Not being locked into one distro???? You're destroying the dreams of the distro company CEO's. STOP RUINING THEIR BUSINESS PLANS!
Seriously, someone take a hint: Linux is supposed to be free, and no one should be locked into a particular distro. If you release a piece of software, you need to make it EASY to install on ANY distro, which means using a software installation standard. What's that? None exists? Then use Zero Install because that's the closest one I know of since it can run on top of any and all distros.
is to make distros irrelevant by pushing devs to release cross-distro packages and push for cross-distro package formats so distros are only nice bundles to get up and going quickly and nothing else.
Yeah, god forbid a Linux desktop ever do as much as Windows or OS X can. I think KDE has come the closest out of all of them, depending on the distro, to being loaded with as many features, but they all still lack in critical areas. These are areas in which Windows is still, after decades, p0wning Linux.
The biggest problem is standardized Linux packaging. All these distros think it's lovely to keep users in their walled gardens, and it's extremely sad when Linux users don't mind. Sure, it's nice to have a cute little manager connect you to their walled garden, and some gardens contain a fairly big wealth of programs, but they are still gardens with walls. This leads to fragmentation due to the same programs being customized and fucked up in different ways instead of everyone being on the same page by downloading the default program from the developers directly. Great, now I have to look on *Fedora's* forums to figure out this Apache issue, or *Ubuntu's*, or *SUSE's*, etc, instead of being able to learn and rely upon real standards. Most importantly it leads to the inability for users to share programs directly, and instead they have to grovel and rely on their service-based repository while at the same time laughing at cloud OSes and saying they will never become reality. Guess what, Linux has been mostly a cloud OS because of walled garden reliance and a lack of packaging standards for a long time now. You only own your OS once you don't rely on a repository and your software is truly mobile and modular. Sure, you can make a server archive of all the packages, but only for a specific distro and specific version. Woohoo, great flexibility there, not.
Another area is driver management. I don't care of the solution is DKMS, making Linux have actual standardized ABIs for some/most/all driver module interfaces, or what, but the fact that there isn't an existing desktop solution there sucks. It's great when you never have to fuck with things, but that's true of all OSes, until you need to. Then what is the solution? Can users share drivers to get around issues? No, you're fucked unless you're a Linux geek, you're reliant on your distro and bring the problem to them instead of taking it to the actual developers directly.
Instead of helping Linux standards and working on the key problems, Linux users seem complacent to allow distros to put in proprietary solutions. They want to use proprietary leverage to get money instead of being good neighbors in the Linux communities and competing in fair ways while upholding user freedom. Sure, some of these companies have done some good things in different areas, but you can't ignore the bad just because of that.
Until Linux is free and no one has to rely on a single point of failure, a single dictator, a single source for their livelihood, and instead can help the world and Linux community as a whole grow by utilizing and helping with real standards so that software proliferates and helps instead of gets held back and controlled due to a lack of those standards, Linux won't ever be number one on the desktop.
Was to give Microsoft an excuse so they could proclaim that their systems followed standards and were cross-platform. Of course in reality, the standards are always only possible if you're using Windows, otherwise you get only partial functionality, which means it's not really a standard.
A total lack of standards on the laptop building topic aside, this would make all computers bundled with Windows totally useless. Fine! That will push even MORE users over to building their own computers, something which will always be cheaper anyway.
Thanks for that, and this is why pushing for openness to combat all that control and corporate greed is so important. Support all movements for openness!!!
I'd like to start an open car company myself as I'm incredibly sick and tired of overpriced proprietary replacement part costs. Of course, the U.S. government might have to be overthrown first for that to happen due to the depth the existing auto companies have dug themselves into the government.
Yes, it is of course slightly space-inefficient keeping older libraries around if the programs could be compiled with newer ones. If the maintainer of a program stops maintaining it, you might have to keep older libraries around if newer ones have broken their ABI/API. If the choice is between a functioning program or a broken one, I'll take the functional one and so would anyone unless you can find a replacement, and those can't always exist, especially for games for example.
