Farewell, Elementary
cassidyjames.com
cassidyjames.com
> In line with Dani’s wishes, I will no longer be involved in any way at elementary.
This seems just shy of a hostile takeover. I’m not sure I’d be comfortable in running an OS or supporting a company that is now solely owned by one person who felt it necessary to kick someone out who clearly was willing to work for free on the project.
Is there something I’m missing here? I would expect this if there was some huge scandal, but out of the blue this seems like it’s just toxic behavior from a cofounder?
I've no idea about this particular situation or what Dani's motivation was, but "work for free" isn't the right description if it involves retaining half of the company (which has revenue and aspires to be worth more).
I suppose that's a fair interpretation of the situation. In my opinion, though, if the company is heading downwards in revenue enough to suggest sacrificing salaries, the ownership stake isn't worth much in the moment. The value might be earned by working on the project, but I personally value the shares of a company that can't pay its employees exceedingly low.
I interviewed with a finance startup a few years back and everything was great until we got down to benefits. That's when they broke the news to me that even though they were being financed by a cadre of millionaires and billionaires with long histories in the finance and banking system, they didn't have enough money to offer benefits.
LOL. No.
I agree with your sentiment, I’d never work for a place that didn’t offer a basic package.
All that said, I think employers are required by law to provide healthcare?
In a perfect world, basic coverage would be part of what you get for paying taxes by the government.
Health insurance is expensive, but I don't think it's "expensive as hell", I've got an unsubsidized ACA plan for $1200/month for a family of three (two early 40s adults and a 10 year old child; Kitsap County, WA), plus $30/month for dental insurance for the kid. Past employers paid more for better coverage, but this covers the mandatory items, and none of the higher priced ACA plans appear to offer me more things covered or a meaningfully better network, so meh.
A single payer system with no deductible would have vastly better outcomes for employees in terms of costs, ability to seek healthcare before things become catastrophic, and income opportunities. There's literally almost no downside.
Also, you were complaining about the cost of healthcare insurance, and now you're complaining about the cost of healthcare. They're not the same thing, and yet the terms are used interchangeably. :P
Single payer would be nice, but I don't see how we'll get there, when nobody is proposing steps to get us there. I guess there was some push to reduce the medicare eligibility age to 60 last september, but I don't think it went anywhere; I'd like to see a phased in approach of moving the age down by one year ever year for 10 years, and at a faster pace after that, but you'd need to increase funding (which looks to be needed anyway) and pay attention to any other problems that come up.
Wow, that's just so much money! No wonder people with health insurance go bankrupt in the US.
It looks like the core Elementary team was just two people. The length of time between releases makes me think they were understaffed for a while.
A business would rather get support from a larger distro with a larger support team. That market already has a number of established vendors: Red Hat, Oracle Enterprise Linux, Amazon Linux, Ubuntu server...
When I first downloaded Godot, I didn't donate because I had no idea if I'd even use it for long. After a year or so of using it, I came across various reminders to donate, and eventually I did (if I'm completely honest, literally just now). But most people will never come across that prompt again and will forget any need to. Its the same case with Elementary, or any other distro I've tried out. Why pay if I dont know if I'll even use it?
That being said, I can't think of any better way of asking without being obtrusive to the user. I guess you could have some really pointless badge or something you unlock by donating which could help prompt users later down the line, but I cant imagine the FOSS community would be a big supporter of that.
It has been a pretty slick experience, and that's remarkable given the general state of Linux on the desktop (admittedly, I may be more of a command line person, but still).
So maybe instead of "donate", try "pay for getting your very own feature implemented"?
1. If you're effectively hiring a developer to work on a custom feature for you, including the future costs of maintenance and code complexity, the price is going to be very high.
2. The lack of a minimize button was a UX decision. You can disagree with the decision, but if elementary implemented every feature that people with money requested, you probably wouldn't like the resulting UI. Or maybe you would, but elementary would be a very different sort of distro.
The HN discussion on that is more enlightening than the actual item.
So to me at least, elementary OS is functionally dead, even if it's not technically dead. I don't care about the drama or who said what or who owed what person what money - the person who did all the work left, and I expect to stop seeing news about elementary OS soon. That simple.
Spending some spare time developing something useful for others, great. Spending pretty much all your spare time doing so, seems unsustainable but you do you. Having that treated as a given rather than a special committment, with extra demands and workload placed on top, this is a problem. Having to get a second job to support your project, what the fuck is going on.
If you do that and you're happy with it, don't let me interfer. But this doesn't seem like such a happy post.
But that just means we and everyone else but these two people themselves are just picking sides, not necessarily actually seeing and evaluating right/wrong/good/bad/fair/unfair etc.
https://news.ycombinator.com/item?id=30611748
I don't mean to say I agree with that article. (Well, most of it is just reporting which is fine, but there are a few small added conclusions at the end) My own comment on that article was that I read it and arrived at an opposite conclusion from the author. And someone else supplied some enlightening background about history between these two and the author.
