Submitted by Something’s – once again – brewing within the GNOME project. While a mere suggestion for now, and by no means any form of official policy, influential voices within the GNOME project are arguing that GNOME should become a full-fledged Linux-based operating system, and that the desktop environment should drop support for other operating systems such as Solaris and the BSDs. I have a feeling this isn’t going to go down well with many of our readers.
Let’s make the core issue clear first. Lennart Poettering, creator of systemd, an init replacement daemon for Linux, is proposing systemd as an external dependency for GNOME-Shell in GNOME 3.2. Since systemd has Linux as a dependency and won’t be ported to other operating systems (which would be a very difficult undertaking anyway due to its Linux-specific nature), it would effectively make GNOME a Linux-specific desktop environment.
Obviously, not everyone is happy with this idea. Debian, for instance, also has versions using the FreeBSD or HURD kernel, and on these versions, GNOME would no longer be able to run. Other than these niche versions of Debian, it would also sound the death knell for GNOME on Solaris and the BSDs.
Jon McCann, Red Hat hacker and the main driving force behind GNOME-Shell, takes the idea even further. “The future of GNOME is as a Linux based OS. It is harmful to pretend that you are writing the OS core to work on any number of different kernels, user space subsystem combinations, and core libraries,” he states, “That said, there may be value in defining an application development platform or SDK that exposes higher level, more consistent, and coherent API. But that is a separate issue from how we write core GNOME components like the System Settings.”
“It is free software and people are free to port GNOME to any other architecture or try to exchange kernels or whatever. But that is silly for us to worry about,” he continues, “Kernels just aren’t that interesting. Linux isn’t an OS. Now it is our job to try to build one – finally. Let’s do it. I think the time has come for GNOME to embrace Linux a bit more boldly.”
Of course, and I can’t stress this enough, it’s a mere suggestion at this point, and by no means any official policy or whatever. The fact of the matter is this, however: if systemd becomes an external dependency for GNOME-Shell, this is effectively what is happening anyway. GNOME-Shell requires systemd which requires Linux. For all intents and purposes, it would turn GNOME into a Linux-only project.
Another important figure in the GNOME project, Dave Neary, isn’t particularly enthusiastic about this idea. “This would be a major departure for the project, a big kick in the face for long-term partners like Oracle/Sun, and also for other free
operating systems like BSD,” he argues, “Are you sure you’re not taking the GNOME OS idea a bit far here? Are we going to start depending on kernels carrying specific patch sets next? Or specifying which package management system we expect GNOME distributors to choose?”
The argument in favour of just focussing on Linux exclusively goes like this: why should GNOME be held back by advances in technology simply because Solaris and the BSDs can’t keep up with the fast pace of development in the Linux kernel? I’m sure maintaining this kind of portability is a major pain the bum for the GNOME project, as it can, at times, hold them back in embracing the latest and greatest the Linux community has to offer.
All in all, I’m not really sure where I stand on this. I’m sure our readers will post some interesting insights which could help us undecided folk sway one way on the other.
Very interesting if you ask me.
Yes it is, and people who complain about this are too.
If the BSDs and Solaris would have a lot of upstream developers it wouldn’t even come up, but guess what, they don’t.
FOSS is a do-ocracy. If nobody does the porting/upstream work smaller projects will fail. That’s evolution, baby.
Xorg already did this, and my guess is that Gnome will be next in line. Wayland will be the next dependancy in 2012 or 2013.
Sucks if you like other Unixes besides Linux, but my guess is that most of the people who run something like OpenBSD probably hate the Gnome-Shell anyways.
I`m saying this because the linux kernel is reaching a level of quality, that is universal.
You simply can`t solve a problem better. And so it has won the race. And well, I say that also because it is GPL`d.
Why don`t the others come around, and contribute/develop for linux instead
Edited 2011-05-19 23:58 UTC
Might as well be asking why Linux developers don’t “come around” and get better paid jobs at Microsoft or Apple — a company that has already passed the finishing-line in the race you describe.
🙂
+1
I voted “Insightful”
If you need to chase money, do that.
Because it sucks.
Wow. I’m surprised you didn’t get flamed for that. Not that I think you intended to start a riot or anything, it’s just a very pointed opinion. Saying that here is like walking into a multi-denominational church and screaming that one particular faith is the only way.
I do get what you’re saying though; GNU/Linux is riding the high wave of popularity and for good reason. As to whether it’s the best way to solve a problem, well that depends entirely on the problem. For example, QNX (pre-buyout) existed because nothing else came close to being able to reliably run mission-critical realtime tasks. I wouldn’t in a million years want the Linux kernel at the heart of the computers on space missions, or commercial airliners for that matter. It’s way too complex and has too many potential points of failure.
That said, it is quite the versatile tool, able to be molded and shaped to fit a vast amount of unique applications. Like Sun/Oracle is so proud of saying about Java, it powers everything from servers to in-car-infotainment systems to phones to regular old desktop computers. It’s becoming the OS of choice for tablet devices (I know, Android isn’t *full* Linux but see above about malleability). It seems even Microsoft finally sees it as a true competitor, with the news of an aggressively marketed ARM port of Windows in the wings.
As to the GNOME thing, well I personally think that DE has reached its pinnacle and is on a downward path now. I’ve always preferred GNOME to KDE for the simplicity and general aesthetics, but lately I find myself drawn to Xfce for Linux installs. I’ve always liked that project, coming from years of using Fluxbox and Openbox, and lately it seems to have been improving in both features and speed, while GNOME just gets fatter and slower.
And if GNOME does become strictly a GNU/Linux DE? Well I guess that only makes sense, as much as the two have been tied together in the past anyway.
I say that within the context of desktop use.
There is always a better way to solve a problem.
Definitely for something as complex as a kernel. Definitely for something as rapidly changing as the hardware our os’s run on.
You reach a certain point of usability and you are “there” kindof, with a certain computing experience. Anyone who had an Amiga might agree.. Ubuntu beats that these days though, with a tweaked kernel. (be inspired – http://www.paradoxuncreated.com/tmp/.config39)
Did I say Very smooth opengl in 2.6.39 and from earlier I know it can do about 0.33ms latency for audio.. (on mine atleast). Takes a bit of cpu, so the cpuconscious might want to reduce that though, but it says something about how good the system is.
And it`s still improving yes..
OpenBSD is being used on the desktop, in corporate settings no less:
http://www.insidesocal.com/click/2011/04/a-company-helps-bring-open…
This is just that: sad.
The simple fact that someone thought of making Gnome a Linux-only thing is a failure. What about the whole freedom/open/whatnot righteous ideas that are supposed to be part of the FOSS movement? In my mind, the movement acknowledges niche players and tries to embrace them.
Turning Gnome in a single platform project is exactly what people have been booing Microsoft about. And now, some want to follow in their footsteps? How surprising. After the good and all righteous speak, comes the copying of long-decried behavior.
I’ll keep to myself what I think of the systemd guy.
And this truly made my day:
Sorry but it’s just ludicrous, both the quote and the bold part… really, Gnome would be held back by crumbly-decrepit-uncreative OSes? I don’t know Solaris but **BSDs**? wtf? Humbleness anyone? What is so technically demanding in a stupid WM that just grew fat? Let me just reply to this bag of crap by referring everyone to a brilliant comment, despite it not being totally relevant in this context: http://www.osnews.com/permalink?460396
Edited 2011-05-19 19:25 UTC
Well, as the article states, Gnome is a free project and anybody is welcome to port/fork it, so why don’t you stop bitching and start coding?
See, that’s the thing about open source… all this ‘free love’ and ‘free choice’ shit is all fine and good, but then somebody actually has to sit down and do the work, and that’s where it falls apart sometimes. And when you have all these different variants of something that exist because people can’t agree to do things in a uniform manner, don’t act too surprised if somebody says ‘enough is enough’, and drops support for your personal flavor of choice.
Exactly. If the other OS folks want to fork it or implement the feature set it relies on then more power to them. Don’t expect the main Gnome tree to do workarounds because your OS lacks the infrastructure the project needs to work properly
Funny then how much Linux users always whine about how certain software is Windows or OSX only.
Yes yes, and Windows users always whine about OS X not being sold for regular PCs and Mac users complain about games… People are going to bitch regardless, doesn’t mean the devs need to pander to their every whim.
