Fedora, systemd, and changes
Significant changes in distributions brought by new releases often make some segment of the users unhappy—sometimes very vocally so. One need only look at the GNOME 3 arguments that have raged here and elsewhere to see one example. But another change that came along with Fedora 15 is getting a similar reaction: systemd. This replacement for the init process has many technical advantages, along with bugs to still be worked out of course, but its inclusion into Fedora 15 has been met with some rather loud complaints. There is merit to some of them certainly, but many seem to fall into the category of "how dare you touch my init".
init is the first user-space process that gets started on a Linux system and is the only user-space process that is started by the kernel—the others are descendants of init. It is responsible for setting up the system for use by other processes: mounting filesystems, starting services, and so on. Traditionally, much of the work has been done using shell scripts, either with the venerable sysvinit or the more recent Upstart, but systemd is on a path to change much of that. Instead of invoking the shell multiple times as part of the boot process, systemd allows much of that work to be done from within the systemd executable itself, which saves a lot of overhead in terms of starting the shell and executing the same basic shell script code multiple times.
But systemd can still use the existing sysv/Upstart-style init scripts, and Fedora 15 makes use of plenty of them. Over time, the plan is to phase those out in favor of configuring those services directly using systemd's configuration files. While systemd seems to be working well for many users, it isn't surprising that a low-level component might not suit everyone's needs. Several recent threads in the fedora-devel mailing list highlight some of those problems, but more than that, they also highlight just how unhappy some people can get with changes of this sort.
Denys
Vlasenko made a reasonable
point—amongst some flames—when he asked about the memory
usage of systemd: "Granted, systemd does a bit more that "typical" init, but I think
using *eleven plus megabytes* of malloced space is a bit much.
" He
went on to list multiple things that systemd does which, in his mind,
should not be part of the init process.
Eleven megabytes of malloced space did seem somewhat excessive, so Adam Jackson dug in to try to learn more. His numbers showed 21M of malloced space, 11M of which was all in 2064-byte allocations. Since malloc uses 16 bytes for its own bookkeeping, it meant that something was doing more than 5000 allocations of 2048-byte chunks. As Michal Schmidt was quick to point out, those come from a known problem with SELinux. Jackson also noted that Vlasenko wasn't really helping his case by not investigating the problem further:
To be clear, this is an observation about how you're presenting your argument. The original post reads mostly as "this looks like it's doing too much, because of these things that I don't understand but I just _know_ they're not necessary, so obviously this is all crap and everyone who's working on it should be ashamed". Even if you were right, that's not a tone of voice that makes people want to listen to you.
Contrary to Vlasenko's prediction that his memory complaints would be ignored by systemd developer Lennart Poettering, it would seem that work started more-or-less immediately on finding ways to improve libselinux. Dan Walsh released an updated version that sped up the loading of the file context information by a factor of four. Discussion also started on ways to reduce the memory usage of libselinux.
But there is an overarching issue here: it is long past time to complain about systemd's architecture or its inclusion into Fedora 15. As Simo Sorce put it:
New features for Fedora are proposed and discussed before a decision is made to include them. Fedora nearly shipped systemd in Fedora 14 before deciding against it in the final stages of the release process. That process was done in the open and the decision was made by the Fedora engineering steering committee (FESCo) in accordance with how Fedora is governed. Perhaps it was the wrong choice for the distribution and its users, but it's a little late to make that argument now.
Fedora users who are upset by the change have multiple choices, of course. Some other distributions are also adopting systemd, but Ubuntu seems unlikely to drop Upstart anytime soon. In addition, Debian will still maintain Upstart and shell init scripts, even as it starts supporting systemd, because it wants to support non-Linux environments and systemd is very Linux-specific. Other distributions may also stick with Upstart (or still be using sysvinit). Another option would be to maintain Upstart and the init scripts as an option in Fedora going forward.
While I have heard multiple complaints (on two continents) about systemd over the last few months, I haven't heard much in the way of strong technical arguments against it. The same is true of GNOME 3, actually. Those arguments may well exist—and even be compelling—but most of what I have heard has been an overall resistance to change. There is certainly nothing wrong with that, and change for change's sake is generally a bad thing, but the systemd and GNOME developers believe that they are making truly useful changes. There are enough other options in the free software universe that it probably makes sense for those who are unhappy to either engage with the development communities to move them in a different direction or to switch to one of the myriad other choices that exist.
Posted Jun 16, 2011 3:37 UTC (Thu)
by ringerc (subscriber, #3071)
[Link] (3 responses)
The shell script boot process and simple, dumb init supplemented by at, cron, autofsd, etc are clunky and painful to monitor and administer. They integrate poorly with each other and with things like directory-based network management. They're annoying to monitor, and don't provide basic facilities like parallel boot with dependency management. They're slow, and waste memory.
Launchd is one of the few things I really like on Mac OS X, and I think systemd is stepping in the same direction but doing it better. Thankfully it's also avoiding the all-too-common trap of slavishly copying Mac OS X (yes, I'm looking at you GNOME 3 and alt-~), instead drawing from lessons learned by upstart, launchd, etc to produce a better design.
