It's Back To Voting For The Debian Init System

[interested]

To be quite honest I am not entirely sure what Ian is trying to get out of coupling the two issues and pushing aggressively for the current L option. The current L seem detrimental even to Upstart that is his preferred init system. For instance, it seems like an Upstart alternative to logind would have to support all init systems (including sysVinit) to be included in the Debian repositories. This makes such an effort less likely and it is hard to see how you can force any volunteer to do this work.

Edit: changed text for clarity.

The L option makes a lot of sense if you accept the premise that the current and former Canonical employees on the CTTE are actually fighting on behalf of Canonical.

This is how I see it; systemd as init system on Debian Linux are bound to win because of the Chairmans casting vote. Nobody really wants Upstart; according to Debian popcon, there are at least 100x more systemd users on Debian right now than Upstart users. The kFreebsd and Hurd porters doesn't wan't Upstart either (though a statement about this is still a draft, but of course they want OpenRC).

But if systemd is choosen together with the L option, it will actually be banned for programs like Gnome and KDE to actually use any features/API's from systemd as PID 1. All Debian programs must be compatible with sysvinit and Upstart, and that must be extremely important for Canonical, who otherwise would have to maintain and debug and package those programs.

So in a perverse way, a systemd victory with the L coupling is almost as good a victory for Canonical as a Upstart victory in making all the Debian developers work for Canonical's cause.

[asabjorn]

So now we have a few initial votes:

Bdale: D U O V F
Don: D U O V F
Keith: D U O V F
Russ: D U O V F
Steve: F U D O V

Does this mean that the D option (SystemD) has already beaten F (Further Discussion) when we take Bdales double vote into account? Even if all the members that has not submitted a vote yet vote F?

[asabjorn]

The L option makes a lot of sense if you accept the premise that the current and former Canonical employees on the CTTE are actually fighting on behalf of Canonical.

This is how I see it; systemd as init system on Debian Linux are bound to win because of the Chairmans casting vote. Nobody really wants Upstart; according to Debian popcon, there are at least 100x more systemd users on Debian right now than Upstart users. The kFreebsd and Hurd porters doesn't wan't Upstart either (though a statement about this is still a draft, but of course they want OpenRC).

But if systemd is choosen together with the L option, it will actually be banned for programs like Gnome and KDE to actually use any features/API's from systemd as PID 1. All Debian programs must be compatible with sysvinit and Upstart, and that must be extremely important for Canonical, who otherwise would have to maintain and debug and package those programs.

So in a perverse way, a systemd victory with the L coupling is almost as good a victory for Canonical as a Upstart victory in making all the Debian developers work for Canonical's cause.

I agree that this is also the way it has looked like to me. However, that is incredibly hard to prove and assumes bad faith.

If it happens to be true that the Canonical members of the CTTE are not acting in the best interest of Debian, and are letting 3rd parties influence key project decisions then that is a serious issue. Debian maintainers are not Ubuntu volunteers and they should not be have to do free labor for a commercial company unless they desire to do so.

[blackout23]

So now we have a few initial votes:

Bdale: D U O V F
Don: D U O V F
Keith: D U O V F
Russ: D U O V F
Steve: F U D O V

Does this mean that the D option (SystemD) has already beaten F (Further Discussion) when we take Bdales double vote into account? Even if all the members that has not submitted a vote yet vote F?

Basically yes even if the other 3 vote F first and D last.

http://www.cs.wustl.edu/~legrand/rbvote/calc.html

[Marc Driftmeyer]

So now we have a few initial votes:

Bdale: D U O V F
Don: D U O V F
Keith: D U O V F
Russ: D U O V F
Steve: F U D O V

Does this mean that the D option (SystemD) has already beaten F (Further Discussion) when we take Bdales double vote into account? Even if all the members that has not submitted a vote yet vote F?

4-1 in favor of Systemd as the default, irregardless of a double vote would give it the clean majority.

