I’ve become a little cynical of FOSS (though I remain a supporter) because of politics and infighting and constant reinventing the wheel.
A perfect example is on this website if you click the “Stands” tab. So much unnecessary reinventing the wheel there, it’s sad.
In my idea world (everyone has an ideal world), Linux would be standardized in everything but GUI, and the DEs would run their own distributions with their respective DEs.
But of course I’m going to get a lot of replies about how more uniformity among distributions is bad. I’ll read them but I still think we are way too divided for desktop success and and are spending way too many resources copying each other.
Why is that your opinion? Don't they exist because of demand for them? Why else would the exist? And why is too much diversity bad? What are the negative consequences here? The only thing I can think of is projects that you favor are lacking resources but if that's the case which ones do you even mean?
I mean I in the context of being a developer for Linux.
I’d love to support my apps on Linux… but it is hell. Flatpak seems like it will help immensely, but Flatpak has a ton of people on here who hate it.
Secondly, a quality perspective. So many bugs remain because so much effort is going into patching the same bug in countless distributions, rather than using those resources to make a smaller collection of codebases more bug free.
Thirdly, it intimidates newcomers, be they users or developers. Which distribution should I use? If I have a problem, is the distribution at fault?
> but Flatpak has a ton of people on here who hate it.
Not a Flatpak-related reply, but generally speaking: when it comes to tools for your work, don't listen to forums like this. Just use what you like and see how it works in the market. If Flatpak works for you, use it.
Otherwise you'll end up tied into obscure LISP dialects and other stuff that really doesn't work well in the real world.
> I’d love to support my apps on Linux… but it is hell. Flatpak seems like it will help immensely, but Flatpak has a ton of people on here who hate it.
As someone who professionally writes native desktop software for all 3 major platforms I can't say I've had the same experience.
Providing packages for the major packaging systems is fairly easy, though less so when cross-compiling to different architectures. These formats very rarely require any changes. Other distribution methods like flatpak can be provided by those that care.
For Windows you require signatures, dealing with the registry and installer compilers that need wine to work outside of Windows.
On macOS you need a signature and notarisation, and any new requirements they decide to add. Some of which you can't do without a mac.
Agreed. I have written some video games that work across several platforms and in my experience, GNU/Linux was the easiest platform to handle by far.
> though less so when cross-compiling to different architectures
If what you develop isn't very expensive to compile, cross-compiling can be super easy as well - you can just `docker run` a foreign architecture image with qemu-user and have it work exactly the same as it would natively (just much slower).
Having developed commercial desktop software for Linux, Windows and Mac I would say the churn is a much bigger issue on Mac than either of the other two. Developing for Linux is mostly fine.
I don’t mind diversity in distributions but the diversity should be based on config options and build flags rather than separate package trees and custom patches
On Linux success. I think one of 3 things would fix my issues (and thus solve mainstream adoption everyone else) either 1) every computer component manufacture open sources their drivers/firmware or 2) open source drivers get very well funded and supported all around (especially graphics and networking) or 3) Linux standardizes a driver ABI so that proprietary drivers can work on newer versions of the kernel without breaking.
First, constant churn that always breaks backward compatibility forcing upgrades and never allowing any old software to be used without a permanently high maintenance burden.
Second, the constant diminution of quality in FOSS is annoying af.
Third, FOSS has become a religion for many. All things must be FOSS because it’s a holy conquest, and if just everything were FOSS all would be perfection because reasons.
Unfortunately, I’ve seen things only get worse as FOSS has come to dominate the world. These could be completely unrelated, and maybe everything would still have been crap, but I cannot help but think that were people forced to pay for development they’d be more discerning.
> First, constant churn that always breaks backward compatibility forcing upgrades and never allowing any old software to be used without a permanently high maintenance burden.
Is that not what enterprise Linux is for? You just pay for that if you need it. And also what about exceptions like python and such.