>Keep in mind that the author [Bryan Lunduke] is not neutral here: Since he's expressed his position against trans people, and by doing that parted ways with Danielle Foré before (who previously identified as Daniel Foré and participated in his youtube videos), and also you can see that [the article] refuses to use the actual name for Danielle; it's not surprising that he wants to reach a conclusion against her.
Edit: Since there seems to be some confusion, this is about the author of the OP post linked in the comment above, not the OP link of this HN thread.
(edit: nvm, found the tweetstorm from last month https://lunduke.substack.com/p/elementary-os-is-imploding?s=... that goes into some details)
That's never been an obligation wrt. FLOSS development. If you want a supported version of some FLOSS project for your own production use, you can pay someone to be a maintainer.
How do you figure? Companies like IBM/Redhat, Google, MS, Oracle, Amazon and others pay tons of devs to work on OSS. The state of OSS has never been better.
Just because some people can't figure out how to monetize doesn't mean the state of the ecosystem as a whole isn't great.
One, what's the target user? Is it really every type of desktop computing? In that case you're competing with not only KDE but also GNOME, even if we confine ourselves to Linux based desktops. Two major desktops are already a bit much for such a narrow market. I totally get doing something for friends or for the joy of it, but any semi-commercial endeavour pretty much needs to outcompete them both.
It's not like GNOME's problem is too slow development which some healthy competition can fix. Seriously, I don't think I am a completely atypical user when I value most importantly my desktop to stick around for the next decade, and work with whatever new hardware changes that comes around without annoying me too much along the way.
Two, are you developing a desktop environment or a Linux distribution? In the latter case, you're competing with massive projects like Debian and Fedora. The testing required to release a general Linux distribution on just one architecture is enormous. Security response must bascially never sleep if users should be kept from risk. If the answer is "both" I don't know what to say really.
I'm not saying there should never be any more Linux distributions, or that the computing desktop is perfect as it is. Far from it. I'm just saying that competing with the established projects here is realistically going to take a decade or more, and maybe finding a niche other than a general desktop for everyone might be a good start.
Installing Linux is the easiest experience among operating systems on supported hardware.
Updating Linux software is the easiest as well.
Using a desktop environment like Gnome is straightforward too.
Compared to Windows/Mac these simple tasks are hellish — and for nontechnical users, nearly impossible.
So anecdotally, it's a mixed bag. Maybe if I had less Linux experience and more Windows, the story would be different.
Don't know the vintage of Windows you referenced, but in recent experience (Win 10/11) installing a webcam was dead simple, just plug it into a USB port. Printers used to be a gigantic hassle, but at least for the 4 printers available to my Windows laptop little configuration and no added drivers were necessary. (No doubt certain webcams or printers out there could be a pain to install in Windows.)
One example: getting OBS Studio to work on Windows was pretty much a no brainer vs. Ubuntu which needed to have additional library support. While OBS works under Ubuntu using it is still a bit more of a hassle.
In any case Windows seems to have more bugs now than a few years back. I figure added "features" contribute to complexity. Who knows how many more LOC have been committed. I'd be a lot happier with simpler but more reliable systems. It's not the tradeoff Microsoft, and probably various Linux distributions, will be likely to accept.
https://packages.ubuntu.com/impish/obs-studio
You can install the package and the package manager will install all the dependencies. Nothing for you to do.
I am tired of all these fake tales of Linux hardship.
However, on 20.04LTS apt installs OBS v25.0x, not the current version (27.2.x).
What I was talking about was installing v27.2.x on Ubuntu. FWIW the OBS site gives up-to-date installation info, so it's not really difficult for people familiar with unix/linux just some extra hassle. Also using OBS in Ubuntu requires entering a password each time, rather an annoyance. (Sure, ways to work around this exist but another minor hassle to deal with.)
No doubt all this would be a daunting task for users without experience with command-line admin.
And before you install the official driver, you at least get the noveau driver (community supported nvidia driver) by default, which is capable enough to do many, many things.
> Updating Linux software is the easiest as well.
The ease of this stuff varies widely from one distro to another.
When you ask people to change something that they are fine with and used to,
* they think you’re making them do more work by relearning everything
* they’re quick to notice faults, and they’ll jump ship just as quickly as well
Perhaps attitudes like this towards potential users are another reason Linux Desktop still doesn't have much buy in.
I just dont think that most people just care what the thing is running as long as it does the job. Computers are an appliance to many, just like a toaster. If the functionality is close to what theyre used to (windows or mac) then there shouldnt be big adaptility problems until they want to go deeper into the device than web browsing.
The keyword in your statement is "close" but normal people who use the computer as an appliance have extremely low tolerance for "close."