It's going to be a bit of a rough transition and I'm really, really not going to enjoy having to wrangle upstart on some systems and systemd on others. I think it's worth it, though, and with luck Ubuntu will move to systemd in a release or two.
Posted Jun 16, 2011 5:07 UTC (Thu)
by tzafrir (subscriber, #11501)
[Link] (2 responses)
Also note that the less things you do in parallel, the less memory you waste. A bit more memory used at system startup and then released is negligable. SystemD actually uses more memory for the init process that runs the whole time (well: a bit more. Nothing much that I really care about, but still).
Nitpicking aside: from the little I looked into SystemD, it seems nice, and and something I want to use.
Posted Jun 17, 2011 1:07 UTC (Fri)
by phess (subscriber, #74656)
[Link] (1 responses)
Posted Jun 17, 2011 1:32 UTC (Fri)
by phess (subscriber, #74656)
[Link]
tzafrir, please please please: it's "systemd", NOT "SystemD".
Posted Jun 16, 2011 10:02 UTC (Thu)
by amtota (guest, #4012)
[Link] (55 responses)
Summary: systemd has made my system boot faster and hasn't caused any problems yet, gnome3 was so bad I *had to* switch to kde.
Posted Jun 16, 2011 14:01 UTC (Thu)
by me@jasonclinton.com (subscriber, #52701)
[Link] (51 responses)
[citation needed]
> gnome-3 blames it on anyone who dares criticise their "design" decisions.
[citation needed]
> gnome 3 seems to have been designed to remove features people were used to (in the name of simplicity) and replace them with eye-candy.
[citation needed]
Posted Jun 16, 2011 14:35 UTC (Thu)
by patrick_g (subscriber, #44470)
[Link] (6 responses)
With Gnome3 how can I install the CPU-freq applet ?
Posted Jun 16, 2011 14:51 UTC (Thu)
by mjg59 (subscriber, #23239)
[Link] (4 responses)
Posted Jun 17, 2011 0:35 UTC (Fri)
by baldridgeec (guest, #55283)
[Link] (1 responses)
Posted Jun 17, 2011 0:50 UTC (Fri)
by mjg59 (subscriber, #23239)
[Link]
Posted Jun 19, 2011 18:49 UTC (Sun)
by nicooo (guest, #69134)
[Link] (1 responses)
Posted Jun 19, 2011 20:56 UTC (Sun)
by mjg59 (subscriber, #23239)
[Link]
Posted Jun 16, 2011 14:53 UTC (Thu)
by me@jasonclinton.com (subscriber, #52701)
[Link]
Posted Jun 16, 2011 20:10 UTC (Thu)
by amtota (guest, #4012)
[Link] (5 responses)
Posted Jun 16, 2011 21:08 UTC (Thu)
by me@jasonclinton.com (subscriber, #52701)
[Link] (4 responses)
Posted Jun 17, 2011 12:51 UTC (Fri)
by nix (subscriber, #2304)
[Link] (3 responses)
Posted Jun 17, 2011 13:08 UTC (Fri)
by me@jasonclinton.com (subscriber, #52701)
[Link] (2 responses)
Posted Jun 17, 2011 22:15 UTC (Fri)
by jrn (subscriber, #64214)
[Link] (1 responses)
Personally I find these threads seem completely petty. Jason, I think it's great that you offer advice when people get stuck for lack of knowing some useful feature. The rest of the comments defending or attacking GNOME 3 seem unproductive except for venting and amusement value --- who cares if your complaints about the platform fell on deaf ears? Code will win out. And who cares if the person complaining used 1000 logical fallacies? They are providing data about how the software is used by a real person, which is a precious thing, no matter how easy it is to attack it.
Posted Jun 23, 2011 14:51 UTC (Thu)
by nix (subscriber, #2304)
[Link]
Posted Jun 21, 2011 2:04 UTC (Tue)
by PaulWay (guest, #45600)
[Link] (1 responses)
Perhaps treating this tangential comment as some great argument against GNOME 3 that needed to be rebutted in full is showing your sensitivity on the subject more than proving how baseless the criticism is?
Perhaps we should just be focussing on systemd here?
Have fun,
Paul
Posted Jun 21, 2011 12:52 UTC (Tue)
by me@jasonclinton.com (subscriber, #52701)
[Link]
Posted Jun 21, 2011 8:29 UTC (Tue)
by marcH (subscriber, #57642)
[Link] (35 responses)
Latest news: LWN comments do not make a scientific, peer-reviewed journal.
If you really cared you would have made at least a small attempt to find the information yourself. It does not sound like you tried (sorry if you did).
Posted Jun 21, 2011 13:00 UTC (Tue)
by me@jasonclinton.com (subscriber, #52701)
[Link] (34 responses)
Posted Jun 21, 2011 13:29 UTC (Tue)
by nye (subscriber, #51576)
[Link] (32 responses)
>> gnome3 removes the alternative
Is it untrue that Gone 2 and Gnome 3 can't be installed at the same time? If they can, it would be worth pointing that out as there are a lot of people who believe otherwise.