Apples and oranges. As the other commenters said, everybody is free to port GNOME to their favourite OS. When Photoshop/AutoCAD/whatever become open source (even if only officially supported on Windows/OSX), and Linux people keep whining because Adobe/Autodesk don’t do the dirty work of porting these apps for them, then we talk.
Edit: corrected a few typos.
Edited 2011-05-20 03:47 UTC
Actually I had no idea gnome was doing some conscious effort to support anything else than GNU/linux. I do understand they were even having problems standardizing with Ubuntu an Xorg.
Ultimately it’s the application developers that choose if they want their package to run on one platform or on many. If the mainstream developers are wanting to make their applications depend on the features that each of them like, instead of what is portable, I am afraid GNOME will end forking according to what each linux distribution out there does.
I would’ve thought the common unix-like interface held gnome together but who knows … maybe the resulting forks are even cool in their own way: gnome-redhat, gnome-suse, gnome-ubuntu (and perhaps some other linux’s consolidating or disappearing).
The BSDs and Solaris? .. well, they just won’t disappear.
What I gathered about this is Ubuntu hoped like hell they could get their new Unity desktop to not be a steaming POS before 11.04 was scheduled to drop and instead, with about 2 weeks left realized that was gonna happen in the foreseeable future,(hence why they didn’t just push it back a few months like 6.06) so they quick whipped up a Gnome fallback, that only came out half baked.
They should have had a split and launched with a solid Gnome implementation under the name Gubunutu.
Obviously, you haven’t spotted enough of my previous comments: it turns out that I go against the most common trends here on OSnews. In particular, I don’t drink the FOSS kool-aid. And I do think that software patents aren’t the viruses some have come to believe.
So I guess my comment wasn’t sarcastic enough to not appear as just “bitching”.
And no, although I have an admiration for the BSD teams, it only stems from the release notes, the documentation and the looks of their source code. I’ve fiddled with the FreeBSD and PC-BSD, as well as each and every Ubuntu release. But I’m a Windows user 100% of the time. So no, I’m not the BSD fanboy you wish I were.
That’s what the <sarcasm></sarcasm> tags are for. They should make that a part of the HTML standard
Lol, so the poor guy tried to troll and he couldn’t. ROFL.
)
Amen to that. It doesn’t mean developers should ignore the opinions of non-developers – but ultimately, they’re the ones doing the work.
Yeah, it sounds like that you shot someone in his head, then tell the judge that I did shoot him in his head, but it was him who chose to die.
People, don’t get carried away with opinions when you haven’t encountered this problem. I have. I was working on a file manager written using GTK and I was considering using GIO to be consistent and portable, instead of Linux’s inotify, which is really really much greater. With inotify you can monitor a directory and when something happens you don’t have to read the whole directory to see what changed. Inotify tells you exactly what happened with exactly which file. I can’t describe how much code I had to add to handle keeping things sorted and stuff (in a way I consider very hackish compared to then cleanliness of inotify).
It occured to me that MacOS has a huuuuge advantage from this point of view (let’s compare with Apple so we can avoid the Microsoft flames). It is, in fact, on the opposite side of the spectrum. It has a very clear direction, no need to support zillions of configurations, no cruft, just the code the needs to run to do what you tell it to.
Linux is, obviously, on the other side of the spectrum. If it is to become anything of what people expect it to (year of the Linux Desktop and everything) it needs to sway a little towards the middle at least. You can’t expect to win an car race with all the 16 cars you sponsor (say) out of 22, refusing to accept anything less. You can focus on one and then you can hope for something. You can help the other teams catch up with you but you need to focus on your main one.
Um, actually if you are racing 16 out of the 22 cars participating in a race, you’re likely to win.
By the way, the reason this stupid WM grew so fat is just because it tries to run on so many platforms, and does that supplying only the largest common denominator of technologies. Something that provides what inotify does, obviously, not available on all platforms GNOME supports.
Are You fscking kidding me? FreeBSD can’t keep up? Sorry but we had fully working kernel event driven hardware abstraction layer daemon called DEVD which handles permissions as well and also can ‘do’ actions based upon devices appear/disappear and more, FreeBSD has is since 5.0, what was the year then, 2003? It was already there ready to port anywhere with the MOST PERMISSIBLE LICENSE AVAILABLE, the BSD license, but what Linux idiots did? The first created HAL shit, later udev shit, latele U* shit (Udisk/Upower/U…) that still is not able to do these simple things that FreeBSD’s DEVD did in 2003, and guess what, it still does and it even did not changed since then, not like in Linux where whole ‘ecosystem’ changes from every odd ‘stable’ kernel release
… and FreeBSD can’t keep up? Can’t keep up my ass :ASD
More by phoenix on DEVD:
http://www.osnews.com/thread?459128
Edited 2011-05-19 21:10 UTC
So if BSD had all of this wonderful stuff before Linux did, why is it that it’s the Linux people introducing their new Linux APIs into Gnome? Why not the BSD people?
I am not a FreeBSD developer, nor a GNOME developer, but after using Linux for several years and now FreeBSD (and other systems) for several years I would say that FreeBSD developers does not jerk off too much about their code tellin’ everybody how great it is and how much everybody should embrace it, like Linux developers.
They have the problem to solve, they solved it with DEVD and that’s it.
None from FreeBSD team went into the GNOME mailing lists and shout ‘hey look what great piece of code we have here, make it as a dependency for GNOME, its so great that You should even consider making it a dependency on Linux platforms, who cares that it does not runs there …’ *
* direct reference to what Lennart exacly did
Edited 2011-05-20 05:14 UTC
Yup, they wrote DEVD, and the Linux guys wrote udev as a rough equivalent.
The difference is that the Linux guys went further and wrote things that took advantage of this new feature – and that’s why Gnome works much better on Linux than it does on anything else. Because the Linux guys are interested in the top-to-bottom stack, and the BSD guys don’t seem to be.
Seems that ‘lack of resources’ is not a popular term at Linux circles, well buddy, it unfortunately is at BSD ones.
Lets not compare apples to oranges here, put all that money that You have put into Linux instead into FreeBSD project and watch FreeBSD doing circles around Linux, but Linux has hype which help get that money.
Sounds like BSD people prefers to “satisfice” instead of optimise seeing negative rate of returns. (bad way of putting it but thats the idea).
Both communities do their jobs fine anyways. I’d rather run BSD-based NAS/routers, since there aren’t much technological advancements in the sector, and Linux-based thin-clients and application servers due to the rapid updates. Best of breeds, I guess.
Speaking of comparison. Did we already knew that Open Source development is about helping and volunteering? Donations are always welcome, but I do not believe the idea that because BSD devs lacks the money, they can’t afford to _volunteer_ their precious time to make GNOME/KDE/whatever work better in their system. They keyword is to volunteer, not to wait for any donations.
You mean, Linux devs shouting at BSD mailing lists for their superior code? Hardly.
Wrong approach. The point is if the BSD developers have enough free time, they do not need to go to Linux dev mailing list(KERNEL/low-level stuff) but go instead to GNOME/KDE/whatever DE and bring in patches to make them work/integrate better in their OS. Have they done that?
* direct reference to what Lennart exacly did [/q]
and Lennart suggested just what phoenix suggested at the end of his comment. you just need to read whole talk.
define core abstraction of systemd interfaces which are later accessed by gnome and put them into small separate solution. which gives you:
– linux already has those so it just works
– makes them available for implementation in systems that do not have those options, they just need to implement their underlaying layer. but still much simpler than complete reimplementation in the way linux does them
same method as using abstract methods in programming. what is so strange about that?
IMHO, this move would be great. define abstraction to some feature that focuses on most users and you can focus on implementation that works. but, in OSes where it doesn’t work, they just need to implement requirement in underlaying layer.
instead of looking at this like “everything but Linux will lose support”, try looking from other side. coding by lowest possible denominator is always slow, inefficient and barely works. not to mention code is unreadable with #ifdefs and hacks.
now enter proposed abstract interfaces. all one needs to provide is basic fall back in the start and later reimplement it correctly. my mind tells me, that no one knows better how to do that better than FreeBSD developers alone. note that if this would be done in udev/devd time, FreeBSD would not need whole implementation, they’d only need to satisfy needs from devd
What about GEM/TTM and KMS? What about UFS being the shittiest file system in use right now (if we don’t consider fat16, ofc). ?
Linux guys created HAL and Udev and various init systems because they have the manpower and they can. After all, that is Linux about: reinventing the wheel.
