The tradeoff is that we are pivoting away from work we had been doing on desktop applications and will cease shipping LibreOffice as part of RHEL starting in a future RHEL version. This also limits our ability to maintain it in future versions of Fedora.
We will continue to maintain LibreOffice in currently supported versions of RHEL (RHEL 7, 8 and 9) with needed CVEs and similar for the lifetime of those releases (as published on the Red Hat website). As part of that, the engineers doing that work will contribute some fixes upstream to ensure LibreOffice works better as a Flatpak, which we expect to be the way that most people consume LibreOffice in the long term.
I’m no fan of Flatpak for a multitude of reasons, but at the same time, I can’t blame Red Hat and other distribution makers for not wanting to maintain a complex set of packages such as LibreOffice. This does give me pause regarding my current use of Fedora on two of my three machines, as I do not wish to rely on Flatpak for anything serious.
Generally, the idea that the OS vendor should “maintain packages” is absurd. Apps should be self-contained packages. That’s how it is in Android and MacOS for example. And even in Unix, we’d put stuff in /opt/app_name and consider it a self-contained package, there would be no pesky dependencies and dependency hell.
But you can always rely on the communitah to create a problem and then create a highly complex solution for it (Flatpak, Snap etc). And also start an eternal format war in the time inbetween (deb vs rpm).
Meanwhile, Steam is slowly evolving into the default package manager for Desktop Linux, and Wine is slowly evolving into the default stable API for Desktop Linux. Can you imagine if Desktop Linux achieves mainstream success that way, with a proprietary DRM client as the default package manager and the Win32 API as its stable API? Richard Stallman’s beard and Eric S Raymond’s mustache would both spontaneously combust.
kurkosdr,
Well, the thing is people complain about both ends of the spectrum (including you). It’s fine if you want self contained packages, we can use them today, but it means we need to tolerate less efficient use of resources. For better or worse, the distro manager is the only entity that can really coordinate things. This isn’t a limitation of linux itself, it just has to do with the fact that without coordination 3rd party developers will not be in sync…different library versions, different security patches, etc. When left to 3rd party developers, it’s quite likely some resources will be wildly out of date. Obviously opinions are mixed about which tradeoffs are worth making, but we should recognize that letting all applications bundle their own resources does come with cons.
Flatpak and snap are there to distribute apps like android/ios do. Also it sure seems rich to complain about flame wars when you’ve been pouring fuel on that fire.
I disagree. I find it very frustrating how much breaks in my windows VMs and old laptops. You can blame me for running unsupported versions of windows, which is fair, but then you have to give up the argument that windows is a stable platform. It is a moving target that most software developers follow when they update their development tools & environments.
In my experience the main barriers for linux aren’t technical, they’re political. Even when writing portable code or using a portable toolkit that generates linux applications literally for free, many software publishers are focused on operating systems with more market share. And while it irks that it reinforces monopolies, I do understand they’re following the money. Most application developers won’t even target macos with nearly 10X the market share of linux, so why would they target linux, or heck the BSDs, haiku, or anything else? For better or worse market share matters in business decisions, and this alone has been a big reason why most application developers won’t support linux. Wine is there to help bridge the gap because we live in a world where monopolies are a reality, unfortunately.
The verdict is in on this one: As long as the major apps (that are likely to be hit with exploits) get updated with new versions of the app containing updated versions bundled libraries (either by the OS like Android does it or by the app itself like it’s done on Windows), nobody cares about outdated libraries on other software. Do you really think I care that my copy of Need For Speed Hot Pursuit 2010 has outdated dependencies? Of course not. Nobody will bother making an exploit in the form of a save game, and I won’t bother loading it, So, it simply isn’t worth exploiting. And it only connects to specific locations on the web too. Even file processing stuff like old media muxers and demuxers (for example, a tool called AviAddXsubs comes to mind), who’s going to spend time to exploit that? But you know what those outdated bundled libraries do? They make that software work without dependency hell issues.
But no, such imperfection cannot be tolerated by the more OCD in the communitah (who somehow manage to capture all decision-making positions).
kurkosdr,
I am skeptical about 3rd party developers doing as good a job with security updates as the repo managers. Who knows, I’d very much like to see the numbers on that.
Maybe or maybe not. Say for example it bundled an XML library with a vulnerability that was exposed to the network. Would that matter to you?
Hint: years ago this actually happened to us for real. We reported the fault in msxml to microsoft, and they confirmed it. Hopefully you can at least appreciate where I’m coming from on this.
I honestly don’t think most people would share your opinion. If the software truly is obsolete/useless then remove it completely, but otherwise if users are installing the software then I still think there’s value in fixing the known vulnerabilities to protect users. Of course if it’s proprietary software this introduces it’s own problems, though that’s a different topic!
You are welcome to do that. All I’m saying is that there are pros and cons.
Are you trying to evoke/imitate eric cartman? Just like you, mispronouncing “y” as “ah” used to be a running gag on south park.
https://www.youtube.com/watch?v=NiiUQKFz3HU
“but it means we need to tolerate less efficient use of resources.”
That is the price of not standardizing on a single desktop and instead wasting hundreds of thousands of hours on competing desktop environments. You can either have awful dependency managers, or static linking of everything even if it’s already on your system. Torvalds himself doomed any chance of the success of desktop Linux by encouraging this competition. Now he cops out with “well Android/Chrome OS….” like all the other fanboys. Good at programming, terrible at business decisions.
As for Linux + Wine SDK as a main system, by the time that becomes popular it would only a brief stopgap before ReactOS starts gaining momentum.
dark2,
I don’t agree. I would argue the repos actually handle it quite well given their underlying goals. Honestly, that linux repos manage to do this as effectively as they do on as large a scale as they do is impressive.
Maybe you don’t think the goals are worth the effort, and I think that’s fair enough. As sukru said, this may have been more important in the past when memory and disk were far more limited. Now that we have several magnitudes more maybe it detracts from the need to be efficient with resources. The rise of snap/flatpak/appimage are likely a result of this reassessment. There’s still the matter of security updates, I believe leaving it to outside projects will realistically lead to worse practices. Maybe you can argue “hey it should no longer be a repo’s responsibility to do this”, but that’s a change that needs careful contemplation.
Linux is much bigger than one person. You can call people like me linux fanboys, but to be fair I’ve criticized it a huge number of times. I just think criticism needs to be objective and thought out. All too often I see linux being criticized on with shallow arguments “linux sucks because I don’t like linux” that fail to consider why things are the way they are. More thought needs to go into the pros and cons of alternative possibilities. You are free to criticize the self contained packages, and you are also free to criticize the centrally coordinated repos. There are weaknesses with both. But at the end of the day if you are just criticizing without proposing any realistic alternatives, then you don’t have a viable solution that’s any better than what we’ve got.
So if you think you have a solution that’s better than what we have, then lay it out on the table so it can be debated.
We’ve been waiting for that for an awfully long time. If reactos could somehow succeed more power too them I say! It’d save me from the need to keep windows environments around for the few things I still need them for.
Legally I’m afraid reactos may be in a precarious position to fight microsoft should they wish to enforce their copyrights. The US courts have decided that APIs are copyrightable (not just implementations, so the onus would be on reactos to build a fair use defense for why they should be allowed to violate microsoft’s API copyrights. I’m not even sure they could afford the millions it would take to go to court. This is only hypothetical for now, but microsoft could avail themselves of their legal options if reactos were to become a competitive threat.
Android and MacOS are not good equivalents when comparing them to linux distros, since they both are single vendor closed ecosystems. So there is, officially, a single point of entry for apps there; their app stores.
Maybe we operate in different circles, but I don’t know a single person who uses steam on linux as a packaged manager.