The other two claims both seem to be self-evidently true based on wide numbers of reports. Those reports may well be unfairly biased, but given their volume it would be more appropriate for you to *disprove* them. So far Gnome 3 has come across mostly as a huge wankfest and a better form of damage control might be to make it look like you give even the slightest damn about users' concerns, rather than just getting angry at everyone.
Posted Jun 21, 2011 13:52 UTC (Tue)
by me@jasonclinton.com (subscriber, #52701)
[Link] (31 responses)
As for LWN, fedora-devel, and gnome-shell-list, yes, there's a vocal group of people who hate change who are indeed making a lot of noise. All told, it seems to be approximately 50 people. Occasionally, there's a legitimate complaint that we have the opportunity to address but they are largely being drowned out by flaming to the point that--for example--the gnome-shell list is almost unusable.
And no, I'm not going to speculate as to what the GP's actual problems with GNOME 3 are if indeed he has any. If he has an issue he'd like to bring up, I'd be happy to address it.
Posted Jun 21, 2011 14:23 UTC (Tue)
by corbet (editor, #1)
[Link] (30 responses)
I believe that the GNOME 3 griping has gone far enough; I have also come to think that I have probably contributed to it more than was appropriate. GNOME has made its choices and plans to follow them through; users can either go along or switch to something that works better for them. I have chosen to follow along for now, partly in the hope that, as happened with 2.x, things will improve in the subsequent releases, and partly because I believe the GNOME folks are trying to do interesting things. When I read stuff like the above, though, I find myself questioning that decision.
Posted Jun 21, 2011 15:24 UTC (Tue)
by me@jasonclinton.com (subscriber, #52701)
[Link] (29 responses)
All that I can really hope to achieve in a lion's den such as here is to correct factual inaccuracies and call people out on their bias and hyperbole.
I mean, we're (GNOME) trying make computers not suck and the kind of conversation that occurs here as a point of usability is fucking EMACS bindings. That's a difference of culture and no amount of persuasion is going to change that. If that's the kind of thing that gets your blood boiling, we aren't going to make you happy.
We care about users because *we* are users. We want *our* computers--for us, our friends and our families--to not suck. That is where we're going.
Posted Jun 21, 2011 15:52 UTC (Tue)
by corbet (editor, #1)
[Link] (20 responses)
There was a good post in a different thread about how hard those bindings are to support in GNOME, and why it might make sense for them to go away. That comment actually made sense. OTOH, saying that such complaints are not "legitimate" and applying expletives to them suggests contempt toward the users who care about such things. See the difference?
Posted Jun 21, 2011 16:12 UTC (Tue)
by me@jasonclinton.com (subscriber, #52701)
[Link] (19 responses)
Well, no. I don't care about your EMACS key bindings, sorry. But I'm not sure *why* you care at all what I think.
Posted Jun 21, 2011 16:54 UTC (Tue)
by corbet (editor, #1)
[Link] (1 responses)
I didn't even raise that issue, you did. I simply pointed out that it's not surprising that people care. What I'm trying to say is that you should not put down people who care, even if you can't meet their needs. There's a reason why I pointed to Matthew's comment saying that supporting Emacs key bindings may not be possible or worthwhile; he did it without putting down the people who use those bindings. You ignored all that and chose to attack me instead.
Every one of those millions of users you somehow aspire to will have a nit-pick issue somewhere. There's no way you will ever be able to satisfy them all. But you will certainly alienate those users if you tell them that they are out of line to even raise the issue.
Oh well, I'm done.
Posted Jun 21, 2011 17:15 UTC (Tue)
by viro (subscriber, #7872)
[Link]
Posted Jun 21, 2011 19:09 UTC (Tue)
by marcH (subscriber, #57642)
[Link] (14 responses)
Speaking of untrue hyperboles...
> about an issue that isn't even *visible* to the vast majority of people, much less widely known.
Granted, Emacs key bindings concern a minority of people. However this is not any minority. It is the most able minority, the one with the best bug reports and patches, the one that installs and maintains Linux desktops for relatives and friends. The minority that originally spoon-fed GNOME and donated limited but free QA and support, all the tedious things that cost an eye to proprietary companies. You seem to have recently become rich enough to turn your back on them and their "fucking" Emacs key bindings? Good for you.
Posted Jun 21, 2011 21:36 UTC (Tue)
by paulj (subscriber, #341)
[Link] (13 responses)
It could be that those weren't the barriers. It could be that actually what has kept mainstream users from GNOME desktops is more the lack of application compatibility/interoperability (this is what my family tell me anyway). In which case, no amount of UX design is going to help.
If the applications is the main reason, I'd perhaps blame some of that on the constantly shifting environment that GNOME/Linux offers. Further, the rewrites of sub-systems in Linux, that cause endless, differing regressions in every major release of distros, have chased away a lot of the userbase. Those are people who once advocated Linux on the desktop, but now advise others against it. Many people in my LUG now seem to have shifted to OS-X over the last decade, others use Windows. Even Linux distro employees often seem to be using OS-X for their daily desktop...
Anyway. We're only the stupid users, the holdouts.
Posted Jun 21, 2011 21:55 UTC (Tue)
by mjg59 (subscriber, #23239)
[Link] (12 responses)
Posted Jun 22, 2011 6:03 UTC (Wed)
by paulj (subscriber, #341)
[Link] (11 responses)
However, my comment wasn't about emacs keybindings particularly.