BSD guys doesn’t have the resources to reinvent the wheel and have thousands of software and libraries that to basically the same thing so they have to stick with only one software per task. That makes them a bit jelous.
UFS2 not ufs; Serves quite well here; you mention jealous; quite not the case here; I review the freebsd forums and a linux forum thread titles daily so am not misinformed. As to not having resources, having a larger community and having the potential for a better desktop, correlation does not mean causation (in this case, in my opinion … although having installed Linux distros, 4 of, and I rate them highly… but draw your own conclusions…) …
GEM is being ported to the BSDs and before that KGI was ported. Hmm UFS … do you really know what you are talking about? Ext2/3/4 are just hacks inspired on it and softupdates are REALLY cool!
I’ve used Hal on BSD … is someone really proud of that? Udev is based on BSD’s devfs and the init stuff is also based on stuff that has been available on the BSDs. No innovation there.
I’ll take your word for that, but making the wheel squared won’t help you climb stairs.
Shittiest You say … interesting.
Maximum UFS2 file size is 32 PB while maximum file size for EXT3 is 2 TB only … and UFS2 is shittiest?
I would like to also hear on which properties You have measured that impressive shittiestness …
FreeBSD can also do not have to use UFS at all now, it can boot from ZFS pool … and Linux has nothing even close to what ZFS offers (do not even try to maze unexperienced users with BTRFS).
Linux developers are like monkeys that with infinite amount of time will be able to write works of william shakespeare, so linux developers with infinite amount of time will be able to write PROPERLY DESIGNED and WELL THOUGHT operating system …
Linux developers are like monkeys that with infinite amount of time will be able to write works of william shakespeare
Taking into account genetic drift, one already has.
All nice and dandy… but I fail to see you point:
Having a comparable or even better technical solution is -irrelevant- unless someone from the BSD camp stands up and implements the required support with GNOME (or any other DE for that matter).
This is FOSS 101 – not rocket science: Unless you’re willing to scratch your own itch, no-one will be doing it for you…
– Gilboa
Heh… Do you really expect a decently designed OS to adopt a gross hack like systemd just to run GNOME?
NO WAY!! Keep it linux-only!
I assume that you have come to this conclusion following your vast experience in using systemd under the yet to be released Fedora 15 for… 2 days?
*Sigh*
– Gilboa
Edited 2011-05-20 22:31 UTC
Heh.. that was funny!
I have years of experience managing the alternatives .. and they work just fine. systemd may work very well for a toy system but is very non-unixy and keeps growing new atrocities every day.
Do I need really fast reboots? Not really, even on my desktop I don’t reboot more than twice a day and then it’s the graphic environment, not the init procedure, that takes more time.
It may take some more seconds to boot.. but I’ll stay with KDE … Thanks !!
Wasn’t meant to make you lough, but OK, thanks, I guess.
So was DOS, Windows 3.11 and Linux kernel 2.2.
Like you, I’ve got far too many years of experience in developing application (and servers) that run on Unix and Linux.
And each and every time I had to manually develop some way to handle socket activation and service dependencies.
Now, given my very limited experience with systemd I can’t really comment on it – but I can say that it’s time to kill SysV and friends.
As much as I dislike Windows service manager’s way of doing things (the management and control interfaces are far too complex and the service registration and logging can easily get damaged leaving you with a dead service with no way of knowing what’s wrong) – their dependency tracking is far more suitable to this day-and-age than the ancient runlevel used by SysV.
I don’t think you fully understand what systemd is trying to achieve.
Fast boot is a ***minor*** part of it.
As the uptake of systemd like alternative init systems increases, it’s just a matter of time till KDE faces the same problem GNOME is.
(For the record, I’m mostly using KDE/Fedora 14 and I reboot my machines [including my netbook]) only when there’s a major security fix that requires a new kernel [practically never])
– Gilboa
You clearly have no idea what systemd is all about. It’s not just a neat replacement for init like upstart but a serious attempt to massively parallelize daemon startup, introduce a nanny for the daemons, reduce the clutter from all the bash scripts in /etc/init.d and much more.
systemd is heavily inspired by Apple’s launchd which has replaced everything from rc.d to cron.d in MacOS since 10.4.
Bringing something like launchd to Linux is just awesome and helps to dramatically reduce boot times, increase reliability of daemons and reduce the effort to maintaining init.
systemd is so much better and cleaner than sysvinit and upstart and I am already using it on some of my Debian systems.
With FreeBSD still being stuck to sysinstall as their installers or CVS as their primary revision control system, I am not surprised that they are so fiercely fighting any progress.
I don’t have any problems with the BSD people. But people from Linux community have their right to choose a certain direction development for their software and BSD people shouldn’t be bitching about that, sorry.
Adrian
You probably havent heard about sysinstall replacement called bsdinstall, here have a read:
http://wiki.freebsd.org/BSDInstall
FreeBSD uses SVN for development since 2008.
No, I haven’t. But from the link you provided it seems that bsdinstall is just supposed to be an intermediate solution. It’s surely an improvement, but does it compare with the modern installers like debian-installer or anaconda provided on Linux?
Is it already used by default when I download and install a current stable release of FreeBSD? We’re not talking about stuff which is still in the making but actually being used
.
Which isn’t really an improvement over CVS for a large scale project like FreeBSD. I think most developers agree, that for really large projects which have many branches in the sources and lots of developers, a distributed RCS like git is a much better choice.
Adrian
Yes, its temporary sollution, there is work in progres to provide ‘fully capable’ installer.
Talking ‘modern installer’ I assume that You want a graphical (QT/GTK) installer that a definite newbie will be able to click-out the install process, there is also such installer for FreeBSD, the PC-BSD installer (http://pcbsd.org) written in QT adn it alows You to install PC-BSD and if You do not need all preconfigured stuff, it also allows You to install PLAIN FREEBSD version.
The PC-BSD installer uses pc-sysinstall as a backend (which is set of created POSIX sh(1) shell scripts) and frontend is in QT, the ‘complete’ goal of text/curses based FreeBSD installer is to use that backend for the installer.
Its already in snapshot builds for the 9-CURRENT FreeBSD, the installer is planend for 9.0-RELEASE which will happen somewhere between 2011/06 and 2011/09 probably.
I would like to comment on that, but I am not a developer, the ‘biggest’ things that I write are mostly shell scripts or wrappers so I do not have ANY experience with any large scale code repository.
No, it’s not actually the graphical UI which make debian-installer/anaconda more comfortable but the control logic and the UI design in general. I do not have any problems with text-based installers, in fact, I prefer these.
I just always felt that sysinstall was quite old and quite cumbersome to use. Sure, if you have used it for a while, you will get used to it. But it definitely deserved an overhaul
.
A very similar approach to the one used in debian-installer/anaconda. In fact, you can virtually use these installers in the same way on the text console as you would use them using a Qt/gtk interface.
Sounds promising. I will definitely give it a try once it’s released!
Well, being a hobbyist FOSS developer with my own project and having worked as an administrator on a Linux network at a university as a student’s job, I have used several RCSs and I felt RCS/CVS the most awkward to use and git to be just the easiest and most comfortable RCS. Last but not least, git is incredibly fast which is really a big advantage when working on large projects
.
Adrian
It is mate, its a lot of old code, I do not use sysinstall (or any other installer) to install FreeBSD at all, I use more ‘manual’ way with which I am able to control every bit of how the installation will look: http://forums.freebsd.org/showthread.php?t=12082
As someone noted, PC-BSD has a very cool installer but bsdinstall has other objectives: we still want to have something that can be installed using a serial console and that will work on non PC-platforms.
[/q]
Well it’s a matter of developer’s choice: FreeBSD is developed in a more centralized manner than other projects, and even before SVN, FreeBSD was using a combination of CVS and perforce. I do recall linux didn’t have any RCS for many years: that was a lesson they learned from the BSDs (and indirectly from SCO).
Fortunately MacOS X is a BSD camp and wont follow suit but I wish you luck pushing such brilliant ideas into Windows then.
Which was ported to FreeBSD and lives happily in the ports tree, should anyone really need it.
Nice, I didn’t know that!
So, there should be a chance to write some “glue code” which would make GNOME use launchd on *BSD instead of systemd, shouldn’t there?
Didn’t someone already mention in this thread that Poettering himself said that he would rather see people write a BSD-specific version of systemd instead of putting too much bloat into the systemd sources by porting it to non-Linux platforms?
With that in mind, I think the situation with GNOME depending on systemd is probably far less problematic than many would think =).