> Third, FOSS has become a religion for many. All things must be FOSS because it’s a holy conquest, and if just everything were FOSS all would be perfection because reasons.
This seems dismissive since there are actually well-thought out arguments for these things posted everywhere, they are not based on faith nor worship. This dismissal is tantamount to you saying you don't care about the same things that they care about. How could this possibly effect you negatively?
> Unfortunately, I’ve seen things only get worse as FOSS has come to dominate the world. These could be completely unrelated, and maybe everything would still have been crap, but I cannot help but think that were people forced to pay for development they’d be more discerning.
I wouldn't go as far as religion. It is not blind. Some in the FOSS community are ensuring what they see as a UN charter of inalienable human rights for what people can do with code that has a presence in nearly every arena of human life. That presence will grow not diminish.
Diminution of quality and constant churn isn't a very good reason to stop allowing people to look at code and fork it. And you see these terrible qualities magnified in monolithic programs like JIRA or EPIC, which is fully closed and runs the entire healthcare backend for the US and some nations abroad. EPIC and other programs like it are a rentseeking disaster. Replacing it with an open source equivalent would be a huge win for downstream costs and extensible interoperability.
I see your point, and don’t entirely disagree, but I am skeptical. If EPIC became open source you’d now need more tech staff to maintain your installations too, as the maintenance burden would creep ever larger, and at some point a business decision would be made to use closed source SAAS again anyway.
This is observed behavior. The churn in the rest of the ecosystem forces everything to be in a state of constant releases, patches, and upgrades. The alternative is that you depend on outdated code and then must maintain it yourself or just freeze your system forever… which is very much worse unless you’re talking about something non-networked.
> Third, FOSS has become a religion for many. All things must be FOSS because it’s a holy conquest, and if just everything were FOSS all would be perfection because reasons.
If FOSS were actually a religion, the holy conquest you describe would be the most adorable and least bloody holy conquest in the history civilization.
If FOSS were actually a religion, then what is stopping you from joining the least bloody and most adorable religion in the history of civilization?
>I’ve seen things only get worse as FOSS has come to dominate the world
You know how the IT world looked before FOSS? Relentless rent-seeking with the software controlling access, combined with a very high barrier of entry, financially.
I also think people give too much weight to the developer being paid for the software or not. There's shitty closed source software and also shitty FOSS stuff. Brilliant minds can work on either. Many eyeballs on the code haven't prevented Heartbleed, for example, and the high paid devs at whatever company also make mistakes and maintain huge leaky legacy codebases all the time.
The big thing with anything digital is that it can be copied very cheaply. This is a game changer because if I have something digital, we all can have something digital, which is not a case, for example, with a piece of bread, which gets smaller the more we share. This quirk or the digital space can give rise to a movement where a single contribution can empower literally every other person, with no additional expense on the contributor's part. This, coupled with ideology that emphatizes personal empowerment, and a lawful document that ensures further participation in the ecosystem, is what made IT accessible to many, without resorting to piracy.
I'm sure folks can be annoying af about this. But the world hasn't gone to crap, it has always been crap, and the free software / open source movement is one of the larger influences that made it so much better.
I certainly agree that certain aspects harm its own adoption, primarily from choice anxiety which can be overwhelming. That said, there is limits to what I think should be standardised. Nothing stops standardised approach being incorrect, and inertia allowing problems to fester.
I've noticed that is a a problem in open source video games. To be fair it doesn't easily translate into the FOSS realm but maybe we think too much of converting existing commercial methods of development to FOSS. An open source dreams like platform, where there is standards could be more effective.
That said that doesn't mean everything should be unique. We should allow time and experience to generate good robust concepts then allow those to be easily utilised.
> In my idea world (everyone has an ideal world), Linux would be standardized in everything but GUI, and the DEs would run their own distributions with their respective DEs.