Posted Jun 22, 2011 14:39 UTC (Wed)
by mjg59 (subscriber, #23239)
[Link] (10 responses)
Posted Jun 22, 2011 15:08 UTC (Wed)
by paulj (subscriber, #341)
[Link] (9 responses)
The comment was actually about which users GNOME is trying to appeal to, and possible other reasons why they have failed to attract that group.
(As an aside, the few younger Linux desktop users that I know of seem all to have eschewed GNOME entirely and are using much more esoteric setups (xmonad, etc) than the few older, grumpier Linux desktop users here - who use GNOME. Anecdotal / very limited data-set though, of course).
Posted Jun 22, 2011 15:28 UTC (Wed)
by mjg59 (subscriber, #23239)
[Link] (8 responses)
Posted Jun 22, 2011 15:52 UTC (Wed)
by paulj (subscriber, #341)
[Link] (5 responses)
Posted Jun 22, 2011 16:25 UTC (Wed)
by mjg59 (subscriber, #23239)
[Link] (4 responses)
Modern desktops are pretty different to the traditional X11 environment. For a bunch of reasons (including an increased number of situations in which applications need passwords, websites that pop up timeout windows on idle sessions, a wider range of things that provide notifications) it's now far more likely that a window of some description may appear under your mouse. It's not clear what the right thing to do in that situation is, and most of the traditional approaches can result in awkwardnesses like you suddenly typing your password into an IM client. Getting this right is important, and if it's something you provide in a visible UI then you have to be sure that you can get it right - exposed options shouldn't work for the 99% case and be pathological in the 1% case.
Having said that, while I don't think focus-follows-mouse is a high priority for people right now, I'd be kind of surprised if we don't see a certain degree of what happened between Gnome 2.0 and 2.2. Some features that vanished were reworked to an acceptable level and reappeared again. In this case I think if someone were willing to work through the problems inherent in focus-follows-mouse and come up with an implementation that doesn't utterly fail in awkward corner cases then that would stand a good chance of integration in some form. Again, not speaking for the developers or designers or anything.
The panel bar stuff is (as far as I understand it) a deliberate design decision rather than a technical limitation. Having it be modifiable via extensions is intended to indicate that you're deliberately diverging from the intended environment, and as a result things may behave slightly differently and if so that's your problem rather than upstream's problem. But yes, the extensions API is still unfinished. I think the aim is for it to be considered stable in the 3.2 timeframe.
Posted Jun 22, 2011 17:17 UTC (Wed)
by mathstuf (subscriber, #69389)
[Link]
I solve this by having it so that no window can *ever* steal keyboard focus (the exception is if the currently focused window disappears, but with tiling, it is usually pretty apparent from the windows changing for the new geometry). I use sloppy focus and one of the things that I did not like about GNOME 2's (CentOS 5.6) FFM implementation is that it will change focus if the mouse moves over a window some minimum distance (though not if it's stationary) where I'm used to it only happening when crossing into a window across one of its borders and click-to-focus was the better choice. I could see where if the previous behavior was kept, this problem would happen, but if better focus stealing prevention existed (even just "Default", "When different application", or "Never") or sloppy focus were used, it wouldn't be that much of an issue.
Posted Jun 22, 2011 21:26 UTC (Wed)
by paulj (subscriber, #341)
[Link] (2 responses)
Re the panel bar, if that's a temporarily missing feature, does that seem like good release engineering to you? We've seen since this "Ah, but that will be added back later" intent before, e.g. GDM and the configuration tool for it, but I bet there are more examples. It turns out that intentions don't always result in code.
OTOH, if it's a deliberate design decision then I remain confused at how GNOME 3 aims for consistence while effectively encouraging massive code divergence amongst its users!
Posted Jun 22, 2011 21:35 UTC (Wed)
by me@jasonclinton.com (subscriber, #52701)
[Link]
Posted Jun 22, 2011 21:40 UTC (Wed)
by marcH (subscriber, #57642)
[Link]
Aaaahh... the "great GDM rewrite", one of the best free software stories ever. Glad you mentioned it
https://bugzilla.redhat.com/show_bug.cgi?id=433649
Posted Jun 24, 2011 17:29 UTC (Fri)
by rwmj (subscriber, #5474)
[Link] (1 responses)
Loss of emacs keybindings is very annoying, and sadly it's something that my move to XFCE can't fix.
Posted Jun 24, 2011 18:10 UTC (Fri)
by mjg59 (subscriber, #23239)
[Link]
Posted Jun 21, 2011 20:46 UTC (Tue)
by Julie (guest, #66693)
[Link]
We (GNOME) are trying to take in to account the user experiences of tens of thousands of users (hopefully millions if we deserve to be rewarded with them) and you want me to beg for your forgiveness about an issue that isn't even *visible* to the vast majority of people, much less widely known.
Well, no. I don't care about your EMACS key bindings, sorry. Speaking as someone with previously no preference for any particular DE, if you _are_ representative of the Gnome community as a whole, your attitude just decided which one I'm going to avoid. And I don't even _use_ Emacs.
