TR Forums

bthylafh wrote:

I was bored and decided to give Arch another chance. That lasted about fifteen minutes once I started going down the list of how to install completely manually. Partitioning is no big deal, nor mkfs or installing the bootstrap packages, but once I got to setting up grub manually I decided I wasn't that bored.

Arch is a good example of a distro I'd probably use if I didn't require utmost stability. Bleeding edge doesn't appeal to me as I'm not exactly a Linux hobbyist or light desktop user. I need something near enterprise workstation grade and I find I get that out of CentOS. Plus I don't dual boot into Windows. I've exclusively been a Linux user since 2000 and enjoy keeping it that way.

If I was way more of a casual user, I could probably get away with using anything that uses the latest and greatest without too much concern.

Answering someone else's question, going on a tangent, and having that tangent be the OP's point: My body is ready.

I administrate a half dozen Linux machines, I am strictly small fry, but hoping to hear more from a new position where I'll touch about a hundred, but less often.

Tangent/original thread idea: I've been continually distro-dabbling since first post, and I have finally gotten an Arch derivative under my belt. My solid fallback is currently Linux Mint 17 (because Ubuntu LTS base is pretty solid), but I'm spending more and more time in Manjaro lately. I think the underlying truth everyone has been either trying to explain or trying to understand is that there are two parts to distro selection. The first part is what packages are available, how well they are tested, etc. This would determine between, say, something in the Debian stable/testing/unstable, the Ubuntu LTS or non-LTS, Arch, etc. This determines how long between upstream kicking out UberSoftPackage 1.2.0 and your distro having it available.

For example purpose, I like to stay very close to the edge, as close as I can get without having constant breakage. In my case, this means I'd run Debian unstable, or Ubuntu non-LTS, or Arch.

After you've determined the distro family you're drawing from, everything else is technically optional. You can take just about any distro and twist it around to the way you like things, adding and removing desktop environments, adding and removing packages, changing defaults, etc. On the other hand, the further you are from the defaults, the harder it is to maintain, since you're moving against the upstream defaults, it's more things you have to reset/reconfigure on a reinstall or new install. For that reason, you should generally find a variant or sub-distro that matches your chosen layout as best you can.

For example, I'm currently switching from Linux Mint to Manjaro as my base distro. Now I don't have to have as many overlays/PPAs/etc to get the packages I want. Manjaro tracks upstream Nvidia drivers more closely, which I like, since there's really not a lot of distro-level changes to the Nvidia binary-only drivers possible, and also Manjaro has optimus support baked in, which is one less thing for me to track/monitor/fix.

I'm about to try installing gentoo again, right now, simply because it's meant to be customized. I'm going to investigate using it as a personalized server OS.

I still like RHEL and the clones, but it's a lot of work to shoehorn things for newer development onto it.

I'm trying out PC-BSD because of all the Sturm und Drang over systemd. More impressed with it than I expected to be so far. The UI's pretty well thought-out, for instance, and the package management seems nice.

Still not a fan of KDE, but it's pretty OK with Xfce.

bthylafh wrote:

Still not a fan of KDE, but it's pretty OK with Xfce.

At least KDE is very tweakable. The defaults for many of the UI settings drive me nuts too, but it's all configurable.

Mostly I really hate the default "start menu" and PC-BSD 10.1's implementation doesn't give me an easy way to have a KDE 3.5-style menu instead.

I just dock a really narrow panel up the left side of the screen with little launcher icons for all the stuff I use semi-regularly. I rarely need to open the menu. The Unity-haters will go "Blechh, that silly launcher bar is one of the reasons I don't like Unity!" but with today's widescreen format monitors it actually makes sense -- I think it is a good use of screen real estate. Anyhow, IMO the real dain bramage with Unity is the way it puts the menu bar for the application up at the top of the screen (and other annoying behaviors), not the location of the launcher bar... most of KDE's behaviors seem reasonably sane (or can at least be made so with a bit of tweaking).

It *is* a bit of a resource pig though.

% sudo pkg install cinnamon cinnamon-desktop

I'll give that a try; I've gotten pretty used to Cinnamon on my laptop.

edit: aaaaaaand crash. Well, never mind.
edit 2: seems to be a bad interaction with VirtualBox. C'est la vie.