Desktop success is pure business. Windows was a piece of shit back in the 90s, especially from Win 95, and yet, it was everywhere. Bill and the gang made sure that everyone who has a PC has Windows installed, by the countless business tactics they employed: bundling Windows with the PC, equipping schools so that they teach Windows and Office, and a myriad of other background deals and other doings with the software and hardware participants on the market. And also they didn't fight software piracy too much.
And since Windows is ubiquitous, people who want to target the PC target the Windows platform. This includes hardware makers, software shops, etc. Contributing to the effect that "everything just works" on Windows.
So for Linux to catch up on this, would mean to make the system more of a baseline for the other participants on the market. And this is not achieved on a technological level primarily. Linux is already there, for a long time now.
Just wanna mention I'm impressed on how smooth the conference is so far. Crazy to see a fully free conference, organised largely by students with so many parallel tracks going as well as live Q/A and going smooth as butter.
The "FOSS on Mobile Devices" devroom[0] got so many CfP submissions that it could have become a conference on its own :) Check it out if you're interested in mobile GNU/Linux!
I would expect it to be faster than that. Much like last year the talks are pre-recorded, and most were available within days of the conference ending (on Sunday).
Keep an eye out on https://video.fosdem.org/ and in the mean time there's plenty of great talks from previous years to catch up on.
I've been wanting a capability based OS for a decade now. I was hoping to catch a glimpse of the genode command line, and how powerboxes work. I've tried Genode in a VM, but there doesn't seem to be a command line anywhere, it's all parts, and no actual OS.
Genode itself is literally calling itself an "OS framework" for a reason, and a shell isn't really a core component of an OS vs something that runs on top of it. There is "Sculpt" as a basic desktop OS based on it to show off some of it.
67 comments
[ 6.0 ms ] story [ 139 ms ] threadFOSDEM is a free event for software developers to meet, share ideas and collaborate.
The description is accurate and far more important.
A perfect example is on this website if you click the “Stands” tab. So much unnecessary reinventing the wheel there, it’s sad.
In my idea world (everyone has an ideal world), Linux would be standardized in everything but GUI, and the DEs would run their own distributions with their respective DEs.
But of course I’m going to get a lot of replies about how more uniformity among distributions is bad. I’ll read them but I still think we are way too divided for desktop success and and are spending way too many resources copying each other.
Why is that your ideal world? What's wrong with diversity and competition?
Too much diversity is also a bad thing.
In my opinion, we have too many distributions for not enough people.
I’d love to support my apps on Linux… but it is hell. Flatpak seems like it will help immensely, but Flatpak has a ton of people on here who hate it.
Secondly, a quality perspective. So many bugs remain because so much effort is going into patching the same bug in countless distributions, rather than using those resources to make a smaller collection of codebases more bug free.
Thirdly, it intimidates newcomers, be they users or developers. Which distribution should I use? If I have a problem, is the distribution at fault?
And so on.
Not a Flatpak-related reply, but generally speaking: when it comes to tools for your work, don't listen to forums like this. Just use what you like and see how it works in the market. If Flatpak works for you, use it.
Otherwise you'll end up tied into obscure LISP dialects and other stuff that really doesn't work well in the real world.
As someone who professionally writes native desktop software for all 3 major platforms I can't say I've had the same experience.
Providing packages for the major packaging systems is fairly easy, though less so when cross-compiling to different architectures. These formats very rarely require any changes. Other distribution methods like flatpak can be provided by those that care.
For Windows you require signatures, dealing with the registry and installer compilers that need wine to work outside of Windows.
On macOS you need a signature and notarisation, and any new requirements they decide to add. Some of which you can't do without a mac.
> though less so when cross-compiling to different architectures
If what you develop isn't very expensive to compile, cross-compiling can be super easy as well - you can just `docker run` a foreign architecture image with qemu-user and have it work exactly the same as it would natively (just much slower).
First, constant churn that always breaks backward compatibility forcing upgrades and never allowing any old software to be used without a permanently high maintenance burden.
Second, the constant diminution of quality in FOSS is annoying af.