Posted Jun 22, 2011 13:41 UTC (Wed)
by roblucid (guest, #48964)
[Link]
The GUI is critical to user experience, they can spend 20-99+ hours a week in it. Just like KDE 4.0 which was obviously not ready for end users, but got shipped anyway without (except in openSUSE 11.1) ability to switch to/from KDE 3 desktop. Of course it made ppl confused and angry, they felt threatened; because their work environment sucked!
Compare with Linus's attitude, of avoiding breakage to user land.
Posted Jun 22, 2011 6:50 UTC (Wed)
by elanthis (guest, #6227)
[Link] (7 responses)
Many of us GNOME 3 opponents do not oppose change. I've been advocating some very large changes to GNOME for years, several of which finally happened for GNOME 3 (after I was told they'd be "too disruptive" years ago). Like simplifying the way overly complex and overly configurable panel worked (this is the one thing that really sucked about GNOME 2, in my personal opinion) and focusing on just delivering a panel that actually worked out of the box the way the majority of people would expect it to. :)
I also had zero problems changing my entire OS and workflow from Linux + GNOME 2 to primarily using Windows 7 (which believe me, was a much much much bigger change than switching to GNOME 3, especially as I'm a developer and MSVC is utterly different in pretty much every conceivable way than the Vim-based workflow I'd become been used to over the last 11 years).
Change is not the problem. Not at all. Change for the better is welcome. Change that makes things worse is not. Even if some of the change _is_ better, the thing is an experience is only ever as good as its worst part. If you make GNOME 10x better in almost every way but then make it hellaciously obnoxious to do one particular task that a user has to do frequently then the final opinion the user will have of the desktop experience is that it's hellaciously obnoxious.
Why did I make that change to Windows 7? Because GNOME 3 _does not work_ for me. It is fundamentally misdesigned in a few key spots, the fixing of which would essentially turn GNOME 3 into a very different experience. Those spots would be small compared to the overall whole that is GNOME 3, but they're the ones that get run into over and over and over during the work day. I'm also -- as a developer -- vehemently opposed to the "do it all over from scratch" mentality that led to abandoning gnome-panel and Metacity for the new shell toolkit and Mutter, but honestly I'm just as opposed to quite a few technical choices in Windows 7 but got over them, because they're not being compounded by a UI that gets in my way rather than helping me get work done.
GNOME 3 has physical usability problems. I've complained about the hot corner numerous times. I trigger that thing constantly by mistake. The hot corner feature is almost the very first thing I turn off in Windows 7 for the same reason. Only, I _can't_ turn it off in GNOME 3. Because it's apparently assumed that I'm _wrong_ for having difficulty controlling a mouse cursor with high levels of accuracy all the time. (As a side note, this is not any kind of disability, more an environment thing combined with the need to use a class of apps that have yet to invent good trackpad/touchscreen controls.) The applications popup thing also causes a LOT of extra wrist movement as I have to go all the way to the upper left corner to open up the menu, the click an Applications button that doesn't look like a button at all, and then often move all the way down to select an application. On a 30" screen. Repeatedly. It sucks to have to do that. A menu like every other OS aside from those intended for tiny mobile devices offers the concept of pointer locality, which means I only have to move the mouse a little distance instead of nearly 30". (Making those menus bigger and hence with larger target areas is good; XFCE's and the old gnome-panel applications menu were absolutely too small. But instead of fixing the problem in GNOME, instead we got a whole new UI design that just trades it for a new problem that wasn't there before.)
I've brought these up to you (and others) before, calmly and politely, and you (yes, you personally) just called me an idiot (that exact word) who couldn't deal with change. So good luck with your role as a Marketing Team member. I'm sure ignoring legitimate quantifiable complaints with The Vision and calling those people idiots will do wonders for the GNOME 3 image in the long run.
Posted Jun 22, 2011 12:11 UTC (Wed)
by nye (subscriber, #51576)
[Link] (1 responses)
This. OMG I want this on a t-shirt.
This is actually the most insightful thing I've read this *year*.
Posted Jun 22, 2011 22:06 UTC (Wed)
by neilbrown (subscriber, #359)
[Link]
So if it ends really well, that can soften the memory of the worst part.
However I doubt that would look nearly as good on a T-shirt:
> An experience is only as good as:
:-)
Posted Jun 22, 2011 13:27 UTC (Wed)
by me@jasonclinton.com (subscriber, #52701)
[Link] (3 responses)
Posted Jun 22, 2011 15:51 UTC (Wed)
by marcH (subscriber, #57642)
[Link] (1 responses)
> No, not really. [references to angry posts on LWN]
So... GNOME is using LWN as a bug tracker after all?
Posted Jun 22, 2011 15:54 UTC (Wed)
by me@jasonclinton.com (subscriber, #52701)
[Link]
Posted Jun 22, 2011 22:01 UTC (Wed)
by elanthis (guest, #6227)
[Link]
Why not? There was nothing bad with that language. I made some suggestions that would have a rather large impact on the UX and the codebase and the powers that be said it would be too much work for changes they didn't believe were desirable by the users of the project. It was a reasonable response. One I disagreed with, but not one that gave me any reason to be upset. I'm now just a little amused by it, not angry, because it's ironic in hindsight. :) Believe me, simplifying the panel is my favorite part of GNOME Shell. I just now believe (very, very strongly) that it has been grossly over-simplified to the point of uselessness.
