LWN: Comments on "GNOME 3.22 released" https://lwn.net/Articles/701386/ This is a special feed containing comments posted to the individual LWN article titled "GNOME 3.22 released". en-us Thu, 23 Oct 2025 14:49:31 +0000 Thu, 23 Oct 2025 14:49:31 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net GNOME 3.22 released https://lwn.net/Articles/702069/ https://lwn.net/Articles/702069/ nix <blockquote> I've looked out of morbid curiosity with where things are going now and then, and each time it seems the corrections to the obvious pathologies of the present become separate and worse pathologies of the future. </blockquote> You think this is specific to GNOME? It's more a property of all human systems built without attention to formal properties (whether human-built, like software, or systems made of humans, like political or economic systems). The only way out that I've ever heard of is to put a floor on it by having some objective measurement that is simple enough to be ungameable (that houses you build should not fall down, say) -- and even this is only a partial solution. Wed, 28 Sep 2016 12:49:53 +0000 GNOME 3.22 released https://lwn.net/Articles/702024/ https://lwn.net/Articles/702024/ flussence <div class="FormattedComment"> Here's a similar perspective, which seems to agree with you:<br> <a href="https://lwn.net/Articles/667610/">https://lwn.net/Articles/667610/</a><br> <p> I have to agree too. Krita is an outstanding piece of software because it focuses on *being good*. I can think of examples of applications that only cared about “taking back an entire userbase from the competition” above all else — and somewhat succeeded — but that's not a sustainable long-term attitude.<br> </div> Tue, 27 Sep 2016 19:11:57 +0000 GNOME 3.22 released https://lwn.net/Articles/701975/ https://lwn.net/Articles/701975/ johannbg <div class="FormattedComment"> What I was trying to ask ( which apparently I could have worded better ) has little to do with current upstream linux open source application developers but more to do with selling point for the future ones since the problem the Linux ecosystem and open source software is facing sooner than people might realize is how to recruit fresh minds and contributors from the phablet generation and the means to get them ( and keep them ) engaged in at least the component that make up the core/base OS level so in the long run it can continue to keep it self relevant and sustain itself.<br> <p> <p> </div> Tue, 27 Sep 2016 14:42:23 +0000 GNOME 3.22 released https://lwn.net/Articles/701973/ https://lwn.net/Articles/701973/ halla <div class="FormattedComment"> There is an appimage because that's what runs everywhere. There's a snap because Michael Hall provided the scripts and making a snap and uploading it takes only a few minutes, so it's not a big burden on me -- but I don't test them, I only make them. Contributions are always welcome. There's no flatpak because nobody has made one yet.<br> </div> Tue, 27 Sep 2016 10:27:06 +0000 GNOME 3.22 released https://lwn.net/Articles/701972/ https://lwn.net/Articles/701972/ halla <div class="FormattedComment"> I was not presenting an argument -- not everything people say is an argument -- I was telling you why I am making Krita available on Linux. You asked, I told you why I do that, why I care about it, and you just have to simply accept that.<br> <p> (And sure, one day, when I have the time and when there are interesting android devices that support pressure sensitive stylus input and that have enough CPU power, I might port Krita to Android. Or someone else might do that. Tizen... Probably not, I don't care about Tizen, just like I don't care about the BSD's. My choice, no argument needed, you don't have to try to convince me I'm wrong.)<br> </div> Tue, 27 Sep 2016 10:22:43 +0000 GNOME 3.22 released https://lwn.net/Articles/701969/ https://lwn.net/Articles/701969/ anselm <p> Why do some people insist on composing music for the bassoon when there are so many more pianos around? Both are musical instruments but you reach a larger target audience with piano music – and think of all the other keyboard instruments you could throw into the mix and their artistic creative possibilities, so you end up winning people who play your music in the millions and billions, not in the mere hundreds or thousands, right? </p> Tue, 27 Sep 2016 09:43:56 +0000 GNOME 3.22 released https://lwn.net/Articles/701954/ https://lwn.net/Articles/701954/ johannbg <div class="FormattedComment"> I have to ask why would you want to provide additional