Actually, I like FreeBSD for adopting a lot of nice things from the enterprise world like zfs and launchd. BSD is obviously better off in this regard when they don’t have to reinvent the wheel (btrfs, systemd) but can simply use software which is already around.
Adrian
Edited 2011-05-21 22:18 UTC
Nice, I didn’t know that! [/q]
Actually I just crosschecked and it’s not in FreeBSD’s ports tree but the FreeBSD port was folded into the Apple version maintained at MacOSforge.
When the port was made the old APSL was not very friendly but now they have fixed that. It’s free software so PC-BSD, Debian’s GNU k/FreeBSD or even OpenIndiana and linux can use it.
Well, its the direction GNOME has been going anyway!!! Let them!!!
You hit the nail on the head when you said that they have been taking GNOME in the same direction as Windows. If a project insists on implementing the very worst methodologies in the history of computing then maybe that project does not deserve to exist!!!
Here are some examples of abandoning traditional UNIX/Linux methods in favour of more windowsy methods.
They use gconf rather than text formats. What’s wrong with that you ask? When the variables in gconf duplicate those values in OS files then it is a problem. Didn’t you guys pay attention in your first week of DB theory that whenever values are duplicated that they will always get out of sync. Where I ran into this was when I added users as root using useradd rather than the GNOME utilities. The result was that gdm didn’t know about the new users and the new users could’t login to gdm.
What is the most retarded thing about Windows? The friggin registry!!! Just go back to windows!!! Leave my UNIX/Linux alone you morons!!!
Another thing about the GNOME community is the way they can’t friggin wait to get MS tech into Linux and thereby fall into the trap that is being set by MS. I’m of course referring to Mono and C#. Again, if you friggin want windows you know where to go!
I’m finished my rant now!!!
After that rant I can’t see it happening anyway. If it isn’t cross platform then the BSD guys will adopt KDE (Yeah!!!). Can you see Oracle putting KDE on Solaris and GNOME on Oracle Linux? Wouldn’t they want a unified desktop for both systems? Combined with the abortion that GNOME 3 is turning out to be they would either stick with version 2 or fork the project. Hopefully that would kill it – the sooner the better!
Bullshit. There are a few people in the community who have an interest in Mono/C#, but a) they’re a small minority who don’t represent anyone but themselves, b) even that minority has long given up on getting the majority to adopt Mono, and c) there’s not single line of C# code anywhere in core Gnome.
now, go and read proposal again.
systemd proposed unified interfaces to solve duplication, which doesn’t impose new problem, but rather solves one. major reason why gconf came to be is simplification for developer.
another thing you failed to grasp is yours so called text configurations. they would be great … if there would be one standard to read/write them. look httpd.conf, smb.conf and inittab. hell, there is no standard how you read/write them. enter xml and you get configurations unified. gconf is nothing but text files formated by xml spec.
try looking at s-c-t from ximian and read why it never saw the light.
personally, my biggest wish is that every project would adopt xml and also provide converter from old config standard to xml or back. everyone would be happy, those editing manually and coders who write gui tools.
You bark in vain. The reason Gnome devs borrowed strongly from MS is because they have seen more value in things borrowed than in “traditional UNIX/Linux methods”. If you think Windows registry is lame, than you must acknowledge than keeping settings in flat text files spread all over the file system is medieval and dumb.
And yes, C# is better for general use than Python and Bash scripting will ever be.
Well, freedom also means you are free to decide to support only one kernel, or only one OS, and it doesn’t mean creating free software to suit everybody&dog’s wishes.
Me, I don’t really care how they decide. GNOME OS would essentially still be Linux as we know it, just with a specific subset of current technologies, maybe having the change to become better because of the narrower field of concentration for development. They stay, everything will remain as it is now, they move, things will change but we’ll have another option of a tailored Linux OS.
Of course some people might feel left behind, using other kernels, toolchains, etc., but hey, there are still quite a number of desktop enviroments out there. It’s not like they would take away all your options.
And again, you are free to fork the current Gnome and go ahead from there.
Again, remember, it’s free software, and while you can have opinions on how some project goes, in the end, it’s not up to you to demand or decide anything. Be happy that those people provide you with free and nice alternatives and possibilities for choice. Simple as that. On a sidenote, I never could understand regular users revolting when their demands (ridiculous in itself) don’t get attended to.
On a sidenote, I never could understand regular users revolting when their demands (ridiculous in itself) don’t get attended to.
Well, end users are the weakest group decision wise with any technology, so demands is just a tad exaggerated. Let’s call them strongly vocalized wishes.
The only power an end user has with any technology is choosing to use it or not. Whenever there are rumblings around considerable changes, it is the current userbase making clear that the current course of the technology is a strong reason to consider dropping it. It is up to the project to decide if they want to hang on to the current userbase or not.
So revolting is also a severe exaggeration. The “worst” an end user can do is abandon a technology if it doesn’t suit the needs anymore. It’s not like end users have forcefully removed project leaders if they didn’t like their decisions.
While I largely do not disagree with your post and its sentiment we should strive to keep hyperbole to a minimum. GNOME didn’t start as a WM; it didn’t really have its own WM until Metacity came along quite late in the process.
And neither is GNOME just a WM, it’s a desktop which provides not just a WM but also loads of applications, applets, panels etc.
Adrian
Lennat said (here: http://netsplit.com/2010/05/27/dependency-based-event-based-init-da…):
“Did you never get the idea that maybe nobody else is using your event-based design, because it is simply broken? (For the reasons I pointed out in my original blog story.)”
You mean it’s not Linux only already?
Snarkiness aside, they should do this. I really want to see where this goes. There are too many good opportunities for storylines and testing hypotheses here to ignore.
Is the Linux hubris justified, is Red Hat behind it, what could they actually do, how would everyone else respond?
> You mean it’s not Linux only already?
No, it’s not. And if it will be, I’ll switch on using KDE across all my machines, including Linux and OpenIndiana ones.
The point is, it pulls in so much crap from Linux-land that it might as well be Linux only. It might be “cross-platform” now, but only because other people took the time to port it and it’s dependencies. It’s not like the GNOME team designed it to work across multiple platforms.
For proof of it’s Linuxness, look at what happened when Xfce moved from Thunar VFS to the GVFS. Xfce lost BSD support, and now people are trying to port all the crap that GVFS needs to work. That’s not cross-platform. If it was cross-platform, Xfce 4.8 would run on the BSDs out of the box, and people wouldn’t have to waste time porting redundant junk.
Whenever I need a DE for FreeBSD, GNOME doesn’t come up; it’s too Linux centric. Run KDE, Fluxbox, Enlightenment, or until recently, Xfce.
GNOME going Linux-only would only be a loss to the GNOME project, Oracle, and maybe Red Hat, depending on how this turns out.
It’s a discussion, those happens all the time.
Not surprised that it’s Mr “Write-for-Linux-only-and-fsck-everyone-else” Poettering who would suggest such a thing though.
I don’t really see what you’d gain from making GNOME Linux only though or why on earth you’d want systemd to be a hard dependency.
Also, since when is GNOME the OS core? Why would system settings have to be Linux specific?
Pretty much nonsense arguments.
If your desktop environment isn’t an OS and the kernel isn’t an OS then what exactly is an OS? I’m thinking there is more than one legitimate answer to this question.
yes, OS definition is broad. You can’t use a kernel by itself, as you can’t use the programs without a kernel. (I’m over simplifying it and ignoring RTOS, special purpose OS) so we often considered the GNU userland together with Linux kernel a “GNU/Linux OS”.
While this is good for servers, it’s not good at all for desktops. Okay, you can assume you have GLibC and POSIX, but you can’t know what sound server do you have, or which graphics system, ipc and all. Then we used to write generic system that talks to multiple (or none) of these…
This effort can be called “lost” because it could be used to improve the applications or libraries themselves. We used to have lots of half-working applications since developer focus incorrectly switched from “making it useful” to “make it usable anyware”, so yes, it barely provided its core features but would run on a toaster 😀
For a while developer of various systems, not only GNOME but also E17 (which I’m a developer) started to assume some things and don’t care about portability. Assume DBus, assume X11, assume Linux. When people came and complained about not working on other systems, these could be considered… particularly if the complain comes with patches 😉
So I don’t see GNOME doing any radical change other than being clear and stating what was indirected told before: developer focus is on the system they use, Linux.
I see no one could say it better than yourself. I agree.