Maybe GNOME today is all about spending time debating and considering every major change that every person ever suggests... but it shouldn't be.
> Just four days ago you claimed that you were an XFCE user. And nary three months ago you claimed you were forking gnome-panel after which two GNOME developers bent over backwards to kindly offer you a git account to continue to do so.
Turns out that some people do own more than one computer, and do dual-boot or use VMs. Crazy, right? :)
In my case, I have a laptop that exclusively runs Windows 7, and is what I use the vast majority of the time. My main desktop at home is dual-boot between Windows 7 and Fedora+XFCE (after I gave up on GNOME). For work, I use Windows 7, and XP when I have to.
I use Linux at times still as I maintain some Fedora packages, like to occasionally use Valgrind on some other code I maintain, and I like to tinker with things (the systemd stuff is really cool and I'm enjoying that, and the Mesa+Gallium work is awesome if slow-going at times). This remaining Linux install will transition to a VM as soon as it becomes possible to run OpenGL 3.2+ dependent code on Linux without requiring Catalyst or the NVIDIA drivers (which don't run in a VM). If I get a Mac, which I'm debating, I may not even bother with the Linux install anymore, and I'll pretty much only ever use Linux as a server from then on. We'll see. Again, that'll be a massive change for me, as I've been a desktop Linux user (exclusively, save for a very brief foray into OS X with a Macbook G4 I ended up not liking) from 1999 until 2009.
> Metacity and Mutter share about 95% of their code. Mutter is a fork; it is not a rewrite. GNOME Panel could not be modified to use a scene graph.
It rewrote parts of it, namely the compositor. Instead of taking Metacity and architecting the compositor bit to support no compositor, RENDER compsoiting, and GL compositing, the old compositor was just left to rot in Metacity and rewritten to use Clutter in Mutter. So now we have on GNOME the choice of a GL-only compositor or a 1990's desktop experience. (I filed bugs on Metacity's compositor, they were closed as WONTFIX.)
And of course the plugin system was not added to Metacity. So even though the new shell does absolutely nothing in any way that requires GL, you can only get the shell experience on Mutter, because the plugins for the overview modes only works in Mutter. So again you're stuck with maintaining gnome-panel despite opting to rewrite it from scratch solely because of some really bad architecture decisions. At risk of starting a completely different flamewar: KDE got this right, why didn't GNOME?
The gnome-shell was a rewritten panel. It doesn't do anything the gnome-panel doesn't save for some very simplistic overview layouts (which certainly are doable in Cairo). It does quite a bit less, in fact. What exactly did it need this rewrite for? Why does there need to be a whole new toolkit that looks (and occasionally behaves) differently than GTK, and is that justification for the duplication of code and extra maintenance required? If the sole desire was for some smooth transitions and such in the menus, why not bring GTK into the 21st century and add support for it there? I'm sure regular applications would love to have a more modern and pretty toolkit, too, without needing to rewrite from scratch for Clutter+JavaScript.
I'm not sure why I keep asking these questions. This must be the 20th time, and I never get an answer that has any kind of actual technical justification to it of any kind. At best I get links to old half-thought-out mailing list posts that can be summed up as "it would be _cool_ to do this instead, so let's do it!" Ah, FOSS.
> No, not really. [1] [2] [3]
Yes, it's amazing how much I dislike the project now after it proved how its new frontmen deal with people who dare to voice complaint. I don't start out angry and venomous, but it turns out I'm kind of a jerk when people insult me. Guess I'm not a big enough man to just turn the other cheek.
> You have me confused with someone else.
Unless someone else was posting with your full name, I don't think I do. This was some time ago, last summer I believe.
Posted Jul 3, 2011 21:52 UTC (Sun)
by bronson (subscriber, #4806)
[Link]
cd ~/.local/share/gnome-shell/extensions
Press ALT+F2 , type 'r' and press enter. Gnome Shell will restart and the hotspot will leave you in peace.
I'm really glad Herman Boomsma wrote this extension.
Posted Jun 21, 2011 19:18 UTC (Tue)
by marcH (subscriber, #57642)
[Link]
Please explain why the hell would anyone waste his time posting factually untrue hyperboles lost in an LWN comment. I do not think Apple or Microsoft or Google or anyone else fears GNOME 3 that much already. I do not think they care about the LWN public either.
The only reason why so many people here express their passionate but honest opinions is simply because they *care* much.
Posted Jun 17, 2011 15:11 UTC (Fri)
by southey (guest, #9466)
[Link] (2 responses)
Posted Jun 17, 2011 15:28 UTC (Fri)
by jake (editor, #205)
[Link] (1 responses)
I didn't say that there aren't any such arguments, just that I hadn't heard them.
jake
Posted Jun 17, 2011 20:37 UTC (Fri)
by southey (guest, #9466)
[Link]
Posted Jun 16, 2011 12:14 UTC (Thu)
by quintesse (guest, #14569)
[Link]
Posted Jun 16, 2011 14:52 UTC (Thu)
by cdmiller (guest, #2813)
[Link] (10 responses)
"it is long past time to complain about systemd's architecture"
Actually, the architecture is the issue and the complaints will keep coming (as long as folks have time to chime in occasionally). Vlasenko and others have very valid points about a simple stable init vs a do everything init plus service manager plus kitchen sink. Systemd will have to mature and gain a proven track record before it becomes accepted by experienced systems administrators, among whom stability and uptime is more valuable than whiz bang features and faster boot times.
Posted Jun 16, 2011 15:12 UTC (Thu)
by jake (editor, #205)
[Link] (3 responses)
who presumably won't be running Fedora to begin with ... unless they like being required to upgrade every 13 months or so ...
i don't doubt that the complaints will keep coming, but they seem a bit misguided at this point ... Fedora has made its choice, endlessly arguing that it shouldn't have made that choice (or that systemd is flawed at the architecture level) is fairly pointless it seems to me ... either proposing that the change be reverted for Fedora 16 or working on fixing the actual problems that are being found would seem like the right way forward at this point.
jake
Posted Jun 16, 2011 15:31 UTC (Thu)
by vonbrand (subscriber, #4458)
[Link] (2 responses)
I for one suspect SystemD will be at least an alternative in RHEL 7, and also in the next SLES...
Posted Jun 17, 2011 12:56 UTC (Fri)
by wookey (guest, #5501)
[Link] (1 responses)
Posted Jun 18, 2011 8:24 UTC (Sat)
by tzafrir (subscriber, #11501)
[Link]
Posted Jun 16, 2011 15:29 UTC (Thu)
by vonbrand (subscriber, #4458)
[Link] (5 responses)
I'm sorry, but the whole point of OTOH, when my netbook has 2GiB RAM, a few dozen MiB "wasted" is not that big a deal.
Posted Jun 16, 2011 20:45 UTC (Thu)
by jond (subscriber, #37669)
[Link] (4 responses)
Posted Jun 18, 2011 22:18 UTC (Sat)
by elanthis (guest, #6227)
[Link] (3 responses)
Posted Jun 20, 2011 3:12 UTC (Mon)
by ThinkRob (guest, #64513)
[Link] (2 responses)
Could I afford a VPS with 128 MB? Yes. Do I want to pay my provider another few bucks a month just because the development world takes the attitude of "everyone can spare a few more MB for *my* app"? No.
Posted Jun 20, 2011 19:13 UTC (Mon)
by intgr (subscriber, #39733)
[Link] (1 responses)
Great news then, systemd revives the long lost Unix socket activation technique. Your daemons don't have to stick around all the time, wasting memory even when nobody is connected, but are only launched when necessary.
When the SELinux bug gets crushed, your system will end up using *less* memory with systemd. :)
Posted Jun 20, 2011 22:51 UTC (Mon)
by mathstuf (subscriber, #69389)
[Link]
I can attest to this. My x86_64 SELinux-disabled Rawhide desktop boots with ~160MB of RAM used. With F14, it used at least 350MB. Of course, my 32bit installs similar to the F14 setup booted with less than 90MB used (and rebooted in < 45s on a 900MHz Atom). Still not sure what was causing the huge discrepancy between 32 and 64 bit.
Posted Jun 16, 2011 16:21 UTC (Thu)
by intgr (subscriber, #39733)
[Link]
For the record, systemd without SELinux here uses only 3.5 MB (RES=3524). That seems pretty reasonable even for embedded systems. I'm running systemd on Arch Linux.
Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
systemd comparison with GNOME 3 is slightly unfair
* systemd has been an alternative for almost a year, on release gnome3 removes the alternative (the old style desktop is now pretty much unusable)
* systemd developers seem to listen to criticism and deal with it by trying to solve the problems, gnome-3 blames it on anyone who dares criticise their "design" decisions.
* systemd was designed to bring measurable improvements to the init system (faster, leaner, etc) - and does achieve its goals, gnome 3 seems to have been designed to remove features people were used to (in the name of simplicity) and replace them with eye-candy.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
>>[citation needed]
The fan of my old laptop is broken and for me it's critical to be able to change the frequency of my CPU in case of hot temperature.
With Gnome2.x it was just a "right-click and choose the applet" but with Gnome3 I fail to see how I can add this applet.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
>> gnome-3 blames it on anyone who dares criticise their "design" decisions.systemd comparison with GNOME 3 is slightly unfair
>[citation needed]
Oh the irony!
Looks like you could learn a thing or two from the way the systemd devs dealt with criticism.
>> gnome3 removes the alternative
>[citation needed]
If you had not selectively edited my post, the answer (mine) was there for all to see. The gist of it: missing features and crashes.
>> gnome 3 seems to have been designed to remove features people were used to (in the name of simplicity) and replace them with eye-candy.
>[citation needed]
oh my, the list is far too long, but "google is your friend" and many have already been listed here on LWN so you won't have to go too far (and there's a good chance you've already read many of them - so again, why the denial?)