image formats of Krita to the current Appimage provided one? <br> <p> There does not seem to be any pre install requirement in the OS ( unlike flatpak and snaps either distribution or end users need to jump through hoops to be able to use it ) as in users just have to download it, give it executable permission and run it so things cannot be any simpler than that from an end user perspective. <br> <p> On top of that you would be signing yourself up for added load as in having to test the application on all provided formats Appimage, Flatpack and as an Snap so why would you do that? <br> </div> Mon, 26 Sep 2016 18:10:20 +0000 GNOME 3.22 released https://lwn.net/Articles/701942/ https://lwn.net/Articles/701942/ johannbg <div class="FormattedComment"> Your philosophy/licensing argument is not compromised by providing it as an Android or Tizen application. <br> Both are Linux platforms the only difference being that you reach a larger target audience on those platforms and you can throw all kinds of mixed devices into the creative mix for them to play with and getting artistic about or with so you end up winning users in the millions and billions not in the mere hundred or thousands right.<br> </div> Mon, 26 Sep 2016 17:47:33 +0000 GNOME 3.22 released https://lwn.net/Articles/701940/ https://lwn.net/Articles/701940/ halla <div class="FormattedComment"> "Now here's a much more serious question people should be asking themselves what's the incentive for upstream application developers today (open or not ) to write an application for an platform that has less then 1% market share and is as fragmented as there are the number of the stars when they can focus on two platforms IOS and Android and reach users in the billions?"<br> <p> I use Linux, and have used Linux since 1993 or so. I care about free software, I care about software freedom.<br> <p> Art is the oldest human endeavor that still persists. It's the thing that binds humanity together. If all anyone has is a cave wall and river full of rust-red clay, everyone has the same opportunity. <br> <p> I care about enabling people and giving them the means to participate in the wider world; that is I why I work to make it possible to participate when it comes to creating art in this post-cave-wall world.<br> <p> And that's why I care about Linux, and why I am developing Krita on Linux. Sure, the Windows users are in the majority now, and Krita will never be on iOS because Apple forbids it, but that is no reason for me to desert.<br> <p> By making Krita really good, on Linux, I have won users by the hundreds, if not thousands, for Linux, and for software freedom.<br> </div> Mon, 26 Sep 2016 17:21:59 +0000 GNOME 3.22 released https://lwn.net/Articles/701938/ https://lwn.net/Articles/701938/ halla <div class="FormattedComment"> Sure, and I totally agree with that.<br> </div> Mon, 26 Sep 2016 17:15:50 +0000 GNOME 3.22 released https://lwn.net/Articles/701936/ https://lwn.net/Articles/701936/ farnz <p>Key to it is the phrase "bugs happen". If there's a bug that only happens if your total horizontal screen width is greater than 8192 (2**13) pixels and you're using a graphics tablet with at least 5000 lpi resolution, then the distro is unlikely to catch it - that's an extremely specific hardware setup needed. On the other hand, that's exactly the sort of hardware setup that an artist (Krita's target users!) might use - two high-end displays next to each other, plus a tablet for drawing on. <p>Given that bugs happen, I'd rather isolate the bug to something that I'm updating because I <b>want</b> the update, than to something that I need to have work, but where the update has no immediate benefits to me; once I'm not in a hurry any more (e.g. I'm past my deadline), I can update the system as a whole. Mon, 26 Sep 2016 17:13:00 +0000 GNOME 3.22 released https://lwn.net/Articles/701925/ https://lwn.net/Articles/701925/ johannbg <div class="FormattedComment"> There is something seriously wrong with the distribution if updating it ends up breaking the entire OS install and setup for it's end users but I get what you are hinting at. <br> <p> In the case of krita, the krita users who are accustom to using the already provided appimage those users aren't going to suddenly switch to flatpak or snap to get their application from. <br> <p> Since the implementation of these solutions is just as fragmented as downstream distribution packaging, upstream aren't going to provide images in all formats so they will just settle for a single format ( either appimages,flatpack or snap ) just like they settled on supporting single distribution or none of it ( those that are pushing for those solutions are the ones doing all the leg work providing these applications as flatpak or snap ).<br> <p> Now here's a much more serious question people should be asking themselves what's the incentive for upstream application developers today (open or not ) to write an application for an platform that has less then 1% market share and is as fragmented as there are the number of the stars when they can focus on two platforms IOS and Android and reach users in the billions? <br> <p> The next evolutionary step mobile/tablet devices are going to take is not going to be moar cameras but fixing screen sharing to tv's and computer screens on the mobile/tablet devices directly through the device itself or as an add on docking station and once that happens you can kiss your traditional desktop farewell since users have suddenly found themselves with an device fraction of the size of an laptop and an pc which fits into their shirt pocket, with all the application they use at the fraction of the price of an laptop or an pc and they need and carry with them at all times.<br> <p> so who has the sales pitch that manages to sell application developers today the reason why they should be developing application for Gnome, KDE or the rest of DE's instead of focusing strictly on android or ios instead and let's hear it.<br> <p> <p> </div> Mon, 26 Sep 2016 16:58:55 +0000 GNOME 3.22 released https://lwn.net/Articles/701932/ https://lwn.net/Articles/701932/ farnz <p>Any other application that's got developers who both care about non-technical users and consistently improve it (rather than simply maintain it in minimal-bug condition) will do in place of Krita. Circling back round to the <a href="https://lwn.net/Articles/701515/">top of the discussion</a>, the query was "why would I prefer to only get updates to one or more applications, rather than the entire system in one go", and the same arguments apply to snaps, appimages, and flatpaks - in all cases, there are reasons to prefer to update just one application at a time, and leave the base system untouched until I've got a bit more time to fix any collateral damage from a big update. Mon, 26 Sep 2016 16:19:36 +0000 GNOME 3.22 released https://lwn.net/Articles/701930/ https://lwn.net/Articles/701930/ halla <div class="FormattedComment"> Well... To be honest, there is no Krita flatpak yet, and I think the reason are the runtimes. With appimage, runtimes aren't necessary, and while your Linux distribution needs to have general support for snaps, snaps don't have dependencies either. And I think that the need for a runtime is one of the reasons nobody has contributed a Krita flatpak yet. I would welcome one, but it seems harder than snap or appimage.<br> </div> Mon, 26 Sep 2016 15:57:18 +0000 GNOME 3.22 released https://lwn.net/Articles/701928/ https://lwn.net/Articles/701928/ farnz <p>If all I'm changing is the Krita flatpak, why would the runtime flatpaks change? Equally, if I'm changing the runtime flatpaks, why would the system packages be affected? <p>The benefit here is a degree of isolation between the application and the system - I can update the two out-of-step. You can get much the same effect using system packages, debootstrap, a set of bind mounts, and some code to enter the per-application chroot every time you run it, but flatpaks mean that someone else has done that for you. Mon, 26 Sep 2016 15:49:10 +0000 GNOME 3.22 released https://lwn.net/Articles/701927/ https://lwn.net/Articles/701927/ anselm <blockquote><em>On the other hand, if I simply grab the Krita flatpak, it's reasonably well isolated, and hopefully the Krita devs have confirmed that the bundled libraries in the flatpak and the runtimes it depends on all work together for Krita - I thus don't have to spend time fixing issues with my WM just because I've updated Krita.</em></blockquote> <p> That's of course presuming that the QtQuick library is actually part of the Krita flatpack and not part of the lower-down Qt “runtime” dependency that whoever made the Krita flatpack had decided was so basic that it was going to be installed everywhere the Krita flatpack would conceivably be used, in which case you're not better off than with a plain Debian installation (because upgrading the Qt “runtime” might break your WM). </p> Mon, 26 Sep 2016 15:43:20 +0000 GNOME 3.22 released https://lwn.net/Articles/701922/ https://lwn.net/Articles/701922/ farnz <p>Actually, that feels to me like an argument in favour of flatpaks; I want <b>just</b> a new version of Krita ASAP, so I'll take the hit on my data charges. I do not want to spend 30 minutes trying to fix LXDE after a QtQuick update, downloading random things to try and get back to a state where my computer actually works. Thus, while my total downloaded data is higher if I use flatpaks, it's tightly focused on the things I want to do. If I use a traditional distribution, I'm stuck not wanting to update Krita (even though the new version will save me more time than the cost of a gigabyte of data), just in case I break LXDE, KDE, or whatever DE I use. Mon, 26 Sep 2016 15:23:20 +0000 GNOME 3.22 released https://lwn.net/Articles/701916/ https://lwn.net/Articles/701916/ johannbg <div class="FormattedComment"> "I think you're trying to optimize the wrong problem -"<br> <p> I would have to disagree since the evolution here on these parts is that everyone is moving more and more to just use mobile bandwith through mobile hotspot from their mobile device(s) for anything other than their home network/wifi since it's easier to simply switch that switch on that device and have internet access which is available anywhere ( as long as there is mobile connection ) rather than requesting access to corporate networks or signup through something online for wifi access so any unwanted and excessive bandwidth download ( in the background by the OS ) which is not directly related to users online activity or trigger ( like manual updating ) is frowned upon. So I would expect that people will prefer smaller distribution optimized package update ( rpm or dept that have been optimized to deal with this from the time when modem was still a thing ) rather than having to deal with a full blown application updated and all it's dependency's in a huga package.<br> </div> Mon, 26 Sep 2016 15:18:51 +0000 GNOME 3.22 released https://lwn.net/Articles/701881/ https://lwn.net/Articles/701881/ farnz <p>I think you're trying to optimize the wrong problem - I have fast internet and a fast computer, so the time cost of updates is <b>not</b> about the size of the updates, or the number of updates, but about the human time it takes to deal with the consequences of an update. <p>In that respect, flatpaks speed things up; when I update the Krita flatpak, I get changes to Krita and nothing else. If I update "just" Krita and its supporting libraries in a traditional Linux setup (say Debian, for the sake of argument), then discover that the new Krita update depends on a new QtQuick library that's supposed to be ABI compatible with the QtQuick that my window manager uses but isn't actually compatible (due to a bug on one side or the other - it doesn't matter to me whether it's the WM at fault or Qt at fault here), I then lose time debugging and dealing with the issue in QtQuick and my WM (which, by the nature of these things, is almost certainly specific to my setup, as the WM developers wouldn't have released it if it didn't work for them). I can't just downgrade QtQuick, because then I lose the new Krita whose .deb dependencies point to the version of QtQuick it was built against, nor can I stay on the new version, because then my WM doesn't work. <p>On the other hand, if I simply grab the Krita flatpak, it's reasonably well isolated, and hopefully the Krita devs have confirmed that the bundled libraries in the flatpak and the runtimes it depends on all work together for Krita - I thus don't have to spend time fixing issues with my WM just because I've updated Krita. When I have time to cope with the consequences, I can update everything, and deal with any unintended fallout. Mon, 26 Sep 2016 14:26:20 +0000 GNOME 3.22 released https://lwn.net/Articles/701829/ https://lwn.net/Articles/701829/ smcv <div class="FormattedComment"> <font class="QuotedText">&gt; ostree is a bit like git: files are stored by their hash. This means that if a file is used by multiple flatpaks it only needs to be downloaded once and stored once.</font><br> <p> This also means that reproducible builds (as in &lt;<a href="https://reproducible-builds.org/">https://reproducible-builds.org/</a>&gt;) can help ostree and flatpak to do better at this.<br> </div> Sat, 24 Sep 2016 21:33:29 +0000 GNOME 3.22 released https://lwn.net/Articles/701810/ https://lwn.net/Articles/701810/ MrWim <div class="FormattedComment"> <font class="QuotedText">&gt; With flatpacks, you're going to have to download copies of every program which uses that library.