IMO, an OS is an interface between people, hardware, and software. You can’t get much more precise than that.
To be more clear (and I’m not even GNOME hacker, but e17 one) non-linux systems lacks lots of desktop-interesting features (ie: udev), software stack lags behind (ie: xorg) and mainly lacks developers interested in desktop.
Sure, the second topic (lag) can be fixed if the last point is fixed (more developers). One can argue the first issue (lack of features) is not an issue but different opinions, but why the developers of the Linux front should suffer because the lack on the BSD side? This would be called priority inversion on other situations 🙂
Also Linux itself is not perfect. But given developer interest we’re working on making it better, see the recent commits to aid desktop use cases and even systemd: cgroups, extended vt controls, etc. They don’t exist in BSD, but they didn’t in Linux as well. People worked on getting them ready in Linux, while nobody cares about this in the BSD land. So why bother about BSD in the desktop land?
I don’t really care if GNOME focus only on Linux. They can do what they want. What does annoy me is that this comes after god knows how many years of listening to Linux users and devs endlessly whining about how this and that software does not support Linux. Now, all of a sudden, it’s ok to fsck everyone else because *Linux* needs to evolve. A bit two-faced.
To everyone who is arguing that GNOME would be fsck-ing everyone other then Linux, thus being “two-faced” after all the complaints of Windows/Mac stuff not being available on Linux. You’re making a hypothetical argument just for the sake of the argument.
Since when did any FreeBSD user complain about a Linux only program? They ported it themselves.
Linux users didn’t complain about opensource programs on Windows. Only closed source once. Driver developers don’t complain about hardware without drivers for Linux. They complain about hardware without documentation.
You are comparing two different things.
You can help port gnome to whatever fringe platform you like. It’s free and open source software.
But you can’t do that with proprietary and closed source software. All you can do is urging the developer to release a version for your favourite platform. That’s why some Linux users complain.
That’s not the point. To paraphrase: Why should proprietary vendors be held back by the shortcomings of Linux?
Same argument. Open or closed source is immaterial.
Like someone said: the people who write the code gets to decide.
Edited 2011-05-19 22:43 UTC
I never said they should.
If you are a developer of proprietary software and decide to not release a Linux version (for whatever reson you might have) and there is no free alternative, some Linux users will complain because they have no choice but to dual boot or try to get it to work under wine.
If gnome decides to drop support a platform however they are not “fscking” their users because if there is enough interest in that platform users can just get the sources and maintain the port themselves.
They don’t have to. Just make them open source (as GNOME is) and Linux developers will happily take care of the rest from there, thank you very much.
To not make their applications open source or not is their decision, to support or not support Linux is their decision.
Maybe it’s just me but I have no issue with not all software in the world being open source.
Dude, you contradict yourself. Let gnome guys do whatever they want and stop trolling please
I have no issue either, and I use closed source software myself. My point is that you are comparing different things.
The comment that started this discussion, I think, is the one that said something like this:
With applications like Photoshop, Linux “whiners” don’t have that possibility. And I agree, it is Adobe’s decision, but the point is, IF Adobe open sourced it, I really doubt Linux developers would keep complaining instead of coding. If they did, then (and only then) you could talk about double standards.
But you certainly have an issue with Gnome choosing to focus solely on Linux…
Not really, I dont even use GNOME.
Do you somehow assume that repeating the same text over and over and over again will drive the point home (Thus far I counted 4 different instances) or are you simply trolling?
– Gilboa
There’s a big difference between not including support for systems you don’t personally run and not accepting compatibility patches. Few projects promise to maintain portability absolutely but to refuse to attempt to be portable where it doesn’t hurt makes no sense. Add all the Linux-specific code you like! Throw it all inside #ifdef if you have to and gracefully degrade the experience that people see on e.g. Solaris. But don’t stop supporting Solaris just because, that’s just mean.
They probably should stop supporting Solaris just to thrown Oracle under the bus. To be quite honest, I wouldn’t be surprised if that was part of the strategy.
Not that I see Solaris being anything more then a headless appliance OS going forward.
Why the hell do you think OS X cares about udev? Sorry, but udev is not something OS X is lacking. Not by a long shot.
I never said that OSX lacked udev (or as someone else wrote in this thread: inotify). I just said that instead of writing code for GNOME (KDE, E17, …) considering that udev may or may not be available, we should all focus on the systems we use, and it’s clear the majority is Linux. If there are non-Linux developers, fine, they can “waste” their time reworking the internals to make both work.
To be clear, and using your example: Why MacOSX would care about GNOME? Why GNOME would care about MacOSX?
“Bad GNOME, that’s a BAD GNOME” *Whack* (as the news paper is swapped against the nose of GNOME)
*face palm*
Really quite a dilemma for me, even without this latest discussion. I’ve been using Linux since 1994. Currently, I don’t like Gnome 3 or Unity. My experience with KDE 4 has been bad, even just a month ago with Fedora. Basically, it looks like I’m going have to try out some of the other DE’s. I’ve used XFCE, Blackbox, and many others. I just really like some of the Gnome Applets (like Freq Scaling and the Gnome Weather applet). I may just have to use Mint or one of the others that is sticking with Gnome 2.x. However, that’s probably not a good long-term plan.
It really seems like users are taking a beating these days. First they decide we don’t need menus anymore. Then, the conventional taskbar/start menu is no good. It really seems like change for change’s sake.
Of course, I’m old.
Agree with you. The way Gnome has been heading this past year its becoming completely irrelevant for me.
Time to try out XFCE and E17 again methinks.
I’ve been using Xubuntu for xfce and Bodhi for e17. Both are really good distros.
Xfce has an applet which with run GNOME applets in the system tray, so you wouldn’t miss anything.
Awesome, XFCE, here I come. I used XFCE a few time, before and after it switched to GTK. I know it has a lot of good applets itself, and I love Thunar.
I’ll install it on my Laptop shortly. Thanks for the heads up!
Yeah, I’ve got that same problem, I’ve used almost every MS and Apple OS since DOS and the Apple IIe started using Linux in ’04, I really don’t like any of the new UIs coming out from anyone though, Gnome2 was IMO perfect in terms of usability and simplicity, KDE3 was good as well if using the MacOS mode, but I can’t stand the mutilation of naming by sticking “K” on everything.
What I can’t stand though are the new UIs like Gnome3, KDE4 and Unity, which just seem to be changing things for no real reason that benefits the USER, but I guess they do give the UI devs some reason to exist instead of moving on to something else, like oh, I dunno, building GUIs for things like the OSS GPU drivers that resemble something like AMD Catalyst Control Center/AMD Overdrive, Riva Tuner or the multitude of useful command line only tools out there that would make things easier for the non geek to use Linux and OSS software to it’s fullest potential.
So now the problem is what DE should I try next? I guess I should give E17 a go, since Xfce, Lxde, IceWM, BlackBox etc all have missing features I find very useful in getting things done faster/simpler.
They should work more on developing portable abstraction layers. That way they don’t need to worry about such issues.
Yes, but then there is the question of who will actually port these abstraction layers to other platforms.
For example, a simple thing like interrupt handling can change radically from platform to platform:
http://www.osnews.com/comments/24763
If you do the abstraction only on Linux, then you’ll be abstracting Linux, not a general API.
They did. But it’s the Linux people who did it (HAL, Polkit, upower/udisks), and people from other platforms coming in a few years later. Someone still needs to do the work, and the overwhelming majority of those people are using Linux.
Like this perhaps?
http://phonon.kde.org/
http://solid.kde.org/
Or other OS’es should adopt Linux features in their kernel and distributions. That would benefit a lot of other projects as well
-1, Fat Chance.
This is either the most asinine comment I’ve seen posted in a while, or it’s brilliantly dry humor.
I’m going to go with humor and continue.
They could also use the Linux kernel, the GNU userland, apt, and the Debian repos. It’s brilliant! They wouldn’t have to code anything.
The one horror about opensource support is having to spend hours running autogen or ./config, maybe getting an option wrong or it is missing something but still has to see if the native compile on my x86_64 box running linux has “strcpy” and hundreds of tests which should be obvious, but each seems to take several seconds.
One reason this abomination still exists is to support all the legacy or strange versions out there. There should be no reason for such junk on 99% of the machines – linux isn’t that different (oh, and most autogen/automake/libtool stuff won’t cross compile anyway, so not only is it horrid, it is broken and stays broken).