Yeah, I have no clue how Steam is a package manager outside of the games it will install. It isn’t like you can install LibraOffice through it…
javiercero1,
I would call it a package manager, obviously they specializes in gaming, but the difference is in lines of business rather than in technology.
leech,
I don’t think they actually want to compete with the FOSS repos because there’s no money there. In terms of commercial software, I’m sure they’d love to get a cut. Ten years ago they added non-gaming categories, but it just wasn’t popular with commercial users or developers, so it didn’t go anywhere as far as I can tell.
https://www.pcgamer.com/steam-offering-non-game-software/
I think conventional windows software developers are wary of getting into bed with app stores because there are billions of dollars that developers stand to loose in fees compared with direct sales.
One of the most annoying parts of the Linux community is the tendency to have to wrap any understanding of how computers work into how their understanding of 1980’s Unix computers work. Therefore any service that downloads and installs software becomes a “package manager” because that’s just how all computers must work (even though these things are only downloading single files that are guaranteed to work with no dependencies, but because they understand only Unix/Linux, Win32 is clearly a dependency that might be missing….)
dark2,
The entirety of your argument was an ad hominem attack. Can you be less of a hater and more constructive?
MacOS allows installing stuff from outside the App Store because not every developer wants to give Apple a 30% cut and Apple doesn’t want to risk alienating those big developers.
The infrastructure and integration tools you are complaining about were created almost 30 years ago, when most computers had something close to 4 MB and hard-disks could store something close to #00 MB. Economy of memory and disk space were top priority. It was a reality very different from what we have now, with even users computers having 16 GB of ram and storage capacity in the # TB.
For many developers, it would be prohibitive to keep a server so that users could download their binaries and compilation was expensive for non-trivial apps, not something most modern machines can do in a few minutes, as of now. The service provided by distros was a gift from Heaven.
Fact is, it worked very well for many years so, why drop it now? For me, the current situation is excellent: already tested apps get packaged, when a new distro release springs, I get things updated with a single command under my prompt. If I need a new version in a hurry, there is flatpak, appimage or snap.
Actually, getting everything updated with a single command is so good that it was “copied” by Chocolatey or was migrated/implemented by msys2.
A little bit of historical perspectives help us understand why things are the way they are, not?
If you have a better proposal, go ahead, I am all ears.
64gb and 128gb of ram is rather common today, and 16gb is starting to look on the small side as even graphics cards has in some cases have surpassed that in video ram alone.
Very few distros really push desktop usage anymore. Ubuntu dropped it and with that the other players have quietly followed suit. There are “desktop friendly” distros, but none on a scale with pockets deep enough to make a dent. Amazon, Microsoft, Oracle, Suse, Ubuntu and now Big Blue/RH have all followed this path. The money and marketshare is in servers, so that is where they are headed. Libre office is the first, but there will be many more user applications being dropped soon enough.
When did Ubuntu drop support for the desktop, that is news to the Ubuntu workstation I’m using to browse this site right now….
@javiercero1 Right? They haven’t dropped desktop support, and Debian will always try to be the universal operating system. RH dropping desktop packages is in scope with their ‘will they / won’t they’ attitude toward Desktop usage. Right around 20 years ago they outright said it would never happen… then they started to be the biggest financial / code backers to Gnome.
They haven’t dropped support, but they’ve dropped focus. We saw that with the death of Unity.
Compare current Ubuntu releases to what was it was trying to achieve originally
https://www.osnews.com/story/11635/review-ubuntu-v504-hoary/
Ubuntu are now a server OS first, that can also be a desktop. That niche is now being filled by Mint and other distros using Ubuntu as a base.
A better example of the shift if shuttleworth’s own words ofc https://m.slashdot.org/story/332797
Yeah, but there is a huge difference between expanding the scope of the distro to non-desktop spins, and claiming that Ubuntu is dropping the desktop altogether.
About when they closed bug #1, with serious President Bush “Mission Accomplished” vibes. Effectively stating they no longer care about improving the desktop experience [because they’re making tons more on server contracts and barely anyone actually uses a Linux desktop].
Here’s your native LibreOffice for RPM based distros: https://download.documentfoundation.org/libreoffice/stable/
Works fine in Fedora. Been using it this way for over a decade now.
So i guess Fedora will go heavy on Flatpak and Ubuntu on Snap. Debian likely to stay more traditional for now.