</font><br> <p> I'm not involved with flatpak, but I don't think this applies. My understanding is that in flatpaks have the concept of a "runtime" which includes a bunch of libraries and facilities that are commonly used between multiple apps. For instance some apps might use the "Gnome 3.22" runtime which might include zlib, openssl and other common libraries. If an openssl update is required then only the runtime need be updated.<br> <p> If the libraries are bundled with the apps then the individual apps need to be updated, as you say; but this doesn't mean many 100MB downloads. Flatpak uses ostree to store and download apps. ostree is a bit like git: files are stored by their hash. This means that if a file is used by multiple flatpaks it only needs to be downloaded once and stored once.<br> <p> Note that I'm not saying there aren't disadvantages to flatpaks. I'm just saying that your *specific* example isn't a problem that Flatpaks suffer with.<br> <p> Disclaimer: I have a handful of commits to ostree.<br> </div> Sat, 24 Sep 2016 13:00:56 +0000 GNOME 3.22 released https://lwn.net/Articles/701783/ https://lwn.net/Articles/701783/ xtifr <div class="FormattedComment"> Man, I swore I'd stay out of this, since my general opinion is that flatpacks and similar technologies have advantages and disadvantages, and anyone who pretends they lack either of those is simply deluded. But!<br> <p> Flatpacks actually exacerbate the problem you're complaining about: updates taking too much time. If a single library gets fixed, with a traditional system, you simply update that library. One reasonably quick download. With flatpacks, you're going to have to download copies of every program which uses that library. The complete programs, not just the libraries. So now your under-a-megabyte download has suddenly become a whole set of multiple hundreds-of-megabytes downloads. Which means you're going to be even more reluctant to do all that downloading, and will continue to suffer from bugs and even security flaws in a bunch of your lower-priority apps.<br> <p> Furthermore, you're going to be downloading duplicate copies of a bunch of libraries that haven't actually changed, making the amount you need to download--both now *and* in the future--even *larger*!<br> <p> A better approach to your *specific* problem, I think, would be to put all your potential updates on hold, and then just download the ones you really want right now. Then, when you have more time, you can take off the holds and update everything else.<br> <p> Note that I'm not saying there aren't advantages to flatpacks. I'm just saying that your *specific* complaint seems like one that's better solved without flatpacks.<br> </div> Fri, 23 Sep 2016 23:44:56 +0000 GNOME 3.22 released https://lwn.net/Articles/701696/ https://lwn.net/Articles/701696/ pizza <div class="FormattedComment"> <font class="QuotedText">&gt; I call bullshit. You're trusting upstream but you're not auditing upstream. GNOME as well as most upstreams can be compromised. There's enough ability to sneak in a fix which actually causes a security problem.</font><br> <p> Flatpaks don't really help in that scenario. If your distro is compromised, then flatpaks won't matter. If your package upstream is compromised, then their generated flatpaks aren't trustable either.<br> </div> Fri, 23 Sep 2016 11:35:51 +0000 GNOME 3.22 released https://lwn.net/Articles/701679/ https://lwn.net/Articles/701679/ ovitters <div class="FormattedComment"> <font class="QuotedText">&gt; Flatpaks offer something where there is no want.</font><br> <p> Within Mageia we got loads of backport requests we didn't fulfill at all. There's now a backport release channel, but still lots of backports are NOT being made available. I already mentioned this use case in my original reply btw. Responding with a "there is no want" without providing some additional thought makes me think you didn't read what I wrote.<br> <p> <font class="QuotedText">&gt; *is* trusted</font><br> <p> I call bullshit. You're trusting upstream but you're not auditing upstream. GNOME as well as most upstreams can be compromised. There's enough ability to sneak in a fix which actually causes a security problem. Further, you're doing on and on about Debian. What about the other distributions I already talked about. If you do NOT read what I write, don't bother to respond as it is pointless!<br> <p> <font class="QuotedText">&gt; It's slackwarisation for dummies.</font><br> <p> It's partly intended to make things easier. So your criticism is actually a compliment. Regarding the "residual quality assurance" it again seems you didn't read much about Flatpak.