If it would fix the build system (and I would hope GNU/FSF and their archives follow), then they should be linux only, or at least reduce it to fewer platforms so it could actually build easily with one or two options for platform/cross settings.
If you are going native on linux, “make” should work. Or at worst set a few nonstandard prefix paths. Even cross should work without too much pain.
A lot of the bluetooth and wifi stuff is also broken, or annoying, and I can’t even try to fix things because I have to pull in or configure or do so much other stuff before I can start.
So having complex, slow, and limited builds – so it might theoretically build for other architectures is a bug. If it becomes linux only and becomes simple, fast, and comprehensive, it would be a win, and others could contribute the missing pieces to build on the other platforms.
If they are going to keep autojunk, then why bother limiting it to linux?
You almost got it. But although autotools is showing part of the problem, when you come out of compile-time-detectable library symbols/behavior and go to dynamically changing systems, the situation is orders of magnitude worse.
The idea of a GNOME OS (or E17 OS, KDE OS) would help here, yes. For instance you could just check for that part and assume everything else. “If GNOME/KDE/E17 system then kernel=Linux, libc=GNU, required_libs=present” and yes we could just skip stupid tests in autoconf 🙂
They should definitely go for this. BSD and Solaris are practically dead on the desktop. For desktops (and laptops) the end-game consists of Windows, OS X, and at a distance Linux.
It’s best to put all the effort into making GNOME on Linux as good as possible. This requires as much integration as possible, and should not be held back by the need to support systems with < 0.01% desktop penetration.
“bsd type-commands-into-xterm desktops… are unfavored in the click-icon-popular computing environments that predominate.” I read that as implied, but the vast majority of desktop users haven’t tried typing into xterms yet. … As far as your second paragraph, if it is not broken, why fix it? And have you seen the responses of those who’ve tried gnome2 > gnome3 yet? Some are quite reluctant to switch… I would purely conjecture, a tighter integration would forestall reliability and user satisfaction in the long term.
* do/did users of other OS report errors?
* do/did developers of other OS contribute with code?
* did they get any inspiration of ideas from projects on the other side?
If any of the above questions got a “yes”, it means someone is showing disrespect to honorable ethical values.
I am not putting the fault on gnome or any important/huge community project, for what matters, but I have seen this same pattern been outlined by people that want to rush things out instead of take the needed time to properly understand and tackle specific system deficiencies so that a proper solution can be constructed and the “lets go rewrite everything” syndrome be avoided.
What is so hard about the old isolation/encapsulation mantra? Seems like some people do not know, or want, to work on a bigger community.
Being nice to everybody is nice, but this isn’t about being nice.
If someone gives you something valuable and you give him 1000 bucks for it, then your brother comes and gives you something not as good, although he did his best in providing you with something good, you want to give him 1000 bucks too, which would be nice, but what if you don’t have that money?
And, most importantly, you can help your brother make something better himself, helping him win his own money instead of doing it for him yourself.
Well, I guess I should be a bit clearer. I am not asking anyone to not incorporate the full benefits of system services they want. I am begging them to cooperate on outlining an isolation abstraction layer, publish it on freedesktop.org and go live their lives.
If the specific thing blocks one module to be build, no problem, do not build it. If it is important enough to bother the right guys, for sure, they will plug the hole.
But no, they just do not want to bother themselves and lose their precious time on cooperation, and the whole notification area is a prime example of that.
And to all people that complain about abstraction layers, more often then not, they expose the shortcomings of their own interfaces, and can be engineered to have negligible impact on performance and memory usage. But, oh well, lets rewrite the whole stack again! It is just so fun to do that!
Yup. The only problem is that by doing that will:
-hinder the project
-be unable to use the latest technologies and advancements
-be unable to highly integrate with the rest of the os
If some other guys want the software, they can:
a) port it themselves if they can
b)hire someone to do the work
They will have to either maintain a separate source tree either port systemd because patches won’t make upstream.
Thing is, that’s all nice talk, but someone still needs to do the work. Projects can, and do, support multiple backends. But someone still needs to create and maintain backends for BSD and Solaris and others – and that’s not going to be a Linux developer, if only because they don’t have BSD or Solaris systems.
You sound like someone who subscribes to the BSD coding philosophy.
Doesn’t this mean that in order to run Gnome, one must use systemd?
What if people want to use sysvinit, initng, upstart, runit, or eINIT…?
What if people want to run DOS, what then?
DOS isnt part of Linux. upstart, initng etc all are.
I doubt most users care which init system they’re running. The real question is what if most distros don’t want to switch to using systemd, as it provides no real benefits over some of the other init systems out there and would require all packages’ init scripts to be rewritten?
GNOME should go linux-only, imho, but it should not depend on systemd. Not everyone wants to reinvent the wheel like Poettering seems to enjoy doing. If he wants to integrate systemd support into GNOME, great. If the Ubuntu devs wish to integrate upstart support into GNOME, that’s awesome. If Debian wants to integrate sysvinit support, more power to them. Forcing a hard dependency of systemd is probably one of the worst ideas I’ve heard… but, then again, most of GNOME 3 seems to be one huge bad idea. To those who think forcing systemd on everyone is a good idea, remember what happened when GNOME started depending on Pulseaudio and just how well that’s turned out?
I think the above comment has it pretty much spot on. This isn’t just saying, “Screw other operating systems”, it’s also saying, “Screw Linux distributions that don’t use systemd.” That could be a problem for Debian/Ubuntu/Slackware (I think) and other distros that use other init systems.
Not only that, but the developer has stated firmly that he won’t accept patches into upstream that would improve portability. Which means if someone wants to put in the work to ports things over to another distro or operating system those fixes won’t get applied upstream, making it a moving target. That’s going to be a huge pain for outside porters/developers.
Oh come on, it’s all acceptable sacrifices in order to increase RedHat’s market share.
…What?
Be fair – systemd is a *very* low level project, just one step removed from the kernel, and very heavily tied to kernel interfaces. Trying to make the *code* portable just isn’t feasible, which is why Lennart instead recommends writing API-compatible equivalents for BSD or Solaris.
Did he in fact recommend this? Links?
All I’ve ever heard from him when asked is “I am only interested in Linux and I don’t care about other platforms.” I never heard a more rational response than that.
if you would read mailing list instead of just posting here, you wouldn’t need links. you can read and search your self. but here is a quick review
he stated following:
– systemd proposes interfaces where DE can integrate tightly into system core
– he is prepared to separate interfaces from core in order to ease the portability, but he is not prepared to accept patches for other than linux inside his core branch. interfaces are portable and he is not prepared to work on anything that is not linux.
– every other os should implement their own clean implementation of interfaces to avoid getting implementation riddled with #ifdef. this also means their implementation will be clean and easy to maintain
– all that os needs is to prepare negative fallback and later implement it in correct fashion
this is how i understood his proposal, but i might be wrong. major mistake he made in proposal was subject line. everything else is really great
what bugs me at this topic is that he proposes one thing that would very much save the day when udev/devd fiasco came (instead of udev implementation, only interfaces need to be implemented trough devd), but BSD people didn’t like that one… and guess what, they don’t like this one
Don’t be insulting. Which mailing list? When? If it’s the discussion linked above, sorry. I haven’t had time to get to it yet today and I didn’t expect from the summary that any details on systemd would be involved.
I am referring to his many statements in the past when he has been asked about things like this, both on mailing lists and his blog. A summary of his intent as you understand it is not useful.
You don’t need me to supply links – just try the ones in the article. Lennart repeats the point in almost every reply he’s made on that thread…
If it’s not portable between *Linux* distros it’s a pretty shitty effort.
It´s about freaking time. The people who write the code get to decide. Gnome and KDE are full of workarounds to make sure that things continue to work on a huge number of platforms that no one really uses.
How many number of OpenBSD users use it as their desktop OS? I know some that build firewalls or file servers with it, none that use it as their workstation. For that they use Linux or OS X.
Tighter integration, simpler code are all good things in my book. This is not to say that people from other platforms should not be taken into consideration, but their opinions are relevant only if they are backed up by working code that is clean and integrates well with Gnome.
And if you don´t like where Gnome or KDE are going, you can always fork it. If you have enough developers and users, it should work, but the truth is that those users and developers are almost non-existant outside of Linux
Considering Sun was one of GNOME’s biggest financial supporters, and that GNOME is now the standard UI on just about every commercial version of UNIX (it has pretty much replaced CDE), this would be a real slap in the face to all of the other GNOME players that aren’t Linux. Sun was a huge backer of GNOME. IBM contributed a lot too. HP also contributed to GNOME. I believe HP and IBM still do contribute to GNOME. Not sure if Oracle does or not, but Sun was one of the largest financial backers of GNOME. Dropping support for anything other than Linux would be a bad idea and would alienate a lot of financial contributors to GNOME.