I also understand there are people who just are not willing to learn new things.
I hope i dont sound like im trying to push linux to people. I want people to have choice of many instead of choice of two when they go buy a new PC. And maybe some would be curious enough to try something new.
Of course she cant move to linux due to adobe software, not that i need her to do anything like that. Anyone can use what they want, my point was that most people dont even know theres other choices than windows and mac.
Also ive been using windows all of my life, recently switched to kubuntu because i got tired of the updates. Didnt take me long to learn to use the pc.
In other words, she noticed.
But most people, from what ive seen, only use the internet browser and maybe some built in apps. Keep it there and the lines between OS' start to blur.
Edit: to add, this is usually why i tell anyone interested to use linux is to try the tools available first, then if you get used to them, make the switch. Luckily i had already used same tools on windows (gimp, krita, blender) that also work on linux.
That said, at least both KDE and Gnome have offered fairly comprehensive and complete desktop environment about on-par with a standard Windows or macOS installation for many years now. There's also things like XFCE, LDE, Cinnamon, and probably others – but I don't really keep up.
There are plenty of issues with "Linux on the desktop", but I don't think this is one of them.
Yes, it's distributed ownership, and while that explains why the problem exists, I don't feel it's an excuse. There is definitely room for someone to come in with good product sense plus technical excellence and work to make a really great desktop OS on Linux, it just hasn't happened yet.
The reason why the "year of the Linux desktop" has never happened is because nobody wants to work for free. As soon as you say the word "product" the word "salary" has to follow.
For the "year of the Linux desktop" to happen you'll need dozens/hundreds of these 6-figure employees at every level of the stack (from kernel to desktop environment to apps) to be able to build a coherent system from scratch and being able to ignore/work around all the BS and wastes of time that currently plagues the Linux distro world.
I do believe you'll be successful to an extent, though whether the result can still be called a "Linux desktop" as opposed to something more like Android (Linux kernel but non-standard GUI stack) will be up for debate.
I don’t know specifically what those things are in the case of elementary, but for example one could imagine certain bits of design in a desktop’s file manager being tied to how the distribution’s filesystem is arranged. Sure the file manager would technically work with other filesystem designs, but then the file manager may as well not exist because it brings little to the table that other file managers haven’t already.
So I can see why someone who’s concerned with fine tuning details in the desktop experience they’re crafting wanting to also be a distribution. Unfortunately that’s also a massively tall order due to the time, personnel, and effort required to pull such a feat off.
It also means you can update your software on your own schedule with modifications where necessary instead of being sandbagged into working around someone else's limitations on their schedule.
Furthermore the general desktop for everyone space and the server space are literally the only ones on planet earth that have enough Linux users for a reasonably sized community.
> Two major desktops are already a bit much for such a narrow market.
Basically there needs to be at least one environment that isn't crap. Gnome
- Has a weirdly bad way of handling multiple monitors
- Requires twice as many steps to perform common operations via its graphical interface
- Relies on hot corners which are both possible to trigger by accident and hard to organically discover. It also makes a lot less sense with multiple monitors.
- Spent the first 8 years of gnome 3 constantly leaking memory until it fell over if you left it on too long. This was caused by a weird mismatch between its javascript desktop and its compiled code that was too hard to fix so they decided to just constantly run the garbage collector and marveled at how not horrible this performed in actuality.
- Deprecates features that people use without replacement. See tray icons for a literal decade
- Has repeatedly has major figures in its development promote the idea of removing themeing and extensions in favor of a vanilla gnome experience. An experience that is notably lackluster.
- Has no addon API forcing addons to monkey patch the javascript desktop ensuring such efforts can trivially fail on any version update and addons that aren't actively developed will stop working.
- Worse if they crash they especially under Wayland take down the entire session
- Lacks basic features in its file manager
I could go on.
The basic problem is confusing an ecosystem of compatible software for a company with labor resources that can be allocated to different tasks that is somehow needlessly wasted on developing two different desktops. Eliminate KDE and you don't get a stronger gnome you get a weaker one because someone passionate about KDE doesn't magically become a gnome developer because they aren't reallocated by their manager to work on a similar project you just lose incidental contributions to mutually useful parts of the open source ecosystem and less cross fertilization of ideas.
>This was caused by a weird mismatch between its javascript desktop and its compiled code that was too hard to fix so they decided to just constantly run the garbage collector and marveled at how not horrible this performed in actuality.
No, this is not even close to what happened. There was never any weird mismatch, the refcounting is very similar in both of them. The bug was that the garbage collector wasn't running correctly. Please read the (now 4 years old) blog post: https://feaneron.com/2018/04/20/the-infamous-gnome-shell-mem...
>Deprecates features that people use without replacement.
I would like to know about these desktops that are still under active development and have never removed or deprecated features. I've never seen one. I'm sure if you look hard enough you can find people complaining that Windows 11 doesn't support 8-bit color displays at 640x480.