[Fenrin]

This is in my opinion great leadership from Bdale. [...]

not sure if it is great leadership, considering that it was known that Steve Langasek worked on a different version of the ballot, and his opinion was not heard on the new one.

[Skrapion]

Basically yes even if the other 3 vote F first and D last.

http://www.cs.wustl.edu/~legrand/rbvote/calc.html

Not quite. Under Debian's rules, you can't tie-break the F option. At least one more person has to vote something (anything) above F before it's over.

[KuroNeko]

Well, look's like systemd will win.

Objectively it really has the most features and is actively developed.
Still... In my opinion it's a bad match for Debian. Systemd is actively developed and I never hear of them doing maintenance releases. Usually you would just switch to the next version, if you experienced bugs with it. Also often enough things change. They promise a stable api, but they will integrate kdbus and change things with cgroups and that will probably have consequences, so Debian 9.0 will certainly be different. Not that one can trust, they won't change init systems till then...

I don't use Debian, so my opinion doesn't count. But as I use Gentoo (for fun), I wouldn't mind them choosing OpenRC. It does, what you need your init system to do, it's really simple to use (even I could!) and is not to big of a change. It doesn't matter that much to upgrade the init system to the developer, so it probably also doesn't matter to the TC, but Upstart promised to implement socket based activation, systemd will integrate more features and both there code bases are much bigger, so backporting bugfixes won't be that easy, OpenRC is small and doesn't have much features, and even if it will integrate some, it wouldn't take much effort to maintain a stable fork for Jessies lifetime. Who does need most of the features that systemd has? OpenRC is reasonably declarative, fast and easy to use. (Probably you could even list every option, and how it could break your system on one page!) But that would mean, Debian would have to patch software, that depends on systemd, not good...

So, why would they choose Upstart? That looks like the worst of both worlds! Is it Lennard, friendly upstream, or did I miss something? Events don't sound safe nor simple to me, but I used Ubuntu only in the beginning, and switched to OpenSUSE and Gentoo to learn about my system, so no user experience here.

I'm certainly biased towards OpenRC, but I think, Phoronix Forums aren't a place, where something like that would never happen...

Either way, shouldn't maintenance costs, ease of use and if the init system brings up the PC be the focus, not the number of features?
That would let them choose OpenRC, jump on!

Still just a random rant,
KuroNeko

[justmy2cents]

This is in my opinion great leadership from Bdale. Despite the fact that 6 out of 8 committee members (Bdale, Colin, Don, Keith, Russ, Steve) have expressed that they are uncomfortable with the current T and L, Ian has so far been successful in his crusade to couple the init system decision and these particular options on the policy issue in the collaborative ballot. I therefore agree with Bdale that letting the vote go ahead on the collaborate proposal seem undesirable and unwise and that the default init question should be decided separately.

To be quite honest I am not entirely sure what Ian is trying to get out of coupling the two issues and pushing aggressively for the current L option. The current L seem detrimental even to Upstart that is his preferred init system. For instance, it seems like an Upstart alternative to logind would have to support all init systems (including sysVinit) to be included in the Debian repositories. This makes such an effort less likely and it is hard to see how you can force any volunteer to do this work.

Edit: changed text for clarity.

why L is really simple, but you need to understand that question order is required here (you can probably see that Langasek is whining about that)
if question T/L is first and L wins (upstart ppl were promoting that)... systemd loses everything it holds advantage in, upstart just some random things as it puts them on feature equal standing. then upstart proponents suddenly gains "PORTABLE" as highly valid reasoning on why upstart is best for second question. but, when question "Which" comes before "How", it probably changes both answers

but, one has to understand why upstart guys are so frantic. if debian chooses systemd, it basically spells death for upstart which probably goes bzr way as abandoned project

[Kostas]

So now we have a few initial votes:

Bdale: D U O V F
Don: D U O V F
Keith: D U O V F
Russ: D U O V F

*claps*
Props for being the only one so far that bothered to actually present the current results in a proper fashion.