Linked by Thom Holwerda on Mon 29th Jun 2009 08:51 UTC
Linux "Intel has created a new network management and configuration system for Linux called ConnMan - but not everyone is pleased to see it challenge NetworkManager. Ars looks at the pros and "conns" of the decision to create the new software."
Thread beginning with comment 370819
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[8]: Here we go again...
by FooBarWidget on Tue 30th Jun 2009 08:34 UTC in reply to "RE[7]: Here we go again..."
FooBarWidget
Member since:
2005-11-11

But you'll have to create the rpm anyway.


You keep talking about how this "needs to happen anyway" as if it's a fact that can never be changed. The point is that it shouldn't have to be like this, that the community should strive to unify packaging efforts in order to reduce wasted efforts. Why shouldn't the community do that? Why should this be the mentality instead?

Reply Parent Bookmark Score: 2

RE[9]: Here we go again...
by vivainio on Tue 30th Jun 2009 10:50 in reply to "RE[8]: Here we go again..."
vivainio Member since:
2008-12-26


You keep talking about how this "needs to happen anyway" as if it's a fact that can never be changed. The point is that it shouldn't have to be like this, that the community should strive to unify packaging efforts in order to reduce wasted efforts. Why shouldn't the community do that? Why should this be the mentality instead?


Some thoughts:

- The .deb packaging format isn't going anywhere (way too much infrastrucure exists for that to happen)

- Debian packaging is hard to learn. It could be made simpler by improving the way packages are made, without need to make changes to the package structure as such (e.g. by providing better "debhelper" scripts, GUI wizards etc). I think Ubuntu is already exploring something like this.

- C/C++ build toolchains suck (automake, I'm looking at you). Improving that story seems more urgent to me than fixing the packaging. qmake, scons are better, possibly CMake too (never tried it).

I certainly want to avoid looking like "stop energy" in this matter, though. There are many fruitful avenues for improvement in this area, and many incremental improvements are possible. But trying to make large communities (debian/ubuntu) or communities with significant financial interests (fedora/redhat) to adopt "revolutionary" changes seems very hard.

Reply Parent Bookmark Score: 2