>See tray icons for a literal decade
This never happened, there has been several extensions for tray icons for the entire decade. The real problem here is that all the APIs for tray icons are broken and nobody bothered to implement one correctly.
>Has no addon API forcing addons to monkey patch the javascript desktop ensuring such efforts can trivially fail on any version update and addons that aren't actively developed will stop working.
This has nothing to do with javascript or addons. In general code that isn't being developed has a high probability it will stop working at some point. And this presents a false choice too. I've heard this statement so many times but nobody has ever proposed what a real addons API should look like. AFAIK there are no desktop platforms or mobile platforms anywhere with an addons API that does everything that can be done with GNOME extensions. The closest I've ever seen are things like addon widgets for Android and KDE which aren't even close to being the same thing.
>Worse if they crash they especially under Wayland take down the entire session
That's not really important anywhere, the apps themselves are supposed to implement session management to insure against any system crash.
If I were you I would avoid commenting on these things again unless you have a proposal to make, they cheapen the rest of your comments. It's honestly tiring to keep seeing these long lists of random grievances where the ultimate suggestion is "just throw it away and start over while somehow avoiding all these issues in a way that is not explained". Do you really think having yet another choice is going to make a difference?
>-------------------------------------------------------------- Here’s the problem: C objects don’t track who owns them; instead, they only track how many owners they have. This is the traditional reference counting mechanism, and it works fine in C land because C is not garbage collected. To the garbage collector, however, the C objects would look like this:
The garbage collector has no means to know the relationships between C objects.
The garbage collector, then, will go there and destroy the root one. This object will be finalized, and the directly dependent objects will be marked for garbage collection.
Only the directly dependent objects are marked for the next garbage collection.
The “solution” is basically throwing a grenade to kill ants. We now queue a garbage collection every time an object is marked for destruction. So every single time an object becomes red, as in the example, we queue a GC. This is, of course, a very aggressive solution.
But it is not all bad. Some early tests shows that this has a small impact on performance – at least, it’s much smaller than what we were expecting.
>--------------------------------------------------------------
This is basically garbage engineering that shouldn't have made it out of the design stage let alone been knocking over systems for 8 years.
Moving on.
In essence a substantial chunk of the Linux desktop crowd feels like gnome is crap and has correctly understood that they aren't welcome to fix it because project owners have a very opinionated vision on how the Linux desktop works.
This is probably why 60% of Linux Desktop users don't use it.
https://linux-hardware.org/?view=os_de
If Pop!_OS succeeds in replacing it and Ubuntu ends up adopting it will probably drop from 40% of users to closer to 5%. It would be in serious danger of irrelevancy at that point if it weren't Red Hats baby.
You are welcome to your own opinion on the utility of the gnome desktop but the idea expressed by the grand parent post that the Linux desktop space ought to just settle on a singular environment that cannot even convince the majority of users to use it despite having more corporate money spent on it in a year than other environments in history seems ill founded.
Edit:
> the apps themselves are supposed to implement session management to insure against any system crash
Well designed software environments don't crash in a way that causes the underlying OS or end user applications to die. One time in a decade would be one time too many.
Session management makes perfect sense as defense in depth for data loss not least of which because hardware failure, power loss, battery exhaustion happen but this isn't even remotely a defense against gnome's unreliable software design where features available in other environments as options require monkey patching gnome's javascript desktop wherein any addon up to and including a hokey little weather widget can somehow destroy the entire session because the underlying javascript changed between X.n and X.n+1.
For practical purposes such crashes can cause the loss of user data and punting responsibility to user apps is absolutely ridiculous.
No, this makes no sense and you're still misunderstanding the problem. This issue occurs anywhere in any system where you connect a GC to an outside system. Every javascript implementation has to contend with it. It also can happen in a lot of other places, for example, if you bind a random C API to Python. It's not a weird mismatch. There are plenty of bad issues with mozjs right now but this is not one of them.
>In essence a substantial chunk of the Linux desktop crowd feels like gnome is crap and has correctly understood that they aren't welcome to fix it because project owners have a very opinionated vision on how the Linux desktop works.
Ok, so what? There are approximately zero projects in the Linux desktop that aren't opinionated, and there are approximately zero people in the Linux desktop crowd who don't think some other project is crap. You don't have to spend very long reading HN and slashdot comments to figure that one out. If you think GNOME was the first project to reject changes, then I have some very bad news for you.
>You are welcome to your own opinion on the utility of the gnome desktop
I don't care about the utility of GNOME. I actually think GNOME is very bad and I agree with some of your other criticisms. I just wish you wouldn't ruin your comment with these other same nonsensical criticisms that we've seen so many times that never go anywhere in a productive direction. Stick to the good criticisms, avoid these petty gripes unless you have something useful to add. Every time I can honestly pin these type of comments to people who aren't desktop developers because it's always the same refrain about how they should have designed it better the first time 20 years ago or something like that. That's a boring comment, it has no intellectual value. You can do better than this. And I'm sorry I don't have a time machine, I wish I did.