I wish everything used standardized dynamic paths, like you could have them be an environmental variable, where programs would just query $LIB64 or $BIN64 etc in order to communicate effectively, allowing a system to store files anywhere it wants. Hell, you could make a structure like Windows did if you wanted to, and put all the shared libraries in \Linux, and all the bins and other stuff in \Program Files, hehe. ^^
Ah yes, one of the last major problems Linux still really needs to solve: binary portability. They should provide 7.11 available for download in a universal installation package which contains everything needed, and then at the end of the installation you select which Xorg version you want to boot into by default if you already have a version installed.
When are Linux users and devs going to hunger enough for this kind of freedom so that they all switch to truly cross-distro installation packaging systems like Zero Install?
Windows has installation CDs that are slimmed down as well as ones which contain more software. The solution isn't to have all distros include the same software. You said what the solution is: standards. All you need are standardized ways for the same type of thing to be done across any distro, like program installation standards. The system needs to be able to recognize all dependencies and to easily obtain anything which is missing. Then, who cares if libraryXYZ is missing? Your package manager will get it for you.
The stupid create-your-own-software-universe model needs to die. Programs need to be cross-distro at the very least, and cross-platform preferably.
Everyone: please just say no to any systems which attempt to lock you into a single vendor source for your software when alternatives exist that give you much more freedom. Using Android as an example, it takes away your freedom by locking you into Android-only apps. Why would I choose that over a distro which allows me to run any and all Linux apps? Want to buy something from the Ubuntu Software Center? Hell no, you shouldn't be ball-and-chained to a specific distro, and even if you did find the DEB file you would be locked out of RPM-based distros unless you somehow converted the DEB to RPM, but why should you have to?
The point is that even open source software can make you a slave if it doesn't offer standards, because that is where real freedom comes from. Development time has a cost, so spend your time and money helping out projects which seek to give true freedom to all computer users world-wide. Programs like Zero Install perhaps? Standards groups like freedesktop.org?
Of course driving a semi is safer than riding a bicycle when it comes to having an accident with the typical vehicle on the road, but the more who drive smaller vehicles the more the typical vehicle decreases in size which makes the road safer for everyone. Constantly pushing tons of metal and plastic around on the road wastes insane amounts of energy. There are lots of other ways to get from A to B without having to do that.
It's called not having to do anything at all, if the Linux kernel had an intelligent standardized driver interface. scp driver.ko stupid@box:., ssh stupid@box modprobe driver.ko, hard I know.
Having drivers come with the kernel so that there is more "plug-n-play" out there is a wonderful feature, but no, these are problems that do affect everyone. There are lots of scenarios I can come up with where this feature would be great to have. One would be being able to use new hardware with an old stable kernel easily. Another would be for users to be able to share drivers easily with each other, instead of having to give noobs instructions on how to compile something. Yet another would be so that anyone could package a driver that works with a piece of hardware that works. Vendors would be able to do this for instance. Vendors could also give Linux support much more easily without having to go through an annoying compilation step.
No matter how you look at it, that *feature* in Linux would be exactly that, it would give you more flexibility, require less upkeep, and make support much easier. Oh, that driver that came in that older kernel is crap? Here's this newer one that works, Grandma, just click on it to install. *That* is a feature, and there's no god damn technical reason why a standardized interface allowing for a more modularized kernel like that cannot be implemented. I'm all for open source drivers, but this isn't an open vs. closed argument, having this feature would help *everyone*, regardless of the license of the driver. Just saving the work of having to recompile all the drivers every time there is a kernel revision would be a nice feature. Save some electricity. Geezus.
A driver should never need to be recompiled by anyone but someone who needs to see the code for some reason. There is no reason you can't have a common language between the drivers and the rest of the kernel. Finding a security vulnerability in the *common language* (ABI) itself should be extremely rare, and basically unheard of if the language was designed properly (that's the point of a common language, to not change), but even if you did it'd simply be a matter of depreciating or removing the old language in the new kernel with the replacement language, and requiring modified drivers for those that are affected.