OK, if I have to give juse one, here's a really famous one (it's amazing - and probably not a coincidence that this has happened again):
"Right Click" (see bash.org #462310)
Look around, there are many people who have been long time gnome users who aren't happy at all. You can accept this fact and deal with it, or deny it.
We are talking about devs would may have ideas, insights, experience and even coding skills that can help improve gnome, listen to their grievances with a more conciliatory stance (yes, criticism is hard to take when you have dedicated a lot of time on something) and gnome will be better for it. Or just continue pretending gnome knows better and that "The Emperor's New Clothes" are sublime.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
Jason, do you realize how this kind of language sounds? You have taken people who feel that their workflows have been made significantly worse, disparaged their numbers, told them that they just hate change, and that their complaints are not "legitimate."
Is that really your view of your users?
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
Key bindings matter - they are a big part of one's interaction with the computer. If we want a desktop that doesn't suck, then things like integration with the editor (some of us) use will be important. Of course some people are concerned with "fucking EMACS bindings." It affects their working days in an important way.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
Please point out where I asked you to beg for forgiveness? [citation needed] as you would say.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
Re the panel bar, if that's a temporarily missing feature, does that seem like good release engineering to you? We've seen since this "Ah, but that will be added back later" intent before, e.g. GDM and the configuration tool for it, but I bet there are more examples. It turns out that intentions don't always result in code.
It's the later case. Like people who install Firefox extensions, installing a GNOME Shell extension should be the kind of activity that is well understood as being essentially unlimited in its power to change everything, is the kind of activity which can not be accidentally activated, and has the benefit of being undo-able by simply disabling the extension in the UI for extensions. At least, that's the broad plan. We'll see what makes it in for 3.2.
OTOH, if it's a deliberate design decision then I remain confused at how GNOME 3 aims for consistence while effectively encouraging massive code divergence amongst its users!systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
Gnome-free
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
> Worst part + Last part
systemd comparison with GNOME 3 is slightly unfair
I've been advocating some very large changes to GNOME for years, several of which finally happened for GNOME 3 (after I was told they'd be "too disruptive" years ago)
That's not the kind of language that GNOME contributors use to discuss changes--or at least I haven't seen any conversations of that type since 2006 when I got involved--are you sure you got that impression from an actual GNOME contributor or was it from someone on a forum somewhere?
I also had zero problems changing my entire OS and workflow from Linux + GNOME 2 to primarily using Windows 7
Just four days ago you claimed that you were an XFCE user. And nary three months ago you claimed you were forking gnome-panel after which two GNOME developers bent over backwards to kindly offer you a git account to continue to do so.
Why did I make that change to Windows 7? Because GNOME 3 _does not work_ for me.
Despite the confusion about what you're using, more power to you. I'm happy that you have something that works for you.
I'm also -- as a developer -- vehemently opposed to the "do it all over from scratch" mentality that led to abandoning gnome-panel and Metacity for the new shell toolkit and Mutter,
Metacity and Mutter share about 95% of their code. Mutter is a fork; it is not a rewrite. GNOME Panel could not be modified to use a scene graph.
I've brought these up to you (and others) before, calmly and politely, and
No, not really. [1] [2] [3]
you (yes, you personally) just called me an idiot (that exact word) who couldn't deal with change. So good luck with your role as a Marketing Team member. I'm sure ignoring legitimate quantifiable complaints with The Vision and calling those people idiots will do wonders for the GNOME 3 image in the long run.
You have me confused with someone else. I would appreciate it if you would make a small amount of effort verifying who it is that you're making a personal attack on before you do so.
I don't think you are an idiot and even if I didn't I would say so. But you do take up an inordinate amount of our time with your long rambling rants in which you make specious claims about yourself and your interaction with GNOME. I think we'd all appreciate it if you increased your signal-to-noise ratio a little more.
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
mkdir activitieshotspotdisabler@harmus.gmail.com
cd activitieshotspotdisabler@harmus.gmail.com
wget https://raw.github.com/hermanus/gnome-shell-extensions/ma...
wget https://raw.github.com/hermanus/gnome-shell-extensions/ma...
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
systemd comparison with GNOME 3 is slightly unfair
> gnome 3.
systemd comparison with GNOME 3 is slightly unfair
Fedora, systemd, and changes
But I must say I love the amount of information systemd gives you (a bit too much maybe for the first couple of times until you get used to what everything means).
Fedora, systemd, and changes
Fedora, systemd, and changes
> becomes accepted by experienced systems administrators
Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
init is managing the whole show. And the "simple, stable" init we are accustomed to by now just doesn't cut it. The huge mess of repetitive scripts, which can't do much more than try to start the service and hope for the best because the dependency information just isn't available worked (sort of) in relatively stable environments with an experienced sysadmin on hand to iron out the inevitable wrinkles. That world might linger on in the datacenter, but to really get elsewhere a fundamental reachitecting is overdue. I believe systemd is that (or a large step in the right direction).Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
Fedora, systemd, and changes
> /family www server, FTP server for off-site backups, other misc tasks
Fedora, systemd, and changes
Fedora, systemd, and changes
> Jackson dug in to try to learn more. His numbers showed 21M of malloced
> space, 11M of which was all in 2064-byte allocations.