Bug#727708: Re: Bug#727708: multiple init systems: We have to see it for what it is: Lennart/Linux OS. Not.
ChaosEsque Team <chaosesqueteam@yahoo.com>
>> We have to see it for what it is. Lennart Pottering and his acolytes
>> who work within other projects are essentially forking Gnu/Linux and
>> are creating LennartOS.
Debian is a multi kernel solution. So solution has to deal with the
fact all OS's are differenet.
>> What should be done is to follow their often spoken refrain: fork
>> every project that relies on systemd, xyzkit, LennartStuff:
>> create X.org-concrete (if they go the Lennart way (there's a patch now!))
>> KDE-concrete, Gnome2 Gnome3-concrete. Security fixes would be
>> applied, as-well as legitimate bug fixes (all these are often
>> one to three liners). Otherwise the software would mostly retain
>> its form. Something familiar, mature, stable.
Fork the kernel??? right we all know how successful this turns out
for those making clones of Solaris. Solaris clones have to go SMC
they don't have a option of using a different init system.
Sysvinit came on Linux by being lazy.
Secure software is a science. I am sick of those who say Software is
more an art. Saying software is a art is a nice universal excuse not
todo quality control.
Really the way systemd is going its going to be come like Solaris and
SMC not able to be split. Launchd on Freebsd is also most likely
going to become not able to be split.
Forking every package that depends on systemd is pointless.
Providing a solution where packages can be less tightly bound to
systemd is the only way forwards to your problem.
Future we need to be able to be able to release 1 package for systemd,
smc and launchd at a min. Sorting this out could also see debian
packages directly installable on OS X.
>From the LCA2014 conference there was a horribly porting deb packages
onto rpm systems because the system could not be stopped to change OS.
The Linux world is horribly fragmented. Art side is part the
problem here.
Yes we are going to see most likely a over reaction going the other way.
Lets go over the key points of software.
User interface design. This is majority a science. You create a
theory you go and attempt to prove it wrong when you cannot this
becomes acceptable wisdom to base interface on.
User interface graphics. This is part standards and part art.
Quality Control. This is again a science. Methods and procedures.
Software is closer to maths and science than art. Yes there is a old
believe that software is art but anyone holding that idea normally
ends up making suspect software.
Science does not give you the right to bully. Lennart is laying down
a challenge who is going to step up. Those attempting to stay in
past on sysvinit are waking up its not possible.
There are still those following advancing Linux without following
Lennart. http://linux.conf.au/schedule/30076/view_talk Project
hammerhead for one.
Something that is required is how to make more packages and
applications not distribution bound or init system bound. Generator
for init files is one of the ways out. Libraries wrapping particular
feature requests is another.
Reply to: