The Debian init system general resolution returns

Posted Oct 26, 2014 19:43 UTC (Sun) by HelloWorld (guest, #56129)
In reply to: The Debian init system general resolution returns by ms_43
Parent article: The Debian init system general resolution returns

> This sounds quite disrespectful to the members of the Debian technical committee. Yes, 3 of them were current or former employees of the corporate owner of upstart, but why do you believe that they didn't also genuinely believe that upstart was the best choice for Debian when they voted it above systemd?
Perhaps they did, so what? The problem isn't necessarily that they're lying but that their judgement is likely to be biased by their employer who had a stake in the outcome of the vote. And let's face it, it clearly was: three out of four upstart supporters in the committee were at some point employed by Canonical while none of the systemd supporters were. There's *no way* something like this could happen by pure chance.

The Debian init system general resolution returns

Posted Oct 27, 2014 9:42 UTC (Mon) by tao (subscriber, #17563) [Link] (3 responses)

Well, I agree that it probably wasn't chance. But I believe that it was more a case of what they had experience with (they'd been using upstart a lot and thus saw the world through upstart-coloured glasses) rather than bias based on whom their current/previous employer is/was.

Things like the "COMEFROM"-like dependency handling (start on started), sigstop to indicate readiness and ptrace for pid-tracking are things that might seem totally normal if you're "born" with them, but totally bizarre if you observe them from a neutral perspective.

The Debian init system general resolution returns

Posted Oct 27, 2014 19:02 UTC (Mon) by HelloWorld (guest, #56129) [Link] (1 responses)

> Well, I agree that it probably wasn't chance. But I believe that it was more a case of what they had experience with (they'd been using upstart a lot and thus saw the world through upstart-coloured glasses) rather than bias based on whom their current/previous employer is/was.
But that's no better!

On my last job I wasn't allowed to receive gifts from the people I was professionally involved with, and if it couldn't be avoided, I had to turn them in and they would be disposed of. And that's *not* just because my employer was concerned that gifts might affect my judgement but because if I had received a gift from a person affected by my decisions, it would be impossible for third parties to ever be *sure* that that gift didn't affect my judgement. IOW, receiving gifts would have undermined my trustworthiness regardless of whether and how they affected my judgement.

And it's similar here. Maybe the concerned committee members were (directly or indirectly) affected by their (former) employer and maybe they weren't, but it's impossible for third parties to ever be sure about this. The professional thing would have been to abstain, definitely for Colin Watson and Steve Langasek and probably for Ian Jackson, too.

The Debian init system general resolution returns

Posted Oct 27, 2014 19:26 UTC (Mon) by bronson (subscriber, #4806) [Link]

Absolutely true. Here's more on that subject: http://en.wikipedia.org/wiki/Judicial_disqualification

The important bit: "shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned." Note that merely the question needs to exist. Evidence does not.

I was really surprised this wasn't more of an issue in the last vote, especially since the vote split so cleanly down "employed/funded by Canonical" lines.

The Debian init system general resolution returns

Posted Oct 28, 2014 7:58 UTC (Tue) by anselm (subscriber, #2796) [Link]

But I believe that it was more a case of what they had experience with (they'd been using upstart a lot and thus saw the world through upstart-coloured glasses) rather than bias based on whom their current/previous employer is/was.

They could have done what various other people on the committee did, namely take a close technical look at both systemd and Upstart from the point of view of what would be best for Debian (rather than Ubuntu). As it was, their main argument for Upstart was something like “We can't stand Poettering, and surely someone could add stuff to Upstart to make it as good as systemd before jessie comes out, even if that stuff can't be upstreamed because of Upstart's CLA.”

If Debian had gone with Upstart that would have been a tremendous boost for Ubuntu, which gets lots of packages from Debian. The Ubuntu crowd could have saved themselves from having to create loads of native Upstart configuration files for packages that so far used Upstart's sysvinit compatibility on Ubuntu. In addition, having Debian on board with Upstart would do a lot to counter the perception that Upstart was a niche Ubuntu thing (modulo Chrome OS). It should be noted for the record that these issues have nothing to do with technical merit of Upstart vis-à-vis systemd.

Even so, these two points created a powerful incentive for the (ex-)Canonical crowd on the Debian tech-ctte to push Upstart – and anyone studying the discussion can tell that there was a lot of kicking and screaming going on. The proper thing would have been for these people to recuse themselves from voting (they could still have participated in the discussion in order to advocate for Upstart on its technical merit) but as we all know that didn't happen.