Debian Is Voting Again On Systemd vs. Upstart Debate

[asabjorn]

3 of the 4 Upstart votes came from former and current Canonical employees, the process is rigged folks.

It is quite sad to see how Canonical has fallen from being a good contributor to the Linux and Debian ecosystem to a proponent of divisive in-house solutions (e.g Mir, Upstart).

I do not believe Debian would even consider projects under the Contributor License Agreement (CLA) if it was not advocated by current and ex Canonical employees in powerful Debian roles. I believe this to be a sad display of a conflict of interest that is not properly dealt with.

[GreatEmerald]

For this occasion, and since I read much of the Debian mailing list during the discussions, I decided to remake them in the format of a (non-interactive) Phoenix Wright: Ace Attorney case format:

http://aceattorney.sparklin.org/jeu.php?id_proces=57684

I'm not quite finished, though, this is just the first quarter of the whole discussion, going only to December. There have been a whole lot of posts... I also tried to keep what characters say as close to what people actually said in their messages, just shortened for brevity. In the end I quite like the result, as it's a quick but fairly accurate rundown of the discussions and it makes it easier to get the general mood for the discussions when the people talking are represented. It's also easier to get a feel of the person's personality.

I also learned some things when researching the people in order to write their profile messages. I didn't know Steve Langasek was actually the maintainer of Upstart... That means that there really were no questions about how he would vote from the get-go. Though I find his explanation for his second choice quite interesting. I'm also quite surprised by how well Ian Jackson's personality matches that of early Edgeworth, even down to the way he speaks. I actually made Edgeworth represent him due to having more animations and Jackson being an active speaker throughout the discussions, so the fact that it fits so well was surprising. The choices for the representation of others also generally came down to the availability of animations for them and then to their appearance. That's also why I had to make the Peanut gallery, as there are just not enough people to represent everyone participating in the discussion.

I'll keep updating the case in the future (and probably make a separate case for their opinion statements later on, as the current case is already extremely big). The link updates automatically once I change something.

As for the recent news, it's interesting to see that lately Steve Langasek seems to be losing it a bit. This message comes off as "Did I say GNOME was garbage? Actually, what I meant is that Josselin is!", which is arguably worse. I consider that an ad-hominem attack and a rather grave accusation, and coming from a TC member it's rather disturbing. And then later on we have this, which is also Langasek assuming and accusing others of doing things they did not actually do. In both cases, understandably, objections ensued.

It's good to see that the voting is now starting for real, but even now Jackson's vote is fishy (as noted by Don Armstrong). I expect to see even more objections until this is over.

[Scimmia]

Yep, the reason to put "Tight coupling" under FD is political. They see their init system is going to lose and this is a way to force package maintainers to not leave them behind. If Upstart was winning, you can bet those same people wouldn't see tight coupling as the end of the world anymore.

[GreatEmerald]

Uh-uh, looks like it might be back to "Further Discussion".

I'm somewhat surprised that Steve Langasek would object to starting the vote, given his previous statements in the discussion (that they need a decision and they need it fast). But I guess he doesn't think a few more days would change much. That said, he's not voting FD himself, so it seems that he's just bitter about it, but doesn't actually want to drag this out further.

[Spittie]

Software outside of an init system's implementation may not require a specific init system to be pid 1, although degraded operation is tolerable.

So, if DL win, Debian will patch Gnome to remove the systemd dependency?

For this occasion, and since I read much of the Debian mailing list during the discussions, I decided to remake them in the format of a (non-interactive) Phoenix Wright: Ace Attorney case format:

http://aceattorney.sparklin.org/jeu.php?id_proces=57684

That's just awesome, good job!

[shmerl]

You can see votes here: http://bugs.debian.org/cgi-bin/bugre...?bug=%23727708

[GreatEmerald]

So, if DL win, Debian will patch Gnome to remove the systemd dependency?

I guess that's what Upstart proponents are pushing for. However, Josselin Mouette noted that it would practically be unenforcible. Someone would have to write the patch, and nobody can force the GNOME maintainers to write such a patch. Though if the Upstart maintainers do that, then the GNOME maintainers will probably be pressured into accepting that (yet I still doubt they could be legally forced to accept it). But in that case I'd expect a GR to happen (and it will probably happen regardless, due to how polarised the discussions are).

That's just awesome, good job!

Thanks. I actually learned quite a bit while making this as I reread the discussions more deeply. They have some interesting information (what problems people have with current cgroups, how readiness protocols work etc.), so I made sure to include that into the case as well.

[shmerl]

For this occasion, and since I read much of the Debian mailing list during the discussions, I decided to remake them in the format of a (non-interactive) Phoenix Wright: Ace Attorney case format:

http://aceattorney.sparklin.org/jeu.php?id_proces=57684

Fun, neat idea, but since I tried to follow the discussion, it's somewhat repetitive for me.

[Serge]

I guess that's what Upstart proponents are pushing for. However, Josselin Mouette noted that it would practically be unenforcible. Someone would have to write the patch, and nobody can force the GNOME maintainers to write such a patch. Though if the Upstart maintainers do that, then the GNOME maintainers will probably be pressured into accepting that (yet I still doubt they could be legally forced to accept it). But in that case I'd expect a GR to happen (and it will probably happen regardless, due to how polarised the discussions are).

My understanding is that if "loose coupling" wins (so any of DL, UL, OL or VL), then if GNOME at some point in the future gains an explicit dependency on systemd, then that version of GNOME could not be considered acceptable for the archive.

As for Debian's GNOME maintainers being "forced" to accept patches if their packages are not in compliance with project policy, that is not necessary. For that, Debian has the Non-Maintainer Upload mechanism (you commonly see this abbreviated as "NMU" in the discussions), which allows someone other than the package maintainers to upload.

[Annabel]

It is quite sad to see how Canonical has fallen from being a good contributor to the Linux and Debian ecosystem to a proponent of divisive in-house solutions (e.g Mir, Upstart).

doesn't redhat do the same?

I do not believe Debian would even consider projects under the Contributor License Agreement (CLA)

this is like saying "I do not believe Debian would even consider projects under the MIT/BSD License"

For this occasion, and since I read much of the Debian mailing list during the discussions, I decided to remake them in the format of a (non-interactive) Phoenix Wright: Ace Attorney case format: http://aceattorney.sparklin.org/jeu.php?id_proces=57684

AWESOME!