Third, FOSS has become a religion for many. All things must be FOSS because it’s a holy conquest, and if just everything were FOSS all would be perfection because reasons.
Unfortunately, I’ve seen things only get worse as FOSS has come to dominate the world. These could be completely unrelated, and maybe everything would still have been crap, but I cannot help but think that were people forced to pay for development they’d be more discerning.
Is that not what enterprise Linux is for? You just pay for that if you need it. And also what about exceptions like python and such.
> Third, FOSS has become a religion for many. All things must be FOSS because it’s a holy conquest, and if just everything were FOSS all would be perfection because reasons.
This seems dismissive since there are actually well-thought out arguments for these things posted everywhere, they are not based on faith nor worship. This dismissal is tantamount to you saying you don't care about the same things that they care about. How could this possibly effect you negatively?
> Unfortunately, I’ve seen things only get worse as FOSS has come to dominate the world. These could be completely unrelated, and maybe everything would still have been crap, but I cannot help but think that were people forced to pay for development they’d be more discerning.
That's pretty vague, care to clarify?
Diminution of quality and constant churn isn't a very good reason to stop allowing people to look at code and fork it. And you see these terrible qualities magnified in monolithic programs like JIRA or EPIC, which is fully closed and runs the entire healthcare backend for the US and some nations abroad. EPIC and other programs like it are a rentseeking disaster. Replacing it with an open source equivalent would be a huge win for downstream costs and extensible interoperability.
If FOSS were actually a religion, the holy conquest you describe would be the most adorable and least bloody holy conquest in the history civilization.
If FOSS were actually a religion, then what is stopping you from joining the least bloody and most adorable religion in the history of civilization?
Come.
Join us.
(Patches accepted.)
You know how the IT world looked before FOSS? Relentless rent-seeking with the software controlling access, combined with a very high barrier of entry, financially.
I also think people give too much weight to the developer being paid for the software or not. There's shitty closed source software and also shitty FOSS stuff. Brilliant minds can work on either. Many eyeballs on the code haven't prevented Heartbleed, for example, and the high paid devs at whatever company also make mistakes and maintain huge leaky legacy codebases all the time.
The big thing with anything digital is that it can be copied very cheaply. This is a game changer because if I have something digital, we all can have something digital, which is not a case, for example, with a piece of bread, which gets smaller the more we share. This quirk or the digital space can give rise to a movement where a single contribution can empower literally every other person, with no additional expense on the contributor's part. This, coupled with ideology that emphatizes personal empowerment, and a lawful document that ensures further participation in the ecosystem, is what made IT accessible to many, without resorting to piracy.
I'm sure folks can be annoying af about this. But the world hasn't gone to crap, it has always been crap, and the free software / open source movement is one of the larger influences that made it so much better.
That said that doesn't mean everything should be unique. We should allow time and experience to generate good robust concepts then allow those to be easily utilised.
Try a BSD!
And since Windows is ubiquitous, people who want to target the PC target the Windows platform. This includes hardware makers, software shops, etc. Contributing to the effect that "everything just works" on Windows.
So for Linux to catch up on this, would mean to make the system more of a baseline for the other participants on the market. And this is not achieved on a technological level primarily. Linux is already there, for a long time now.
Kudos to the team!
I hope I do not jinx it by saying this :)
To be fair, most of us were students 15-20 years ago and have even grown a few grey hairs...
[0] https://fosdem.org/2022/schedule/track/foss_on_mobile_device...
So much for FOSS on Mobile Devices
Cuts the second column with no possibility to scroll.
Keep an eye out on https://video.fosdem.org/ and in the mean time there's plenty of great talks from previous years to catch up on.
I've been wanting a capability based OS for a decade now. I was hoping to catch a glimpse of the genode command line, and how powerboxes work. I've tried Genode in a VM, but there doesn't seem to be a command line anywhere, it's all parts, and no actual OS.
All the times on this website are displayed in the local time zone for Brussels which is Central European Time, CET (UTC+1).