If someone says that driver ABI is too low-level to create a standardized interface for it, I call BS. There's no reason one couldn't be created. Programs use standards and get along with other programs all the time, and there's no reason drivers have to be any different. I believe there isn't a bigger push for this because distro companies want programs to be locked into their distos and distro versions to encourage reliance on them. Sure, it is possible to compile, but I believe they enjoy having that artificial barrier there because it makes it MORE difficult for end-users to rely on anyone else but them.
For a community that strongly cares about standards and interoperability, this area, and the area of making cross-distro packaging solutions in general, are totally contrary to those beliefs.
Well I can understand keeping power settings all in one place. Users define their power settings inside of Gnome, so unless this configuration is placed in a standardised format somewhere so that all other programs will always know how to access it, the login manager won't have any knowledge of the user's settings which would be bad. Of course, there should be standardised ways to access ALL information and configurations on a Linux system, so I'd love for that to happen regardless.
I'd consider any Linux distro that had picked up adoption of universal packaging standards. Maybe that will happen soon though? Oh please oh please oh please...
Interesting, I hadn't heard about that. I wonder if it will be fixed in that main kernel version or if it's too big of a change and will have to be fixed in.39 or something.
To summarise, their argument is that LightDM is light on code because it can't do as much as GDM and the others, and if you removed those features from the others they would be light as well.
If that's true and that is the main difference, maybe it'd be easier to strip out, or turn off, parts of GDM if Canonical wants to dispose of certain features to achieve a faster boot time.
I totally agree with most or all of your points, and maybe some of those things will be addressed but I think some might be too contrary to the paradigm. I too was annoyed at the amount of work it took to open up another terminal, but for that particular problem it will or could easily be addressed as it has been addressed for Firefox. For Firefox, if you right click on the icon, you can tell it to open a new window. For Gnome Terminal, Nautilus, and other apps though there is no option in the right click menu, and instead you have to go up to their menu and tell the program to open another window which is much slower. So, I hope they make all the programs capable of having multiple windows have that right click option.
On a completely unrelated note, I think for the "system tray" that they wanted to get rid of which is normally used for apps you want to know are running, but don't want them taking up a lot of room in the window switcher, that they should do what Chrome/Firefox do for "browser apps", and give you the option to condense those long-term apps down into a single icon in the window switcher.
The panel is the launcher, and it's scrollable when it gets full so that you don't need (but still might like) another one. It's a "dock". A dock is a window switcher + app launcher rolled into one, which is what both Microsoft and Apple made default in their OSes, and now Gnome and Ubuntu. Of course, Linux has had docks long before anyone else.;)
The elderly people I've showed Gnome to had no problem navigating around in it either, it came pretty naturally and simply due to it being in general an easy GUI to work with, despite a few rough edges here and there.
It would be nice if applications could use a single API for Linux in general, and the program would be rendered appropriately for the DE the user happened to be in. This would destroy the whole "this app is KDE, this app is Gnome" thing. If you could just standardize the API for every service an application needed, whether it be a clip board, or a key store, or a network service, having some good standards/APIs so that apps could be shared between both DE's more would certainly help the Linux ecosystem.
I am glad that there is a lot of recycling due to apps using the same libraries and back ends though, I just wish the front ends could be recycled too.
As long as the bling doesn't steal tons of their computer's power, most users would prefer to have both bling and productivity. Something that is nice to look at and to use to get their work done at the same time. Crazy concept, I know.
Most user complaints stem from people who used a development release (4.0, 4.1, 4.2, 4.3) of KDE 4 and thought it would measure up to a stable release (3.5).
Maybe they should consider using appropriate labels then for those "development releases". Maybe stick an Alpha there, a Beta here, you know, something helpful.
Regardless, I can't stand KDE4. As mentioned all over, the interface is incredibly cluttered. While I don't like Gnome for not including more easily accessible advanced options which could be simply hidden/buried one level down, until the KDE developers learn to keep things simple and bury their options hardly anyone uses, and basically actually start heeding user interface design and workflow, Gnome will have to continue to be my DE of choice.
Slashdot Mirror
Not being locked into one distro???? You're destroying the dreams of the distro company CEO's. STOP RUINING THEIR BUSINESS PLANS!
