as a developer, I love that decision. (100% compatibility/stability no interefering with new, untested libs coming in via the system's repository)
as a user, I hate that decision on every level. It means that I've to rely on the vendor to deliver security patches, I've the same lib 100times on my system which takes up space (and no, cheap space is no excuse to that. Bad habits stay bad habits). I'm not happy about it.
Similiar to my feelings about the GPL. As a software developer that needs money to pay my rent, pay for my lunch and so on, I hate the GPL. It's retarded in being infectious. Sure, you can sell your software but only one person will purchase it and can then rename it, put it on a website, and sell it for 1/4 of the price.
As a user, I love the GPL because it gives me the ability to do anything with it.
Is it normal to be so split about such things?
Oh noes! It's the rabid wolves out to get Ubuntu again. If the many existing solutions were perfect or just needed a bit of work to be acceptable then why hasn't anyone done it yet? Why fault Ubuntu for trying to solve an obvious problem?
I've learned that Ubuntu is damned if they do and damned if they don't so they might as well ignore you bunch and try to accomplish their goals.
Is that the same 'acceptable performance' as that of the Software Centre?...they already have proof of concept code working with the current system written in Python and acceptable performance...
While i understand canonicals frustration with the current situation (yes its a long standing problem for linux) i think this is a problem that should be discussed with all the major distros. FFS we don't need yet another standard.
There is only one way out of this - to provide well-thought stable interfaces,
and when stuff gets updated to shift them into legacy bindings, till they irreversibly break.
But that would mean that library developers will have new hurd of work to do - for every new version to test how legacy bindings perform with them.
And even this does not guarantee long run, because architectural changes are inenvitable, and those break everything.
So, how about stop being lazy and actually getting responsibility to maintain your software or to opensource it so that others can maintain it.
One law I learned from Linux: Application unmaintained = Application is dead.
You can as well set liberation money on your software, ie "reverse"-kickstarter. Blender was born this way.
If you apply classic proprietary "sell copies" monetarization approach to GPL, don't worry to become a butthurt. Its like trying to catch water with a colander
Last edited by brosis; 05-08-2013 at 04:03 PM.
So if it's all dynamically linked, and they ship their .so files, then it's an OK idea, even if redundant in terms of disk and RAM usage. But by far not if it's statically linked.
More copies of the same lib is not really a problem, you can easilly use a deduplication software in userspace (via symlinks) activated by every installation/disinstallation, or even the emerging kernel ones.
The real issue is that this is all about proprietary software, do you understand this?
In an ideal GPL world recompile a package isn't really a problem, doesn't the distros born because of this?
Have someone who takes care of the dependancies and sort the problems for you?
If even a rolling distro works good, i can't see a problem here, if not to please closed source software.
In my experience, arch works good 99% of the time, the other 1% is because of conflicts with binary blobs.
Selfcontaining packages are just a bunch of data files, executables and libraries with a script that set appropriate LD paths.
EVERY developer can put whatever he wants in that package by reaching all the deps until glibc (game developers usually put just sdl and no much else), so i really CANT see the need of this canonical move; i think they just want to chain developers to their system, instead.
What if in the future the ubuntu base system the devs are linking to will contain a blob? think about that, it is dangerous.
Said that, ubuntu is taking a direction, and i will NOT follow their way of development.
Last edited by kokoko3k; 05-08-2013 at 04:25 PM.