, Arch, or Gentoo.
I've forgot to mention I've moved to Fedora a year or two back for this reason - got tired of rebuilding/PPA hunting for every part of my OS when I want to use the latest version of tool/lib X.
And once dependencies are too old (very often) prepare to be rebuilding 5+ custom libs and figuring out the differences between debian package/path layout and what the library build uses ... so much wasted time.
They have an excellent testing/QA process, especially given the speed at which they're developing - this results in a very high quality.
I have lost hours debugging mysterious crashes because of SELinux, and it is really not safe to have components unexpectedly crashing when they are part of your core infra.
Plus I guess that like every security frameworks it runs with priviledge, it has a lot of lines of code, is hard to audit, and thus highers the surface of vulnerability.
Hint: they use for instance strcmp a lot http://stackoverflow.com/questions/24353504/whats-wrong-with...
Their code mixes if(){} with the if() else (without braces)
They are sometimes using enums, sometimes #define sometimes magic values to refer to constant values.
You should really read the source code.
Definitively above the average of C code in the wild, still having well known code smell.
I would like to see if PVS studio could confirm my intuition.
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux....
I know I've been suggesting it to people who are considering Linux for the first time, with the caveat of disabling SElinux.
With that said, I'm not sure if it is good for non-technical people or people who aren't interested in learning about it. Getting certain things installed can be a bit of work, as they don't support things like Chrome or even Chromium by default. I also view it from the perspective of not knowing how to use any graphical installers (do they exist?). If I was aware of that stuff, I'd possibly consider it for non-technical users, but as it stands, I don't.
Yes, it exists and works pretty good when I installed it. Set up encrypted LUKS, EFI just fine and I found it reasonably pleasant to use.
1) strict licensing restrictions - lots of basic functionality/software missing from official repositories; there is rpmfusion of course, but that's not something my mom should know how to install
2) Very short support. You'll be stuck without updates in no time, and upgrading to possibly broken release (that might radically change things) every few months is ridiculous.
So.. if they were to offer LTS version, I'd happily recommend it to everyone. As it is, it's great if you don't mind the caveats mentioned above (especially everything that stems from 2nd point). CentOS is not a viable alternative, as it is "enterprise" OS stuck with archaic software.
New releases are out after around 6 months. If you find release upgrades risky you can always wait a whole year and stay one release behind (any bug would have been ironed out in that time).
I usually don't update in the first couple weeks and I hadn't any problem with a Fedora upgrade in the last two/three years.
For a long time back in the early teen releases there were no sane upgrade paths which pushed me to Debian.
Probably, if I move away from the convenience of a Debian derivative these days, it will be to FreeBSD.
My reason for choosing Fedora is because of its maturity, major player backing and SElinux implementation.
It works great, I feel like a modern Linux user in Fedora.
It's a very educational experience, but not really something the "average grandmother" is likely to have the patience for, particularly when they're just trying to watch a flash video on Facebook. That's the real strength of Mint, and I'm not sure what best fills that niche if Mint is off the table.
EDIT Write the post install script yourself don't use one of the many flawed ones that a search will show up.
I really don't see why the downvote? The OP wants an easy to use distro that their grandmother can use but seems happy to install it them-self.
FreeBSD has good support for DEs with XFCE and Lumina, anyways. At least I've never had a problem.
I was able to get a compositor (Compiz) working on Mate and this fixed the tearing issue. It had some issues where it just kept crashing and rebooting that I never figured out. So I switched back to Gnome3. Although Gnome3 comes with a compositor (I think? And I think it might be Compiz?), I can't seem to get rid of the screen going black when a video starts.
But it really doesn't make a big difference and it's not a big enough annoyance to matter. I don't really use FreeBSD for media anyway. It's just the auto-play videos on Facebook that triggers the infrequent frustration.
I think Lumina is looking really promising, though. I recommend giving it a try if you have the time.
https://micahflee.com/2016/02/backdoored-linux-mint-and-the-...
Linux Mint sucks, but I'm pretty much OK with disabling autologin to root.