>Well designed software environments don't crash in a way that causes the underlying OS or end user applications to die. One time in a decade would be one time too many.
Then don't install badly designed extensions. If you think that is a nonsense response then please rethink what you're trying to say here. The worst offender for hard system crashes in Linux is still driver bugs. Do you really think driver developers want that to happen?
>this isn't even remotely a defense against gnome's unreliable software design
This doesn't follow. The problem here is not the design, the problem is rather a lack of design.
>any addon up to and including a hokey little weather widget can somehow destroy the entire session because the underlying javascript changed between X.n and X.n+1.
And small changes in the drivers can crash your whole kernel too. I still don't know what point you're trying to make here, besides asking for someone to fix the crash.
>For practical purposes such crashes can cause the loss of user data and punting responsibility to user apps is absolutely ridiculous.
No, it is not ridiculous at all. Nothing is being punted. The apps have always had this responsibility. Some crummy desktop applications are just slow to catch up here. Most of the "serious work" apps I've seen (like word processors, spreadsheets, photoshop, etc) have their own auto-backup and have had it for decades because that's the most reasonable thing to do. Almost all mobile apps are also designed this way because the OS can kill them at any time or the device can lose power at any time. But you knew that already?
I had both a 90s hp windows pda and HTC sensation with 768MB of RAM.
Android apps have a defined point in the android lifecycle where they are supposed to save everything when they lose focus in case they have to fall on their sword which they indeed will have to do frequently.
Dying and being restarted as if nothing happened is part of normal operation because it happens a lot.
Your desktop session should never crash. Exiting unexpectedly should be a risk of battery exhaustion, users closing without saving, hardware failure, or power loss.
The idea of your desktop session just crashing out just shouldn't happen and indeed doesn't happen with i3wm KDE or windows.
The fact that extensions can crash is unremarkable the fact that they can crash the session itself is extraordinary. A windows 95 level of bad.
A common refrain in gnome land is
User: gnome lacks functionality
Gnomie: but it has powerful extensions
User: gnome extensions can crash the session!
Gnomie: You don't have to use extensions!
You either have to deal with potentially unstable extensions or limited functionality. Neither is particularly reasonable.
Characterization of the indefinite and indeterminant lifespan of objects as a normal part of gluing a compiled and garbage collected language is simply normalizing pathology.
It's like a smoker saying well sometimes people's lungs just fill up with fluid and shrivel up and die.
Well yes if you fill them continually with toxic smoke for several decades.
It seems painfully obvious that the dependency relationship expressed in the JavaScript should be mirrored in the c code allowing it to remove the garbage before the system falls over.
This seems painfully obvious when the memory is allocated for an ephemeral purpose like a frame of animation.
Can we not destroy everything related to frame n upon drawing it?
Happens to other JavaScript situations seems to be a special class of problem it's like the cigarette smoker telling you that the meth smoker also has problems with his lungs. Well yes yes he does.
Yes but all of these things still happen anywhere and the apps themselves are the only place that can be fully prepared for that, with cloud-synced web apps this is also not a problem. A lot of Linux users really seem to hate using cloud services though, for those users it's apparent that avoiding data loss is an exercise in self-flagellation.
>The idea of your desktop session just crashing out just shouldn't happen and indeed doesn't happen with i3wm KDE or windows.
Well again, that's a nonsense comparison because none of those have a comparable addons system.
>The fact that extensions can crash is unremarkable the fact that they can crash the session itself is extraordinary. A windows 95 level of bad.
No not really. You seem to think the alternative here to monkey patching is a system that doesn't crash. It's not. The alternative is to load binary plugins into the display server which is even worse and has a higher risk of causing crashes and memory leaks.
>User: gnome lacks functionality Gnomie: but it has powerful extensions User: gnome extensions can crash the session! Gnomie: You don't have to use extensions!
I've not seen anyone here saying that, I'm certainly not saying it. I wish you didn't ignore the part where I said to just fix the crashes in the extensions. I also wish you wouldn't turn this into a conversation with hypothetical users. You're talking to a real person here.
>You either have to deal with potentially unstable extensions or limited functionality. Neither is particularly reasonable.
Yes, welcome to engineering. It's the art of being forced to decide between two bad solutions because you don't have enough time to do it right. If you have a lot of spare time then I suggest you volunteer your time to do it, and figure out how to do it in a way that benefits everybody (not just gnome).