<br> </div> Fri, 23 Sep 2016 08:13:32 +0000 GNOME 3.22 released https://lwn.net/Articles/701638/ https://lwn.net/Articles/701638/ johannbg <div class="FormattedComment"> What I'm trying to say is that downstream distribution are doing desktop environments such as Gnome and KDE etc an disservice. <br> <p> At the hand of downstream distribution implementation, upstream desktop environment community's suffer from brand dilution and fragmentation, loss of contribution and feedback ( since individuals contribute to that downstream and see downstream as it's upstream ) inconsistent user experience ( since downstream implement it all in different manner, with different configuration and out of box experience for end users ) a security risk ( due to difference in testing,quality and maintainability in downstream distributions and the shipped desktop environment release and api levels ). <br> <p> It's not without reason desktop environments on Linux remain so unpopular with novice end users and not an viable platform for companies and application developers as can be seen after what two decades of development of the oldest desktop environments out there Gnome and KDE and if desktop environments and the community surrounding them are genuinely interested in succeeding as an viable desktop alternative for end users one of the things they need to do is that they need to start releasing an full blown os themselves. ( You probably know why the Gnome community has not done this already at least I would be amazed if the discussion of it releasing it's own full blown OS has not come up since arguably flatpak is one step in the direction of doing just that and many of it's core developers are doing changes on the core/baseOS leve as well)<br> <p> Since you mentioned Fedora and Gnome in the same context end users have never been able to experience the same consistent desktop experience of Gnome between Fedora releases ( even to this day they wont get it as in it wont be the same between F24 and F25 they wont get the same set of default application being shipped with it as in the previous release, they wont get the same experience of configuring it and they wont get the same location of applications in it etc ) which is partially to blame to how Gnome has been implemented, maintained in Fedora and partially how upstream Gnome is implementing it's upstream changes and maintaining it's own set of applications. Perhaps Mageia has done better in providing that consistency for end users between releases.<br> <p> Ironically you have the history repeating itself with the same fundamental problem and assoicated patterns forming in the android fragmentation in which you can just swapped out distribution for manufactures with the only difference being there is no incentive for manufacturers to push out new updated release on existing devices since their business model isn’t exactly based on extending the life of your smartphone. <br> <p> The end user reaction from the android user base is that it's being fedup of that bullshit which starts with them starting to change device manufactures in the quest for the better android experience ( which is the same reaction that can be seen with distro hoppers ) then results in them either switching to apple ( which is the same end user reaction I have seen with novice end users that tried Fedora and Gnome and it did not work for them as in their response was "I have tried Linux" not I have tried Fedora or I have tried Gnome but Linux and they moved back to Windows or OS-X and did not want to try another Linux distribution or another desktop environment on the same distribution) or suing manufactures [1] ( which would have happened already if Linux distribution could be sued ).<br> <p> If any of the desktop environments would have succeeded it would have been faced with the exact same problem Google is dealing with in android now so it's quite interesting to see what can be learned from Google dealing with the fragmentation problem since the same approach might be applicable if and then when one or more of the desktop environments finally succeeds. <br> <p> 1. <a href="http://www.digitaltrends.com/mobile/dutch-consumer-group-sues-samsung-over-updates/">http://www.digitaltrends.com/mobile/dutch-consumer-group-...</a><br> </div> Thu, 22 Sep 2016 22:46:58 +0000 GNOME 3.22 released https://lwn.net/Articles/701636/ https://lwn.net/Articles/701636/ ovitters <div class="FormattedComment"> Some distributions do provide a GNOME version of their distribution. E.g. Mageia has such an iso. Fedora workstation is the same. No matter if distributions follow this or not, it's good to make it very clear what the state is of the various programs we provide. Saying that some is pretty much beta quality, etc.