I'm a staid, boring Ubuntu user of many years. I stick to LTS releases only on my desktop and laptop, and really don't want anything more than a stable, pleasant desktop user experience. I even think Unity is just fine (no, really).

So of course, I need something else to tinker with. Enter OpenWRT on a TP-Link TL-WR703N - it's a lot of fun messing around with a CLI in a very limited system (400 MHz MIPS, 4 MB of Flash, 32 MB RAM). Figure it never hurts to have a good, cheap setup to play and learn with.

Odd, I have Cinnamon, VBox, and VMware, all at once, no problems. It's not endemic, at least.

Btrfs was, however, a bridge too far. Maybe next time.

Forge wrote:

Odd, I have Cinnamon, VBox, and VMware, all at once, no problems. It's not endemic, at least.

Something with Cinnamon's 3D acceleration, FreeBSD/PC-BSD's Gallium driver, and VirtualBox interacting in a not-good way, from what I've read. Probably it'd be fine on hardware, or whenever *BSD gets a newer driver.

Btrfs was, however, a bridge too far. Maybe next time.

PC-BSD gets ZFS for free, fwiw.

Well, I've got btrfs working fine now, so pbtbtbt.

Ubuntu LTS for me with gnome-session-fallback to get rid of that horrible Unity thing.

Unity isn't bad. It still has problems, but I don't hate it as much as Gnome3.

Desktop-focused BSD. I do not understand. I can understand FreeBSD on your desktop if you're really BSD-crazy. But desktop-focused BSD.

PC-BSD is a turnkey desktop like Ubuntu, Fedora, Mint, etc. X.org and KDE are installed out of the box, and the user has to do little to no configuration.

So as strange as this is for me to say, I'm a total convert to KDE. I tried it a bunch in VirtualBox testing both openSUSE and Fedora and after some getting used to I've fallen in love with it.

I have basically the same workflow as I did back when using GNOME 2 but at the same time it feels different. Overall KDE is far more integrated and interconnected with all its parts and makes for a much more complete feeling solution. Rather than it being something that's made of many different, slightly non-compatible parts.

I make use of both Fedora and openSUSE for my KDE experiences. I use openSUSE on my workstation where stability is the most important thing and I use Fedora on the laptop. While Fedora leads more towards bleeding edge, I'm more willing to deal with slight headaches on more of a backup system than on my workstation.

Needless to say KDE has completely won me over and really just makes me wonder why I didn't give it a chance sooner than I did.

Funny you should say that now. I'm contemplating possibly jumping ship from KDE to Cinnamon (Mint's desktop). Will probably end up sticking with KDE for at least one more upgrade cycle though!

just brew it! wrote:

Funny you should say that now. I'm contemplating possibly jumping ship from KDE to Cinnamon (Mint's desktop). Will probably end up sticking with KDE for at least one more upgrade cycle though!

Cinnamon is a nice little desktop but more than serious technical issues, my main issues with it were UI and visual. Much like in GNOME Shell, it irks me to no end that neither project have a means of having their fonts handled by the main GNOME settings. I don't like that typeface and font size are controlled by the theme and not your settings which on a really simple level sort of breaks intuitiveness. Users shouldn't have to modify a theme's files just to change the font to match.

jmcknight wrote:

Overall KDE is far more integrated and interconnected with all its parts and makes for a much more complete feeling solution. Rather than it being something that's made of many different, slightly non-compatible parts.

KDE really puts an emphasis on being a cohesive platform, and they've built a lot of plumping to allow applications to seamlessly work together. The major players settling on Gnome as the Linux desktop wasn't a great decision.

"Everyone else is stupid. We're smarter," is the Gnome design mantra.

Flatland_Spider wrote:

jmcknight wrote:

Overall KDE is far more integrated and interconnected with all its parts and makes for a much more complete feeling solution. Rather than it being something that's made of many different, slightly non-compatible parts.

KDE really puts an emphasis on being a cohesive platform, and they've built a lot of plumping to allow applications to seamlessly work together. The major players settling on Gnome as the Linux desktop wasn't a great decision.

"Everyone else is stupid. We're smarter," is the Gnome design mantra.