>Characterization of the indefinite and indeterminant lifespan of objects as a normal part of gluing a compiled and garbage collected language is simply normalizing pathology. It seems painfully obvious that the dependency relationship expressed in the JavaScript should be mirrored in the c code allowing it to remove the garbage before the system falls over. This seems painfully obvious when the memory is allocated for an ephemeral purpose like a frame of animation. Can we not destroy everything related to frame n upon drawing it?
This again is more nonsense. No, you can't because the longer you wait the longer your GC pauses will be. This is a problem with any stop-the-world generational GC and it's only made worse by plugging foreign objects into it. The solution would be of course to improve the GC (there are a lot of ways to do this) but then it becomes a mozjs problem. GNOME Shell is just a downstream and can't do anything about it. You're barking up the wrong tree here.
Please just avoid talking about this unless you can propose a solution to the problem, if you can come up with something good that works everywhere and is easy to use then a lot of other projects will benefit from it. If you can't then this discussion will go nowhere. The most that could be said otherwise is "don't use GNOME" but I think it's safe to assume you were already doing that.
This is where the entire dialog not only jumped the shark but blew it up while screaming Yippee Ki Yay.
When the complaint is that a system is unreliable the solution isn't to pair an inherently unstable system with cloud synced web apps its not run software that doesn't randomly shit the bed. For example fire up a KDE x11 session and simulate crashing software by shooting its window manager kwin_x11 in the face and marvel at the fact that the window manager died but your applications did not. One can return to normal by explicitly calling kwin_x11 or even a different window manager like compiz and continuing like normal.
I3wm doesn't have the same trick being only a wm but its also literally 1/10th of 1% as much code as the totality of the gnome project and it isn't monkey patched as the user is running it.
That is to say other people handle the issue by not randomly shitting the bed knowing that the orthogonal issue of losing user space state and data is known to be imperfect. The designer of gnome can't fix literally every app in existence and thus ought to pessimistically act as if every failure will result in user data.
> Well again, that's a nonsense comparison because none of those have a comparable addons system.
An addon system that is literally used virtually entirely to add back options that were present in gnome 2 and are present in other systems. Users of x11 can pick and choose different bar/dock applications, window managers, compositors, selection of user space applications. They don't need to monkey patch a javascript process insofar as it doesn't materially improve the power of their environment.
> The alternative is to load binary plugins into the display server which is even worse and has a higher risk of causing crashes and memory leaks.
The actual alternative is just not to remove functionality that you need to monkey patch a javascript program to add back.
> This again is more nonsense. No, you can't because the longer you wait the longer your GC pauses will be.
The traditional methodology is to just not write such code in a language with a GC. There is no real good reason to write your desktop in JavaScript in the first place if performance of the GC is a factor.
Not every technology is worth adopting.
But you are correct in that it doesn't appear to have been widely adopted.
We knew there was a disagreement on how to part ways. But this outcome seems totally unreasonable for an open source project.
I've noticed that that in English speaking media it's common to comply with companies wishes for weird ways of writing names, while in Sweden for example, media usually writes Iphone, as that is the correct way to write a name in Swedish (which also capitalises names)
You also have GNOME shipping on Fedora and Ubuntu which most new users are going to download, which can't even get their stuff together enough to put a freaking dock and activity indicator on the desktop without having to completely switch to a different activities screen. Asking for a dock will literally get you hung in the village square as people assert that normal users want to use keyboard shortcuts. Everything on the Linux desktop has to either be old fashion or eclectic and tailored to the tech enthusiasts already using it (one of the things that Elementary was actually solving). I tried getting my dad to use GNOME and he gave up after five minutes to go back to his Mac (I can't even launch an app?). It's a mess. Every other distro is basically being held together by a scrappy group of volunteers that are supposed to compete against millions of dollars of investment and armys of full time employees.
I feel bad for Elementary. They made a good product that was actually sane for people coming from MS/A, but they were never going to turn it into something super profitable or even sway companies like Adobe etc. to get their apps running natively on Linux.
This isn't doom and gloom about Linux. Linux owns computing. It's more about how Desktop is niche and not even worth getting into at this point, and how the author has much to gain by just walking away and moving on to other ventures.
Edit:
You can downvote all you want if it brings you catharsis, but you know I'm right. The year of the Linux desktop isn't right around the corner, just like it wasn't right around the corner last year, or the year before that. Linux desktops will always be for enthusiasts, students who want to learn Linux, or people who have to use outdated hardware, which is honestly the way it should be so that Linux keeps its identity and doesn't become yet another replaceable component in a proprietary stack like it is on Android (I'm looking at you Fuchsia).
And no, it doesn't matter that you're reading this right now on Sway using lynx and your mother and sister are using Ubuntu that you installed for them. You are on HN. You are not the segment that makes Linux desktops a lucrative business idea. A software engineer needs to know how to think about market segments and the normal user use cases and not be completely self absorbed into how the world works for them.