About that.
> It’s important to be aware that UAC elevations are conveniences and not security boundaries.
https://technet.microsoft.com/en-us/magazine/2007.06.uac.asp...
>unfortunately, this is also where we run into some of the limitations of UAC. Remember, there is no effective isolation; there is no security boundary that isolates processes on the same desktop. The OS does include some protective measures to keep the obvious and unnecessary avenues of communication blocked, but it would be impossible and undesirable to block them all. Therefore, Microsoft does not consider breaches of that nonexistent security boundary to be security breaches.
https://technet.microsoft.com/en-us/magazine/2007.09.securit...
In default and probably usual configuration UAC does not represent security boundary between medium-il (user) an high-il (equivalent of linux's root). Only if you bother to run under non-admin account are you protected from escalations.
Leo Davidson then provided demonstrations of privilege escalation between those two integrity levels without triggering uac.
I haven't checked Windows 10, but at least until Windows 8 most machines running Windows had been running most software effectively under root.
After that, you have a trusted signature, and it doesn't matter if it the signature page is defaced.
If you're downloading the signature each time you're downloading a new version of a package or iso, and its SHAs, you're using GPG incorrectly.
I've had many signatures for package signers in my keyring for 5+ years. If and when they replace the key, they let the community know and we update our keyrings.
Yes, if you don't want to do that, fine, but for those of us who are careful users of GPG, it's a huge barrier against malware in packages and distros.
1. Get key id of the key used to sign the hashes. 2. Google for key id. Hmm, no one's used the key id before, that's odd. Not in my web of trust. And it doesn't match the key id mentioned on other forums/scripts/sites. I'll definitely hold off until I can verify the key directly with one of the distro maintainers.
Even if you haven't previously established trust, GPG can be an extremely valuable ally.
As linked in the blog post in my other comment, tails explains things quite well:
Ideally, the keys should be distributed in safe fashion too, so this whole question should be moot.
Though perhaps more trivial when the attacker has helpfully provided your operating environment.
I believe you're confusing a PGP public key with a PGP signature for a particular file (made using said public key).
My sentence about "trusted signature" could have been clearer, too. You have a public key that you trust, that you leverage to verify any new materials signed by the package/distro maintainer. So by extension, you can trust the signed hashes.
I've been using PGP/GPG on a regular basis since 1998, so I'm pretty used to the workflow by now and not at all confused about key pairs and signatures (although I will admit that I sometimes got confused about the point of subkeys until I read this a few years ago: https://alexcabal.com/creating-the-perfect-gpg-keypair/)
I'm not bothered by the licensing issues mentioned, and I'm ambivalent about the namespace issues, but I've been increasingly uneasy for some time now about Mint's security practices. Serving downloads over http and not providing GPG signed SHA hashes like every other distro is fairly irresponsible in this day and age.
This recent security issue, and the poor response to it, are basically the straw that breaks the camel's back for me. I'm moving on to Ubuntu-Mate, since frankly Mate was the primary reason I was using Mint anyway. Serving downloads of the most popular Linux distro from the same machine as is running WordPress is cringeworthy, and failing to take the compromised machine totally offline until it's 100% sure the compromise has been mitigated (through reformatting, including boot sector) shows really poor judgment.
I'm a bit sad to be so critical, since I recognize that Clem has done a lot for the Linux world, and as a Mint user I've benefited personally from his work. But when you're distributing operating systems to so many users, you have to take security seriously. To do otherwise, even on a "hobby" project (although I'm fairly sure it's his full-time job now) is pretty irresponsible.
In many ways, I'd like to pitch in, but based on other interactions I've seen and read about, I'm not sure my input would be welcome, particularly wrt security issues.
Edit: I'm also playing around with FreeBSD for my development environment, since I can use Mate on there. To be honest, I don't really need a DE these days anyway, since I only use terminal and a web browser. I should look into just using a Windows Manager.
Edit 2: Apparently they do provide GPG signed hashes. I've been looking for them each time I've downloaded Mint distros, but never came upon them. So I stand corrected.