Edited 2011-05-19 21:56 UTC
Was, Sun was. Past tense. Sun is now Oracle and Oracle hates FOSS.
And anyways HPUX, AIX and Solaris are going ship Gnome 2.28 + patches forever (until they die). They will not run Gnome-Shell. Never ever.
Oracle loves FOSS … they call it unbreakable and resell it.
You should remove btrfs from your system and vow never to use it instead of spreading FUD.
But that’s a Red Hat fork, not Solaris
No, Oracle loves FOSS as long as it benefits them.
Re: OpenBSD on the desktop, See: http://undeadly.org/cgi?action=article&sid=20110420080633
Its not the first time he cames in with his ‘briliant’ ideas and fscks up everything, PulseShit anyone?
R.I.P
I say go for it, exploit the potential of Linux to the maximun w/o the need to worry of other unpopular platforms.
This assumes that Linux is also unpopular, which happens to be right.
Oh noes, Linux fanboys and advocates, hitting 1% on the desktop barely makes Linux popular so you don’t need to reply to my post. Just do as usual, vote me down.
Hi,
There’s too many different “Unix-like” OSs and not enough common standards between them; which makes it hard for GUI projects to support all “Unix-like” OSs properly.
There’s too many different GUIs and not enough common standards between them; which makes it hard for “Unix-like” OSs to support all GUI projects properly.
[sarcasm]
Obviously, GUI projects should abandon support for different OSs; and OSs should abandon support for different GUI projects. Any suggestion that common standards could’ve prevented both these issues would just be silly.
[/sarcasm]
– Brendan
Considering how much money Sun contributed to the GNOME project, how how much of a strong backer of GNOME Sun was, this would really be a slap in the face to the Solaris community if GNOME decided not to support Solaris anymore.
Edited 2011-05-19 21:45 UTC
Sun’s dead in case you’ve not noticed, and Oracle has done a fine enough job of spitting into the face of whatever community existed already.
Edited 2011-05-19 22:29 UTC
By community you mean the guys who took Solaris and ran it for free? Like in free beer? They still can get their free beer from Oracle, they just can’t decide anymore what ingredients the free beer contains.
I mean anyone outside Oracle in the Open Source world who might’ve given a damn about making their codebase more friendly to the platform in general.
Edited 2011-05-21 21:19 UTC
what solaris community?
i, as i’m sure many others did as well, left when oracle killed opensolaris.
I read the headline… It gets my juices going…
I glance over the comments… It gets me even hornier…
Then, as I hear the drums in the background, my fingers move over the keyboard…
Perving over the screen, I post my ultimate opinion, thoroughly expressed down to the bit..
I feel unique. I am myself again!
YEAH BABY!
Then I go on and click on the next piece of news…
Oracle is no longer interested in supporting GNOME on Solaris (as Oracle has decided that workstations have no place in its product line and that they don’t need a GUI for server stuff) and nobody uses it on BSD anyways. Therefore it kind of makes sense to reserve support for platforms where it will actually get used (aka Linux)
Where did you hear/read that?
Nonsense. Granted GNOME is working hard to drive away serious users with its 3.x version, but I know at least one BSD guy who uses GNOME every day. There are doubtless many more.
“GNOME Discusses Becoming a Linux-only Project”
Alright, that sounds ridiculous, and pretty bad. But…
“…influential voices within the GNOME project are arguing that GNOME should become a full-fledged Linux-based operating system, …”
Now that… is f***ing retarded. And downright f***ing *bad* to the very lowest level. Complete insanity.
Seriously, it’s one thing (and bad enough) to drop portability and focus on one OS/kernel, but they’re taking it to a whole new level with just the *idea* of GNOME becoming a full-fledged OS. IMO, one of the greatest strengths of the major desktop environments is that each distribution (and OS) can tweak their own versions of the environments and adjust the preferences/change the defaults. If GNOME fully controls not only the source, but the distribution of their desktop by literally making it its own “official” Linux-based OS, goodbye everything (left) that is still good of GNOME.
What’s next? GNOME decides to go proprietary and charge for the privilege of using their desktop? Seriously, what the f***?
While KDE really picked up after the major disappointment that was KDE4 (and 4.1, and 4.2…), GNOME is really pulling an Ubuntu here with their insane ideas. KDE redeemed themselves… GNOME and Ubuntu appear to be digging their own graves. At least attempting.
Can’t you read? This was just a suggestion by one of the developers. “They” haven’t decided on anything or proposed this as an official suggestion or whatever.
Notice I said, even “…just the *idea*…”; even if in the end they don’t do this, this is a disturbing thing to even consider. And even more disturbing that it’s not just one person’s idea. Yes, I know it’s not official, but similar ideas spreading between multiple high-up people is not good.
Oracle will be fine. GNOME should do whatever is best for GNOME.
Whatever happens, it will be pretty inconsequential.
You mean that even with an outrageously different UI that is dumbed down (and f***ed up beyond all recognition) to brain-dead and hideously non-functional levels, there will be absolutely no qualms amongst GNOME users–and even non-GNOME users of traditional desktops–if the GNOME head honchos decide to change their project in such an extreme way?
I find that hard to believe. Even companies based on profit–ie. Microsoft and Apple–would have a hell of a time keeping their current users (as well as gaining new ones) if they changed their interface in such a major way.
Familiarity sells.
It doesn’t matter if GNOME changes completely. Distros will find an alternative. Did I really need to repeat this
Nope. Not really.
When Ubuntu announced unity, a shell competing with gnome-shell, everybody was complaining about more options. Now everyone is complaining about less options. Maybe someone should write an article that gnome is not changing in any way so that everybody could be happy.. I doubt..
They released budget figures yesterday here in New Zealand, and since both left and right-wing parties are unhappy, I figure the government has it about right.
And I think you’ll find the same is true with UI design… the best you can do is balance the number of unhappy people at both ends of the spectrum…
Hello GNOME people, dont overlook anything. Today it may be Linux kernel. I respect it. But tomorrow a new powerful kernel will surely come in force, (may be in a few months, who knows 🙂 ) … You will be left out then !!!
It’s a dangerous proposition.
What if a “Free” and better operating system emerges from the trenches 20 years from now. Something like a Free Plan 9 on steroids. Or, God forbid, something really bad happens and Linux goes proprietary.
GNOME is putting all its eggs in one basket. Free Software was _never_ just about technical merit.
Technically speaking, it is extremely scary, that systemd as an init system isn’t POSIX complaint by default and only Linux enhanced optionally. While systemd is the most impressive init system I’ve encountered, this single fact worries me.
I have long dreamed of a fully integrated GNOME experience from boot up to shutdown. Fact is, the current “way” is a mess. Developers just grab a bunch of projects, make them “almost” work well together, and call it an operating system.
As a FreeBSD enthusiast I am only slightly disappointed, and I honestly believe that FreeBSD would have been the better choice here. However I respect their choice with choosing the Linux kernel, and hopefully instead of making just another distro, the GNOME developers really do something completely different and up to par with the level of quality and integration of Windows or OS X. Even if they have to re-invent the wheel, integration is the key!
I look forward to (one day) installing GNOME OS on my desktop.
Edited 2011-05-20 06:36 UTC
We might have a perfect desktop experience finally when Gnome would be closely married with linux kernel and more importantly the service initialization engine. Also GNOME would dictate the way system configurations are stored and if GNOME would prefer registry-like gconf that is nicely integrated with init engine then I would very much support the effort.
Basically there would be no Linux+GNOME anymore but GNOME would become OS. Just like nobody (majority) would not associate Linux with Android or MeeGo although both have penguin hearts.
I believe that actually the very best way forward is to start making specific Linux distros where only kernel is Linux and everything else is very deeply integrated with user interface. Just like OS X is user interface tightly integrated with BSD kernel. And Android is user interface deeply integrated with Linux kernel and mobile hardware.
Edited 2011-05-20 08:14 UTC
Discussing issues like that is good in my opinion, but certainly not the way this discussion happens.
I think at the beginning the perceived problems should be mentioned. What are the problems?
Well then devs of other OSes like the BSDs etc. could have a change to argue, in contrast everything is very vague now and the benefit not clear.
