A correspondence between two key people at GNOME has set a flutter, neigh a quake of level 4-5 in the Opensource world. What could termed to be nothing more than routine technical discussions, bordering on being slightly aspiring in nature, is suddenly eating up the blogosphere of Linuxdom.
How the systemd quake happened
The person involved here is Lennart Poettering, who sent the mail suggesting/proposing systemd become a blessed external dependency for GNOME 3.2. and had a spate of follow-up replies from people on the To list- desktop-devel-list gnome org. And that was how the system quake was triggered at exactly 14 hours and ten minutes on the 18, 2011, May.
What he proposed was a line of thought to overcome a technical point that was putting things in development into slow motion. Interface between GNOME and system being minimal why should not become an external dependency believes Len. As the number of interfaces for GNOME and its components is likely to increase, and systemd is Linux-only, then non-Linux platform replacements seem unnecessary. Additionally, if support for the archs can be dropped and distributions keen on offering some of the features of systemd would be better off compiling system, dropping all except for the mechanism daemons.
He explains clearly that not all Linux distributions are presently live with system. Almost all have migrated to newer versions or are in the transitional stage with additional packages for distribution. Apparently, Ubuntu is a lone exception here but expectations on resolving it by the following year is high.
Integration between gnome-session and system is very lose and can easily be “#ifdef’ed out” for distros/OS that are retro.
The final point is since system has close to zero external dependencies, what is truly required is only Linus, udev, D-Bus and some optional deps.
Telling and justified reply from McCann another GNOME giant
McCann, Jon agrees that the proposal would likely suit Debian and definitely not GNOME. Programming GNOME as an OS will not be viable as the ultimate goal and instead proposes an app building bloc or SDK that has higher levels of user-friendly features.
GNOME should remain a child of open source, unlimited and easily portable to any architecture. Confining it to only Linux and not offering portability features would really be regressive to the open source spirit. Though, kernels may not interest all and there would be limitations on the ease of use, straight-jacketing GNOME is not the answer.
But the thought is positively inspiring and probably the next step would be put all the energy behind a Linux OS. GNOME needs to retain more and allow more of LINUX.
However, Linuxdom is in for some hard thinking and Linux-only would be voted out for an exciting Linux OS or platform.