Seriously, someone take a hint: Linux is supposed to be free, and no one should be locked into a particular distro. If you release a piece of software, you need to make it EASY to install on ANY distro, which means using a software installation standard. What's that? None exists? Then use Zero Install because that's the closest one I know of since it can run on top of any and all distros.
is to make distros irrelevant by pushing devs to release cross-distro packages and push for cross-distro package formats so distros are only nice bundles to get up and going quickly and nothing else.
Yeah, god forbid a Linux desktop ever do as much as Windows or OS X can. I think KDE has come the closest out of all of them, depending on the distro, to being loaded with as many features, but they all still lack in critical areas. These are areas in which Windows is still, after decades, p0wning Linux.
The biggest problem is standardized Linux packaging. All these distros think it's lovely to keep users in their walled gardens, and it's extremely sad when Linux users don't mind. Sure, it's nice to have a cute little manager connect you to their walled garden, and some gardens contain a fairly big wealth of programs, but they are still gardens with walls. This leads to fragmentation due to the same programs being customized and fucked up in different ways instead of everyone being on the same page by downloading the default program from the developers directly. Great, now I have to look on *Fedora's* forums to figure out this Apache issue, or *Ubuntu's*, or *SUSE's*, etc, instead of being able to learn and rely upon real standards. Most importantly it leads to the inability for users to share programs directly, and instead they have to grovel and rely on their service-based repository while at the same time laughing at cloud OSes and saying they will never become reality. Guess what, Linux has been mostly a cloud OS because of walled garden reliance and a lack of packaging standards for a long time now. You only own your OS once you don't rely on a repository and your software is truly mobile and modular. Sure, you can make a server archive of all the packages, but only for a specific distro and specific version. Woohoo, great flexibility there, not.
Another area is driver management. I don't care of the solution is DKMS, making Linux have actual standardized ABIs for some/most/all driver module interfaces, or what, but the fact that there isn't an existing desktop solution there sucks. It's great when you never have to fuck with things, but that's true of all OSes, until you need to. Then what is the solution? Can users share drivers to get around issues? No, you're fucked unless you're a Linux geek, you're reliant on your distro and bring the problem to them instead of taking it to the actual developers directly.
Instead of helping Linux standards and working on the key problems, Linux users seem complacent to allow distros to put in proprietary solutions. They want to use proprietary leverage to get money instead of being good neighbors in the Linux communities and competing in fair ways while upholding user freedom. Sure, some of these companies have done some good things in different areas, but you can't ignore the bad just because of that.
Until Linux is free and no one has to rely on a single point of failure, a single dictator, a single source for their livelihood, and instead can help the world and Linux community as a whole grow by utilizing and helping with real standards so that software proliferates and helps instead of gets held back and controlled due to a lack of those standards, Linux won't ever be number one on the desktop.
Was to give Microsoft an excuse so they could proclaim that their systems followed standards and were cross-platform. Of course in reality, the standards are always only possible if you're using Windows, otherwise you get only partial functionality, which means it's not really a standard.
A total lack of standards on the laptop building topic aside, this would make all computers bundled with Windows totally useless. Fine! That will push even MORE users over to building their own computers, something which will always be cheaper anyway.
Thanks for that, and this is why pushing for openness to combat all that control and corporate greed is so important. Support all movements for openness!!!
I'd like to start an open car company myself as I'm incredibly sick and tired of overpriced proprietary replacement part costs. Of course, the U.S. government might have to be overthrown first for that to happen due to the depth the existing auto companies have dug themselves into the government.
Interesting, never heard of that one, thanks!
Yes, it is of course slightly space-inefficient keeping older libraries around if the programs could be compiled with newer ones. If the maintainer of a program stops maintaining it, you might have to keep older libraries around if newer ones have broken their ABI/API. If the choice is between a functioning program or a broken one, I'll take the functional one and so would anyone unless you can find a replacement, and those can't always exist, especially for games for example.
I wish everything used standardized dynamic paths, like you could have them be an environmental variable, where programs would just query $LIB64 or $BIN64 etc in order to communicate effectively, allowing a system to store files anywhere it wants. Hell, you could make a structure like Windows did if you wanted to, and put all the shared libraries in \Linux, and all the bins and other stuff in \Program Files, hehe. ^^
Ah yes, one of the last major problems Linux still really needs to solve: binary portability. They should provide 7.11 available for download in a universal installation package which contains everything needed, and then at the end of the installation you select which Xorg version you want to boot into by default if you already have a version installed.