<br> <p> Let's first check if distributions want to provide a GNOME version of their distribution.<br> </div> Thu, 22 Sep 2016 20:10:04 +0000 GNOME 3.22 released https://lwn.net/Articles/701624/ https://lwn.net/Articles/701624/ bucky <div class="FormattedComment"> There's occasional want.<br> <p> For example, mysql-workbench is unstable on Fedora with every other release (of either). On the occasions where I actually found out why, it was that some library or other was upgraded in the system, and workbench is developed against older ones.<br> <p> For tiny companies like Oracle that have limited manpower, they could perhaps produce a single Flatpak package that is more likely not to crash.<br> </div> Thu, 22 Sep 2016 19:26:44 +0000 GNOME 3.22 released https://lwn.net/Articles/701574/ https://lwn.net/Articles/701574/ stumbles Instead lets bow to Gnomes balkanization. Thu, 22 Sep 2016 18:38:28 +0000 GNOME 3.22 released https://lwn.net/Articles/701552/ https://lwn.net/Articles/701552/ Wol <div class="FormattedComment"> Which is where gentoo excels :-)<br> <p> Of course, that has *other* problems - my main machine is stuck on the now obsolete KDE4, because I suspect when I try to upgrade to 5, it's going to be a load of trouble for which I can't afford the time at present ...<br> <p> Cheers,<br> Wol<br> </div> Thu, 22 Sep 2016 16:01:02 +0000 GNOME 3.22 released https://lwn.net/Articles/701523/ https://lwn.net/Articles/701523/ pizza <div class="FormattedComment"> Nevermind that many classes of applications require relatively up-to-date system software too (eg kernel and graphics stack) -- no matter how much stuff you cram into your container, you're not going to get OpenCL on stock CentOS6.<br> <p> <p> </div> Thu, 22 Sep 2016 13:32:17 +0000 GNOME 3.22 released https://lwn.net/Articles/701520/ https://lwn.net/Articles/701520/ farnz <p>Because I want a new version of (say) <a href="https://krita.org/en/">Krita</a> now, as there's a bug fix or new feature that will massively improve my working day. However, I don't want the time saved by the bug fix or feature to be eaten up by changes to my web browser, mail client, kernel, or other component that I don't want to update right now. <p>Of course, I do eventually want better versions of everything - but, right here, right now, what I want is <em>just</em> the latest version of one of the 5 to 10 applications I use every day. And note that if (to continue with the example at hand), I want Krita 3.0.1 for the soft proofing feature, I need a build that's under 15 days old - if you built 15 days ago, you're building either an unstable RC, or Krita 3.0.0. Equally, though, I don't want a kernel that's that up to date - one that's a few months old supports my hardware and software just fine, and I'd rather have the bugs I know about than new bugs. Thu, 22 Sep 2016 13:26:30 +0000 GNOME 3.22 released https://lwn.net/Articles/701519/ https://lwn.net/Articles/701519/ pabs <div class="FormattedComment"> All the stuff that Debian does is great, but we still need sandboxing of applications, be they Debian packaged applications or Flatpak packaged ones. There are bound to be bugs in various Internet connected applications that boundaries like containers or Qubes-style VMs can help to limit the effects of.<br> </div> Thu, 22 Sep 2016 13:20:03 +0000 GNOME 3.22 released https://lwn.net/Articles/701503/ https://lwn.net/Articles/701503/ smokeing <div class="FormattedComment"> Flatpaks offer something where there is no want.<br> <p> Specifically, sandboxing (as in chrome) is one mechanism to run untrusted code. Anything pulled from random sites is inherently untrusted. Anything installed from Debian (or any extant distro with a "business sense") *is* trusted, because reputation, QA, well-behaved upstream, identifiable maintainers, and consensual engineering decisions taken by people with a track record (well, mostly, given the systemd fallout). With Debian, it's decades worth of all these efforts continually upheld.<br> <p> With Flatpaks, it's Android-style free-for-all of murky installable.. things with residual quality assurance and murky provenance. It's slackwarisation for dummies.