Shrinking segment, absolutely. Especially for the casual home-user segment. But leading with exaggerated flamebait is not welcome here, as I'm sure you are aware.
Getting volunteer coders to work in the customer-service style that is required to make a desktop for non-technical people just doesn't seem to work. And turning a volunteer community into a profit seeking one usually seems to hit these roadblocks.
We should just accept that success for Linux on the desktop looks like what we have now. Which seems fine to me. Everyone who wants it has it. Unending lust for growth is a trait of megacoroprations and cancers, not great company to keep IMO.
This is patently and blatantly not true. I have installed Ubuntu to my mom and my wife's parents, and they are happily browsing internet ever after, without any malware or crapware. In the past few years I have spent ~0 hours on maintenance of these devices.
Previously, when they were using Windows, it was a never-ending nightmare.
You see, a well set up Ubuntu is perfect for people who need the computer to just work. They just can't break it!
My wife had an older laptop, which had a failing hdd and I used the opportunity to swap for a SSD and asked her to try Ubuntu.
Yeah, she had a lot of questions at first, but eventually she got used to it. She got a new laptop and had to go back to Windows due to her work, but whenever she is just browsing or videocalling anyone, she doesn't mind using the older laptop.
Also, guess which laptop always has issues with the printer?
In my experience any laptop on any system will have issues with the goddamn fucking printer :-)
I had an old dot-matrix printer connected to the parallel port for a very long time. Could "print" with just "cat file.txt >/dev/lpt0". It was great. Downside was I couldn't use it in the evening as it woke up the neighbours.
"More recently, Linux has also been used on the workstations used by animators for drawing and modeling their creations, as the leading producers of animation software have tailored their applications to run on Linux." [1]
[2] is a demo video of Pixar's use of linux for animation and rendering.
[1] https://www.nytimes.com/2002/06/18/business/technology-disne...
[2] https://www.gamingonlinux.com/2016/09/pixar-film-production-...
[3] https://renderman.pixar.com/resources/RenderMan_20/installat...
Also Docks are terrible UI. They are annoying, take up too much space, and you can never really find a good place to move them so they are out of the way.
The Apple Dock was something designed to look cool and fresh in order to sell a bunch of G3 and G4 iMacs back in the day, but it's a step backwards from even Windows 95's taskbar.
Gnome devs staying away from that is one of the more laudable decisions they made in terms of design.
Using Ubuntu, Fedora or any other distro with either a large community or a clear funding path (like PopOS) is great. The “2 guys made a distro and have no funding or way to monetize their work” isn’t a Linux problem it’s a business model problem.
I've been daily-driving KDE/Plasma for... around 7 years now? And I have a really low tolerance for daily technical nuisances. There have been some issues over that period, but honestly not more than I would have to deal with on other platforms.
I also introduced a partner to Elementary a few years back and she has absolutely fallen in love with it despite the shifting set of glitches it has on every update (she calls "personality"). She's clever but not especially technical and had no Linux experience beforehand.
As I see it, there are essentially three desktop platforms to choose from: Windows, Mac, and Linux. Windows is moving in an ad-supported direction that I want no part of and there's even more that I dislike about what Apple is doing (and this from someone who was one of those weird die-hard Apple enthusiasts in the 90s).
So I didn't switch to Linux because I'm fanatical about it, but because it was my option of last resort for having a desktop environment that I felt like I could largely control and use the way I wanted to use it. And, it has turned out that I'm really happy with it, most of the time.
That's probably not going to be a persuasive argument for the average casual computer user. On the other hand, tired tirades against "the year of the Linux desktop" completely miss the fact that Linux is a completely reasonable desktop experience today.
That's just not at all acceptable to the majority of people who want to know they can click a link in Firefox (or Chromium or whatever) and watch the damn video. Perhaps more important: it's a prime example of sidestepping the question that is so frequent in these communities where you don't get answers to your questions, you get told random things sort of related to your question that others have done.
I started this recent test drive after installing Arch on a secondary system intended to be a home server and, within an hour, I had Steam installed, a DualShock 4 controller connected via bluetooth, and I was playing a windows game (Final Fantasy 12 - The Zodiac Age) at 60FPS stable. I honestly couldn't believe how much "just worked."
and i tried using windows 10 shortly after it's release and just couldn't.
people like what they're used to. if you're used to windows, linux is a tough sell. but if you're used to linux, so is windows. linux people like to worry about how it is perceived by the average user, but the average user fails all the time on any operating system.
My next system is going to be a lightweight Linux host with a pile of VMs and something like https://github.com/casualsnek/cassowary, with juicy enough hardware to make it feel like I'm only running one OS.
which is the same as my troubleshooting procedure on ubuntu.