When are Linux users and devs going to hunger enough for this kind of freedom so that they all switch to truly cross-distro installation packaging systems like Zero Install?
Windows has installation CDs that are slimmed down as well as ones which contain more software. The solution isn't to have all distros include the same software. You said what the solution is: standards. All you need are standardized ways for the same type of thing to be done across any distro, like program installation standards. The system needs to be able to recognize all dependencies and to easily obtain anything which is missing. Then, who cares if libraryXYZ is missing? Your package manager will get it for you.
The stupid create-your-own-software-universe model needs to die. Programs need to be cross-distro at the very least, and cross-platform preferably.
Everyone: please just say no to any systems which attempt to lock you into a single vendor source for your software when alternatives exist that give you much more freedom. Using Android as an example, it takes away your freedom by locking you into Android-only apps. Why would I choose that over a distro which allows me to run any and all Linux apps? Want to buy something from the Ubuntu Software Center? Hell no, you shouldn't be ball-and-chained to a specific distro, and even if you did find the DEB file you would be locked out of RPM-based distros unless you somehow converted the DEB to RPM, but why should you have to?
The point is that even open source software can make you a slave if it doesn't offer standards, because that is where real freedom comes from. Development time has a cost, so spend your time and money helping out projects which seek to give true freedom to all computer users world-wide. Programs like Zero Install perhaps? Standards groups like freedesktop.org?
Of course driving a semi is safer than riding a bicycle when it comes to having an accident with the typical vehicle on the road, but the more who drive smaller vehicles the more the typical vehicle decreases in size which makes the road safer for everyone. Constantly pushing tons of metal and plastic around on the road wastes insane amounts of energy. There are lots of other ways to get from A to B without having to do that.
It's called not having to do anything at all, if the Linux kernel had an intelligent standardized driver interface. scp driver.ko stupid@box:., ssh stupid@box modprobe driver.ko, hard I know.
Having drivers come with the kernel so that there is more "plug-n-play" out there is a wonderful feature, but no, these are problems that do affect everyone. There are lots of scenarios I can come up with where this feature would be great to have. One would be being able to use new hardware with an old stable kernel easily. Another would be for users to be able to share drivers easily with each other, instead of having to give noobs instructions on how to compile something. Yet another would be so that anyone could package a driver that works with a piece of hardware that works. Vendors would be able to do this for instance. Vendors could also give Linux support much more easily without having to go through an annoying compilation step.
No matter how you look at it, that *feature* in Linux would be exactly that, it would give you more flexibility, require less upkeep, and make support much easier. Oh, that driver that came in that older kernel is crap? Here's this newer one that works, Grandma, just click on it to install. *That* is a feature, and there's no god damn technical reason why a standardized interface allowing for a more modularized kernel like that cannot be implemented. I'm all for open source drivers, but this isn't an open vs. closed argument, having this feature would help *everyone*, regardless of the license of the driver. Just saving the work of having to recompile all the drivers every time there is a kernel revision would be a nice feature. Save some electricity. Geezus.
A driver should never need to be recompiled by anyone but someone who needs to see the code for some reason. There is no reason you can't have a common language between the drivers and the rest of the kernel. Finding a security vulnerability in the *common language* (ABI) itself should be extremely rare, and basically unheard of if the language was designed properly (that's the point of a common language, to not change), but even if you did it'd simply be a matter of depreciating or removing the old language in the new kernel with the replacement language, and requiring modified drivers for those that are affected.
If someone says that driver ABI is too low-level to create a standardized interface for it, I call BS. There's no reason one couldn't be created. Programs use standards and get along with other programs all the time, and there's no reason drivers have to be any different. I believe there isn't a bigger push for this because distro companies want programs to be locked into their distos and distro versions to encourage reliance on them. Sure, it is possible to compile, but I believe they enjoy having that artificial barrier there because it makes it MORE difficult for end-users to rely on anyone else but them.