Another question is if not Gnome should change and use more abstractions. For that there are examples like KDE Solid, which works on a multitude of OSes.
To me it appears recently there happens much important in the Gnome world without careing about anyone else. Be it other desktop environments and OSes recently. This is really bad as it makes all the efforts void that were put into cross dektop, multiple OS. And much of these efforts were also put into by Gnome devs.
Until recent years, Direct Rendering was both BSD and Linux project, but not anymore. Because for developers it was too much effort to support both with same code.
And now we have similar situation with a larger project.
Truth is that BSD or Solaris aren’t “extremely” popular, and live effectively from being dragged by Linux, in a sense it’s Unix and many linux software also works on that OS with minor porting effort (if any). If for Gnome it makes sense in ease of develompent and for better technical decisions, while BSD and Solaris users are already just a tiny minority, they should do that. From the morality side of things, BSD and Solaris people should realize that they already got very much due to popularity of Linux (kernel and all the surrounding sw), so nobody “owes” them anything.
Uh, exactly what are the BSD’s and Solaris owing the Linux kernel in terms of popularity?
And I’d suggest that if you really think that Abstraction layers will solve all the problems of multiple platforms, you are not a developer.
1. Someone needs to devote a lot of time in developing an abstraction layer, when they could have been spending that time developing cool applications and features.
2. Abstraction layers are by their very nature a case of least common denominator. Yes, you can implement optional pluggable features, but that just reinforces point 1.
If GNOME developers want to spend their time creating cool stuff rather than supporting minuscule platforms, that is their choice. The developers on those platforms can port the software if they wish and exist. It is after all, open/free software.
The reality is that GNOME/KDE/E17/XFCE/whatever all have a large group of whiners, but very few actual, productive developers.
So before you whine about this, you really ought to fulfil these two requirements:
1. You actually USE one of these platforms. I suspect a quite a few whiners don’t.
2. You understand the process of software development and the time added to your process by supporting multiple platforms. HINT: it is not free. It isn’t even cheap.
As a GNOME users since the pre-1.0 inception and a Linux user since 1995 I find talk of GNOME OS most interesting.
Google the proverbial 900lb. gorrila made Linux mainstream with Android. Android does not use X and does not use an X-based toolkit or DM/WM. Any talk about GNOME as GNOME OS is targeted at Android as the most visibile direct competitor.
The X devs are abandonning ship in droves headed towards Wayland. Wayland is 100% pure linux and not portable. Once Wayland is viable the GTK Wayland backend will take precedence and the X backend will wither on the vine. The abandonment of X is the end of an era in which a UX system, which was composed of a combination of kernel+userland+X+wm/dm consituted a whole OS-this interchangeability and code sharing made it possible for a near infinite number of possible permutations ranging from kfreeBSD to Gentoo-Hurd to vie a in vibrant ecosystem of alternate UX-like systems. This is comming to and end, not immediately but foreseeably, for better or worse. Legacy concerns about CDE and Motif are now officially history.
GNOME already attempted the totally modular thing(ie. division between core and desktop, portability and compatibility goals across arch’s and systems) and Meego and Nokia were the results and GNOME got burnt really, really bad. Outside of the OLPC and a handful of other smallish project modular GNOME got no traction.
GNOME has gotten big, it’s grown up and despite ever recurrent growin pains has reached maturity. The old corporate environment which kept X going as the centerpiece of compatibility (X11) is no longer. Sun, once a long supporter of GNOME/X got swallowed up by Oracle and Oracle has absolutely 0 interest in the broader free software community. HP, IBM love Linux but have no real interest in the user facing side of software development, kernel sure, but graphics/graphical applications, no interest.
All the major players in X drivers land have opened up their docs and code except for nvidia. At least part of the shift towards Wayland can be read as a big FU to nvidia, who by playing users against free software developers have been instrumental in keeping X from becomming what it should have become.
GNOME has gotten big enough to weather an outright fork/split represented by the Fedora-based Gnome-shell and Ubuntu-based Unity. Many criticize Ubuntu for going in on their own with Unity/Upstart/LightDM etc. But the reality is Fedora/Redhat has sponsored a stream of *kit(consolekit, packagekit, devicekit, etc.) over the years, each charachterized initially by terrible documentation, non existant support for other distros and an attitude that says “our way or the high way”. So in essence Ubuntu and Fedora are each doing the same thing-just with slightly different rhetoric, fedora basically claiming GNOME for itself(which didn’t start with systemd, but with pulseaudio) and Ubuntu claiming the Linux desktop for itself. Hence the talk of GNOME OS. Systemd may actually catch on and be used by many other distros, it certainly has it’s merits and advantages.
But mores users use Ubuntu than all of the other distro’s combined plus all other non-Linux GNOME desktops(*BSD, Solaris, etc.). Core technologies will still be shared across Gnome-Shell and Unity but GNOME itself is in danger of becomming the lowest common denominator between the shell and Unity. Ie. what is a GNOME developer nowadays ? The answer to this is less clear now than it has been for several years. But it is also less important now. It used to be that there were a number of large commercial distributors who had to each give and take when it came to the free software communities. Now it’s Redhat and Ubuntu, Mandriva is out of commission at the moment, and Attachmate is not going to sink the money into SuSE sufficient to really sway things. While I commend the salaries that have been paid by Intel and Nokia to X/GNOME devs over the past few years in hindsight their contribution has been problematic at best-pitting one group against another and playing corporate loyalty games and infusing GNOME/X development with the pervasive propietary influence of secrecy. GNOME has completely gone from causal free software developer greatness to serious corportate gamesmanship, much as the Linux kernel has become IBM and Oracles playground.
Ubuntu has HP on it’s side and Redhat has Intel on their side. No it’s not quite that clear cut-but close. Intel has strong armed the move from X towards Wayland, which was easy to do because a big chunck of the most influential X devs work for Intel. Intel strong armed the move to a clutter based GNOME desktop, ie. the shell. Sad because E17 was ripe for the picking for that task and all of the problems with proper 3d support in the drivers could have been put on ice. Tungsten Graphics with it’s Gallium architecture is still the odd man out-if X has a future beyond Wayland Gallium and the radeon/nouveau dev’s are it’s best hope because the Intel X devs are no longer playing ball. Perhaps the future is not so bright for Wayland, perhaps clutter isn’t the last word on animated scene graphics on the GNOME desktop and perhaps the talk of GNOME OS will vanish in that great big bit bucket in the sky. Who knows. But given the obvious trends it is easy to see why the idea of GNOME OS is becoming palpable and perhaps even necessary, even if it remains unpallatable.
Moreover the talk about a Linux-only GNOME OS reveals a fundamental shift in self-perception. When you are on the outside and want to get inside, you bend over backwards not to alienate anyone, to include everyone and everyone sings kumbayah, holds hands and hugs trees. Once you are on the inside you condescend to those who were once your friends, if you even stoop so low as to acknowledge their existence. I do not like this development, but I can’t really begrudge it either, the GNOME guys/gals have made it into the big league. Let’s just hope that their egos don’t get to big for their britches.
Fine, let ’em do it. KDE can step in and will probably get more help from the BSDs than GNOME ever did. Luckily for Solaris users it still has CDE, which is actually smoking fast compared to GNOME (I can’t believe I’m using ‘CDE’ and ‘fast’ in the same sentence).
-one kernel
-one GUI
-one framework
-one way to manage startup
-one way to manage settings
etc, you see the picture.
All of these highly integrated and written specifically with that OS in mind.
Linux isn’t an OS, is just a collection of software flocked together.
And yes, making Gnome Linux specific will make Linux more of an Os than it is today.
-one GUI
Like Aero and Classic.
-one framework
Like .NET/Winforms, GDI and WPF. Or consistently having a different widget set in your core OS and your flagship Office suite. Or Cocoa and Carbon.
Or one could add equally important things:
* One way to deal with application installation. Like MSI, Installshield and NSIS. Or drag&drop or installers.
I can’t comment on BSD, but as far as I’m concerned.. Solaris is (sadly) dead since Oracle took over so I don’t care about that bit. I/we need to look ahead, and the future is clearly Linux
Tom UK
The Gnome development team has the right to support – or not support – any OS that they choose. As a long time KDE user (and FreeBSD user) I could care less.
I think this gives KDE a wonderful opportunity to expose many more people to it’s merits that might otherwise not have been if Gnome was the default window manager.