"it is driving me nuts that old tweets get referenced, new ones are not seen, and context is ignored."
Context doesn't travel well in posts of 280 characters, maybe use a more suitable broadcasting platform.
A core member of the GNOME team has accused System76 of being "a case study on how not to collaborate with upstream" following confirmation that the Linux PC vendor is working on a new desktop environment built with Rust. System76 makes PCs, laptops, and servers with Linux pre-installed, using their own distro called Pop!_OS. …
Or a headline grabbing flavor of the month language
Only a matter of time till a nasty mem manager flaw gets found and exploited and the fix will make it no better than c, that's my prediction for something that will happen in the next 5 years...
Don't get me wrong I quite like rust but I just don't get the hype train everyone seems to be getting on, seems like c devs are getting all fomo wanting a bandwagon to jump on like RoR or node..
Rust isn't interpretive. It doesn't depend on garbage collection. It instead works (mostly) at the compiler level enforcing rules about how values are stored in memory to prevent dangling references and other memory loopholes. It requires the programmer to adhere to some additional rules of coding, so it's not automatic; there's a learning curve. It also enforces certain styles of coding that preclude the "shortcuts" that were needed in the past to extract performance or reduce memory footprint; it's kind of the price you pay for memory safety. Anyway, with almost no runtime checking, the chances of a flaw in the memory safety system are greatly reduced.
That description of Rust sounds a lot like the goals for ADA. This language has been around for some time -- its a solid, well supported, mature language that eliminates a lot of common programming bugs by making them impossible to code. I haven't heard of a rush to write Linux system components in it, though.
Reinventing wheels seems to be a universal programming technique.
Rust and Ada have nothing in common besides that one philosophical notion of making writing bugs harder. "Just use Ada" isn't really solving the problems Rust solves and Ada has many downsides for real world development. It's sort of like saying we don't need Java or C# because Eiffel exists.
Ada on Linux is more complicated than "just use the language"; while there's a POSIX binding for Ada (developed by IEEE TCOS in the mid- to late-90s), it binds language features to OS capabilities, rather than exposing OS APIs to programs in the language. Ada tasks vs pthreads vs the way Linux things about threads are quite difficult to reconcile.
Beyond that, the Rust approach to enforcing memory ownership (borrow checker etc) is a substantial advance over how Ada handles memory allocation.
DOD contractors spent quite a bit of time building software in Ada, and it still has some uses. It's not fair to imply that it wasn't tried.
"Only a matter of time till a nasty mem manager flaw gets found and exploited and the fix will make it no better than c"
Tell me all you know about Rust is its name without telling me all you know about Rust is its name. Rust is not a managed language and there is no memory manager to exploit.
I think it's great that they're coming up with something new instead of creating yet another fork of some ancient software that's been knocking around for decades. Starting from scratch is obviously a lot more work, but at least it doesn't carry any of the baggage that comes with old code, like some massive security flaw that affects a ton of different projects and is a nightmare to fix.
Absolutely! But the OP I was replying to stated "I think it's great that they're coming up with something new instead of creating yet another fork of some ancient software", hence my using "ancient" because to build a "new" desktop still includes planting it down onto "ancient" foundation (or unfinished Wayland ones) :-)
It's really not a great analogy though. Would you fork Xorg if you wanted to build a new display server or would you build it from scratch while potentially retaining API compatibility with X11 to let all existing software run? How you choose a starting codebase and how you choose dependencies may be very different.
> Would basing on some of the many alternatives been better?
Since most of the alternatives have their own problems, whether from historical baggage or recent design decisions whose only basis is small groups or even individual devs thinking it'd be cool to do it that way1, I don't think it's a bad decision. But time will tell, as it most likely will be a '15th standard'-type situation, but maybe it'll go well, or maybe it'll fall in a screaming heap.
1: that's not a bad thing in opensource projects, I mean, it's the reason many devs devote their own personal non-paid time to a project. Therefore it is perfectly reasonable that they spend their time on things that they want to do. But for core systems that many people depend on, it's not that great a paradigm ...
I use KDE or MATE when I'm using something UNIXy, so I don't really have a dog in this fight, so maybe I'm wrong but I will say it sure as hell seems like every time someone has an issue with a desktop environment the common denominator is almost always GNOME or its developers not listening to their users.
or OS-X as it was known at the time. Development was actively hostile to non-Linux hosts.
I wish the Pop!_OS folk luck, but there are a lot of "other" applications out there that are going to want to keep doing things their own way. There will be a deal of work involved.
Gnome devs are hoity-toity because they won against the patent troll but seem to ignore the fact that their desktop is the most tweaked and altered desktop in the Linux world. Mate does justice to what Gnome should have been even after jumping to GTK3. Cinnamon is now the correct modern implementation for Gnome that actually tries to be user-intuitive. Xfce is not Gnome per say but it's a lightweight cousin. LXDE is similar but dying if not dead. I have never seen Gnome come fresh out of the box unlike how it is with Fedora which tells you what a heaping mess Fedora is in the first place. Canonical—and many other projects like Ubuntu/Pop!OS—tweak Gnome to fit the bill because by default, vanilla Gnome does NOT cater to the user... it caters to an ideal that barely 1% of the user space actually considers sane. Projects like MATE and Cinnamon don't exist because people out there think they can do "better than Gnome", they exist because Gnome devs tossed the Linux community to the wayside for laughs and giggles. They have a complex like no other and it's not surprising it's come to this point. I don't think we actually need a whole new desktop, and perhaps there is some rightful questioning to be answered by System76, but I can also see how Gnome is not the ideal for getting a complete working desktop out of the box—not to mention patch/hack. People literally have to hack and or come up with their own forked desktop to make Gnome usable. That's the state that we live in and it's pathetic that any major distro actually offers Gnome by default. Utter trash.
See, I was just going to remark on the odd juxtaposition of Gnome calling out other developers as un-cooperative.
But this, this is first class ranting. We do not need a whole new desktop, someone just needs to fix the two fundamental problems with Gnome - it's performance, and it's developers.
I've been around long enough to remember the founding of Gnome and how they were first trying to rally devs and users on various online forums. The project has asshole woven direct into its DNA and will never be another way.
The whole premise of Gnome was that everyone who didn't want exactly what they wanted (a MacOS UI clone originally) was stupid and wrong because vague UI theory reasons. Ditto all existing windowing libs, drawing libs, etc. Not that they didn't have valid criticisms, but it was always black and white and anyone who wanted other choices to exist just didn't get it, the very idea of dev resources going to KDE used to piss off gnome devs. KDE was wrong and therefore wasteful, plus "who wants to look at windoZe lululul". They have the right answer and you will acknowledge them as far as they're concerned.
Stuff like dragging a guy in public until he quits his job for the crime of working on a competitor after you made it clear Gnome won't meet his needs is just... So in line with the Gnome mentality that I'm certain not one person in the entire project will ever get the issue people take with that behaviour.
Well I'm not a fan of system 76 or pop OS...(did they realize that Popos means ass in Greek...), I haven't used gnome recently and I don't it for all of the technical and political reasons... I mean look around at who worked there recently. From where I stand, the same people get passed around through project to organization because of some cronyism. And the product is crap, in my humble opinion.
I mean some places work for innovative results and some well... They just exist to exist.
Gnome and GTK3+ are a PITA. They care nothing for backward compatibility, they care nothing that people use features they then remove, they care nothing that their idea of what is "Correct" is actually "Not correct", but provide no ways of correcting without hacking the code. Sadly we still have to use it.
I'm glad you are happy.
I'm typing this on a Mint/Cinnamon laptop that I use for Zoom/MS Teams and associated things for working online.(*)
I usually use an older laptop with Slackware and xfce4. When Slackware 15.0 finally launches I'll switch over to the default plasma desktop and perhaps get a slightly newer recycled laptop to run it.
All are fine for the end usery kind of tasks I do each day.
I think that these er... debates... are about the programming under the surface, and especially the effect that re-writing key libraries such as GTK3 has on other applications that people want to write.
(*)Typical workflow is 'windows key -> Type fire... -> press return -> log into Microsoft 365(ish) -> delete the pile of auto-generated and irrelevant emails -> access documents I'm working on and click on the next meeting in the calendar -> make sure I've blown my nose and trimmed the nostril hairs
Gotta say, I've also been using Gnome for a few years and once you get used to it, it's perfectly fine. I quite like it (but then again I like absolutely minimal desktop clutter). My tip is learn the shortcut keys.
I use it pretty much the same way I use MacOS. In Gnome, hit the 'Windows' key and start typing, then hit Enter once the autocomplete shows what you want. In MacOS, hit CMD+Space, start typing, hit Enter when you see what you want. So much quicker than using a mouse. I can see how Gnome may confuse / annoy if you are wedded to using the mouse for everything, but personally I prefer the keyboard approach than clicking through multiple fly-out menus (as in KDE). I'm sure KDE has an equivalent keyboard workflow too, but Gnome works fine for me.
I initially installed Gnome because I'm using it on an MS Surface device, so needed a touch friendly interface. But I find it perfectly serviceable with the keyboard and trackpad too. I did try out KDE and Mate for a bit, which again were both fine, but too fiddly for when I needed touch input.
"Davis then said that System76 fixes bugs in its own distro before fixing them upstream in Ubuntu (Pop!_OS is based on Ubuntu)..."
That just sounds like good business. I fix bugs first in my own branch of projects, instead of waiting to see whether my PRs are accepted and then waiting for a new release of the code and then integrating that into my release.
The article takes a journalistically neutral tone, something quite rare to see outside of El Reg nowadays, but wow from the quotes it sure does sound like one Gnome dev's hot take is being taken way too seriously. I suppose the Gnome devs must be quite defensive and fragile at this point, but you'd think they'd have grown thicker skins by now. FFS it's Gnome, nobody likes it; we just use it because it's the default.
I used to use Gnome, way back. Then Gnome 3 happened, so i tried KDE. Bit fo a resource hog, but it seemed to work well. Then the KDE team decided that the backend had become a bit of a mess apaprently, an they decided to do a release of KDE that was broken.
Oookay. I've used Xfce before, I'll switch to Xfce. Many happy years later...
Wah, some bugger's tweaked the scroll bars to be insanely thin with no way to make them wider (I'm old, sight isnt so good these days, and neither is the steadiness of my hands). OK, so iI ecide to make a post on the Min forums, try to register so I can do that and... nada. SIgh. Iwishdistro creators would include a tool suc that one can send a comment to the devs re usability features without having to sign up to forums.
Never mind, i can always try one of the other lightweight desktops when I'm gaming, or Cinnamon or KDE other times. I wonder if KDE nowadays is as nice as it used to be beffore my bad experience with that one release? Hmm - might take a look-see tomorrow..
Personally I would side with System 76 on this one. They are primarily an OEM and make their living from hardware. More and more buyers expect a cohesive, reliable OS and UI when they buy a pc/laptop.
System 76 need to prioritise their customer experience because broken extensions make them look bad and could give them a bad name. And since the Gnome people are all assholes anyway, deciding to make their own DE was the best decision.
They can style it how they want, add features at will and not have to wait for Gnome to do something. It will be way better for customers. After all if a customer must have gnome, they are free to install and Linux distro that had Gnome.
Seems like a win-win for the customer and System 76.