Well, to be fair, KDE 4 was quite the flustercuck originally. I think that drove a lot of people into the GNOME camp; I was one of them. Then came GNOME 3, and suddenly KDE didn't look so bad any more! Over the past few years, KDE 4 has managed to mature into a very functional and effective desktop that can easily be customized to suit your workflow, albeit still with a few quirks. Of course, the KDE team took this as a signal that it was time to start shifting focus to KDE 5, starting the cycle anew.

just brew it! wrote:

Well, to be fair, KDE 4 was quite the flustercuck originally.

Yes it was. The KDE team didn't handle the 4.0 release very well, and graphics drivers weren't in a state where they could handle the pressure KDE placed on them. KDE4 can be pretty unforgiving of wonky drivers or unsupported cards, and it wasn't shy about leveraging the capabilities of video cards.

I think that drove a lot of people into the GNOME camp

SUSE and Mandravia were, and still are, the two major KDE-centric distros. Most distros picked Gnome as their default desktop, and this was going back to the KDE3 line.

KDE 4 has managed to mature into a very functional and effective desktop that can easily be customized to suit your workflow, albeit still with a few quirks. Of course, the KDE team took this as a signal that it was time to start shifting focus to KDE 5, starting the cycle anew.

Qt5 was released too. KDE also serves as a technology demonstration platform for Qt, and the world's most visible Qt project can't lag behind.

Flatland_Spider wrote:

Qt5 was released too. KDE also serves as a technology demonstration platform for Qt, and the world's most visible Qt project can't lag behind.

Yeah, that will inevitably result in some conflict of interest. But at the end of the day, end users want a functional and stable DE, not a tech demo!

The quintessential illustration of this (from when I first switched to KDE a couple of years back) was discovering that I could rotate desktop widgets to arbitrary angles on the desktop. Want your calculator to sit at a 20 degree tilt? Sure, KDE can do that. Try and drag that same calculator to the other display in your dual-head setup? Nope, that'll cause it to crash! It just felt like there was a little too much emphasis on pointless eye candy, and not enough on basic functionality/stability.

just brew it! wrote:

Flatland_Spider wrote:

Qt5 was released too. KDE also serves as a technology demonstration platform for Qt, and the world's most visible Qt project can't lag behind.

Yeah, that will inevitably result in some conflict of interest. But at the end of the day, end users want a functional and stable DE, not a tech demo!

The quintessential illustration of this (from when I first switched to KDE a couple of years back) was discovering that I could rotate desktop widgets to arbitrary angles on the desktop. Want your calculator to sit at a 20 degree tilt? Sure, KDE can do that. Try and drag that same calculator to the other display in your dual-head setup? Nope, that'll cause it to crash! It just felt like there was a little too much emphasis on pointless eye candy, and not enough on basic functionality/stability.

Being the late adopter to KDE that I was, I've read the horror stories and I'm glad that I missed them. The first version of KDE that I gave a serious go at is what Debian Wheezy ships with now which I believe is 4.8 and was immediately impressed with it. In both Fedora and openSUSE I'm running 4.14 and couldn't be happier.

just brew it! wrote:

Yeah, that will inevitably result in some conflict of interest. But at the end of the day, end users want a functional and stable DE, not a tech demo!

I'm not seeing how it would create a conflict on interest. KDE isn't a Qt project. It's a community FOSS project that uses Qt as its graphical toolkit. Qt will accept patches from them, but it's neither is driving the other's development. The KDE team does what they do because they want to.

It's a community run FOSS project! It's one giant on going tech demo.

It just felt like there was a little too much emphasis on pointless eye candy, and not enough on basic functionality/stability.

I'm not going to argue with that. There were some features in KDE3 they didn't implement in KDE4 until several releases down the line, and the tech demo videos they posted demonstrating their progress were pretty useless.

In the bit of testing I've done with Plasma 5 it seems like the transition from 4 to 5 won't be as rough as 3 to 4. Sure, it's still not done yet obviously but aside from a few bugs, I don't see anything that's really going to be a show stopper in terms of migration once it's declared stable.

Plus I like that it's not such a marked difference from 4. The stuff that needed to be improved is improved and the stuff from 4 that's good mostly stays that. I like these upgrades in projects where it picks up where the last version left off instead of re-inventing the wheel then crapping all over your user base like the GNOME guys did and continue to do.