<br> </div> Thu, 22 Sep 2016 13:09:24 +0000 GNOME 3.22 released https://lwn.net/Articles/701515/ https://lwn.net/Articles/701515/ rvern Why is updating the entire OS a problem? I <i>want</i> to have newer versions of system software too, not just applications. Why would I want only the newer versions of the applications? Stability, performance, reliability depend in no small part on the system software; that this is not immediately visible by users does not make me want to upgrade the system software any less. Not updating the entire OS is what causes a large set of problems if you ask me. Thu, 22 Sep 2016 12:57:53 +0000 GNOME 3.22 released https://lwn.net/Articles/701504/ https://lwn.net/Articles/701504/ johannbg <div class="FormattedComment"> What I dont understand is why is Gnome emphasizing so much on distributions providing some kind of "Gnome Experience" ?<br> <p> You do realize that desktop environment ( whatever it is called ) can never have any remote chance of succeeding until they themselves start releasing their own OS under their own brand as in GnomeOS, kdeOS etc. in which they are 100% full control of.<br> <p> That's the only way the community surrounding those desktop environment can design, implement, update and maintain and release the desktop environment as the community agreed upon, developed and intended for their target end user base as well as receive feed back directly back up to it's own community. <br> <p> <p> </div> Thu, 22 Sep 2016 12:42:21 +0000 GNOME 3.22 released https://lwn.net/Articles/701478/ https://lwn.net/Articles/701478/ ovitters <div class="FormattedComment"> Flatpak offers an additional choice. Some distributions might do away with their packaging for apps which have a Flatpak (Fedora maybe). Some might not.<br> </div> Thu, 22 Sep 2016 08:18:58 +0000 GNOME 3.22 released https://lwn.net/Articles/701476/ https://lwn.net/Articles/701476/ ovitters <div class="FormattedComment"> We're being clear in what applications created by GNOME aren't at a good quality. Further, it is meant towards distributions which want to provide a GNOME experience. You skipped over a few essential bits while trying to find some reason to complain about GNOME.<br> </div> Thu, 22 Sep 2016 08:16:43 +0000 GNOME 3.22 released https://lwn.net/Articles/701473/ https://lwn.net/Articles/701473/ oldtomas <div class="FormattedComment"> <font class="QuotedText">&gt; You'll pry Debian from my cold, dead hands.</font><br> <p> Thirded.<br> <p> Look, torquay: you always sound as if there weren't any truth beside yours, but there is.<br> <p> Distribution is trust, and I trust the Debian community (and model) much more than any other. I can't change that.<br> <p> I wish you much fun with your flatpaks, but please, don't try to force them upon me. That'd surely cause grief. In exchange I won't try to force dpkg (and the Debian universe) upon you. OK?<br> </div> Thu, 22 Sep 2016 07:00:18 +0000 GNOME 3.22 released https://lwn.net/Articles/701462/ https://lwn.net/Articles/701462/ torquay <ul> <i> If you want to run Linux with Windows-style software distribution, you go right ahead. But I escaped that hell years ago. </i> </ul> <p> Flatpak is far more advanced than the simple Windows-style software distribution. I don't doubt that rpm/deb/etc are useful for low-level OS components, but the processes they represent cause too many problems for installing up-to-date user facing apps. The problems include balkanization, unnecessary man-in-the-middle work, bus factor, non-scalability, distro-centric, no sandboxing, etc. </p> <p> On top of that, you have popular distros (Fedora, Ubuntu, RHEL, ...) which tend to ship only one version of most software that never gets updated (modulo security fixes). In order to upgrade the version, you have to update the entire OS, which causes its own large set of problems. For LTS distro releases this also causes much of the software to be seriously stale. </p> Thu, 22 Sep 2016 05:19:42 +0000 GNOME 3.22 released https://lwn.net/Articles/701459/ https://lwn.net/Articles/701459/ donbarry <div class="FormattedComment"> Indeed. And with their anti-recommendations of what to NOT install, GNOME demonstrates once again their strategy of "simplification" into nonexistence of utility, while providing postmodernist documentations of the correctness of their "flow."<br> <p> I switched to MATE years ago. I've looked out of morbid curiosity with where things are going now and then, and each time it seems the corrections to the obvious pathologies of the present become separate and worse pathologies of the future. Like someone said of PHP, it's become a fractal of bad design -- and yet it's "design driven"!<br> <p> <p> </div> Thu, 22 Sep 2016 04:43:20 +0000