For a community that strongly cares about standards and interoperability, this area, and the area of making cross-distro packaging solutions in general, are totally contrary to those beliefs.
Well I can understand keeping power settings all in one place. Users define their power settings inside of Gnome, so unless this configuration is placed in a standardised format somewhere so that all other programs will always know how to access it, the login manager won't have any knowledge of the user's settings which would be bad. Of course, there should be standardised ways to access ALL information and configurations on a Linux system, so I'd love for that to happen regardless.
Come back and let us know what you choose. :D
I'd consider any Linux distro that had picked up adoption of universal packaging standards. Maybe that will happen soon though? Oh please oh please oh please...
Interesting, I hadn't heard about that. I wonder if it will be fixed in that main kernel version or if it's too big of a change and will have to be fixed in .39 or something.
http://www.advogato.org/person/mjg59/diary.html?start=296
To summarise, their argument is that LightDM is light on code because it can't do as much as GDM and the others, and if you removed those features from the others they would be light as well.
If that's true and that is the main difference, maybe it'd be easier to strip out, or turn off, parts of GDM if Canonical wants to dispose of certain features to achieve a faster boot time.
11.04 is SO SLOW to boot in comparison to 10.10.
I totally agree with most or all of your points, and maybe some of those things will be addressed but I think some might be too contrary to the paradigm. I too was annoyed at the amount of work it took to open up another terminal, but for that particular problem it will or could easily be addressed as it has been addressed for Firefox. For Firefox, if you right click on the icon, you can tell it to open a new window. For Gnome Terminal, Nautilus, and other apps though there is no option in the right click menu, and instead you have to go up to their menu and tell the program to open another window which is much slower. So, I hope they make all the programs capable of having multiple windows have that right click option.
On a completely unrelated note, I think for the "system tray" that they wanted to get rid of which is normally used for apps you want to know are running, but don't want them taking up a lot of room in the window switcher, that they should do what Chrome/Firefox do for "browser apps", and give you the option to condense those long-term apps down into a single icon in the window switcher.
no launchers in your panel, no additional panels
The panel is the launcher, and it's scrollable when it gets full so that you don't need (but still might like) another one. It's a "dock". A dock is a window switcher + app launcher rolled into one, which is what both Microsoft and Apple made default in their OSes, and now Gnome and Ubuntu. Of course, Linux has had docks long before anyone else. ;)
They exist, they just don't make chips yet, but that is being worked on. http://www.fabathome.org/ http://www.makerbot.com/ http://www.thingiverse.com/
The elderly people I've showed Gnome to had no problem navigating around in it either, it came pretty naturally and simply due to it being in general an easy GUI to work with, despite a few rough edges here and there.
Valid point, but one I'm hoping may be ignored if much of the public ends up transitioning to WebM instead.
It would be nice if applications could use a single API for Linux in general, and the program would be rendered appropriately for the DE the user happened to be in. This would destroy the whole "this app is KDE, this app is Gnome" thing. If you could just standardize the API for every service an application needed, whether it be a clip board, or a key store, or a network service, having some good standards/APIs so that apps could be shared between both DE's more would certainly help the Linux ecosystem.
I am glad that there is a lot of recycling due to apps using the same libraries and back ends though, I just wish the front ends could be recycled too.
As long as the bling doesn't steal tons of their computer's power, most users would prefer to have both bling and productivity. Something that is nice to look at and to use to get their work done at the same time. Crazy concept, I know.
Most user complaints stem from people who used a development release (4.0, 4.1, 4.2, 4.3) of KDE 4 and thought it would measure up to a stable release (3.5).
Maybe they should consider using appropriate labels then for those "development releases". Maybe stick an Alpha there, a Beta here, you know, something helpful.
Regardless, I can't stand KDE4. As mentioned all over, the interface is incredibly cluttered. While I don't like Gnome for not including more easily accessible advanced options which could be simply hidden/buried one level down, until the KDE developers learn to keep things simple and bury their options hardly anyone uses, and basically actually start heeding user interface design and workflow, Gnome will have to continue to be my DE of choice.