I remember talking with a buddy of mine about Linux, and he said the same thing you did about any distro outside of Fedora or Ubuntu: That Linux is held together by a scrappy group of volunteers competing against millions of dollars of investment and armies full of full-time employees. That was 1998 and what do you know, Linux is doing just fine. Linux desktop is still innovating. GNOME might do stupid stuff (though they haven't "literally" hung anybody that I'm aware of) but they innovate. Marketshare hasn't changed much but the size of the market has grown massively. 3% now is a lot different from 3% in 2010, and there are tons of developers and sponsors helping to keep Linux alive on the desktop.
I was the only Linux desktop user in my office of 150 people for awhile. The absolute bullsh*t everybody had to eat from OSX (let alone windows) astounded me. And they just smile and gobble it up because they have a gui in front of their VPN. Let 'em I say. The Linux desktop will do fine without them.
The best example I know of this is GIMP; it has just a few (4 IIRC, could be 3 or 5 too, but thereabouts) people working on it in their spare time. That's it.
Folks complain that it's not as good as Photoshop, and that's true, but if you consider the team size it's actually pretty darn impressive! How many people work on Photoshop? I betcha it's a few more than a handful.
Not all of the Linux ecosystem is just volunteers though; most kernel contributions come from various companies; Ubuntu and RedHat work on a lot of different things (cue systemd discussion), etc.
Seeing these things as "products" where "users" have to choose from in a "free market" is already a flawed outlook. That's not what people are doing or how they're thinking.
I've been using OSX at work and I absolutely hate it. I keep asking myself..why the fuck is my 16GB/512GB Core i7 Mac taking 3 seconds to open the settings UI? Why does the "finder" app suck so much? How come OSX has got some basic usability gaps like windows jumping out of my screen and not having an "auto raise windows on hover" option..? and the list continues. Windows seems better at some of these UX departments in my POV. Anyway, these proprietary Operating Systems are always a subject of corporate greed. They're never 100% in the user's control and can be remotely toyed with - The companies backing them can happily add/remove your software, send advertisements, spyware, bloatware down your throat and you'll just end up seeing your system resources being leeched away along this process. Apple could be better in this department but I cannot just stand the (subjectively) horrible UX of the entire system.
Anyway, linux can be what you want. I've run KDE Neon on my PC for years without having to deal with any problem whatsoever. It's been smooth and easy and KDE gave me way, way more features than Windows and Mac put together. I'm talking about KDE but on Linux you can get anything you want. All at the convenience of knowing that there is no commercial element behind these things. Since donations are how things usually get paid for in FOSS, you're welcome to donate to keep it lively and bustling. I do, from time to time.
Are you being hyperbolic here? I move between Linux, Windows 10, iOS, macOS, and Android and they are all mostly fine. I can’t say I love or hate any of them although there are things that each seems to be best at.
The virtual desktops and full screen mode combined with separate program and window switching shortcuts were really easy to get used to and it feels a lot better to use than Windows. The M1 SoC is also a lot faster than anything I've used on Windows laptops in the last few years. On my old Windows laptop I had the same experience as you on macOS. The settings app sometimes taking ages to load and freeze randomly when installing updates.
If I am honest they all suck in different ways (or sometimes the same ways) and most people will be happy with any of them. Of course some people have software that require a certain OS but I guess that's the minority. The biggest advantages of Windows and macOS are Microsoft Office, Adobe Products and the bigger user base. But the first two matter way less because everything is a web app today.
I don’t think you’re wrong, I think you’re missing the point. I can’t speak for the elementary developers but I don’t think you'd start a project like this hoping to take over the world. You start a project like elementary because you think something like elementary should exist.
I think Linux’s hardware agnosticism is both a blessing and a curse. Yes, you can run it anywhere, but it doesn’t run great anywhere.
I think that could change though. IMHO, Asahi Linux is one of the most interesting things going on in the world of Linux right now. They don’t have to support a lot of different hardware configurations so I think they have a good chance of really getting things dialed in and creating a super polished product.
Since you work on the kernel, maybe you can answer this - does Linux have something similar to timer coalescing? IIRC, that’s one of thing in macOS that helps them with battery life.
Correct, and the drivers that get worked on are the ones that we kernel developers have access to. The situation has gotten better as more and more vendors have contributed drivers over the years and standards have allowed many devices to run on a single driver and we've become more focused on improving and adopting USB, PCI, etc. Plus kernel builds have become more automated as companies like Intel contribute lots of automated builds against tons of different hardware configurations. Still though, Linux is focused on the server and mobile market. That's what people like me get paid to work on, and so it's going to get the best support.
> Since you work on the kernel, maybe you can answer this - does Linux have something similar to timer coalescing?
Yes it does. Also Google has contributed tons, and tons, and tons of power management code (for Android of course). Power management on Linux has improved tremendously. There's still only so much that can be done, because we don't have as tight of an integration into the hardware stack on Desktop. We can't force vendors to implement certain features or make sure everyone implements certain low power state callbacks in drivers.