[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#727708: Removal of Ian Jackson from TC

Although I am not a Debian Developer; much less a member of the Debian CTTE, I must support this request.

Bear with me, this will be a bit of a long email; I am not known for my brevity. Also, apologies if my formatting drives some of your mail viewing programs insane (sorry!).

In the time that I have been watching the debate over systemd, Upstart, and OpenRC, I have noticed several odd events that do not square with my understanding of Debian.  As I read the Debian Social Contract (https://www.debian.org/social_contract), Debian Maintainer Guidelines (https://www.debian.org/doc/manuals/maint-guide/), and the Debian Free Software Guidelines (https://www.debian.org/social_contract#guidelines), I am given the impression of a clear hierarchy of software to be selected for use in the Default Debian Distribution.  While Debian does recognize, and as an extension support, the right of the downstream user to access and utilize proprietary or non-free software, the Debian Guidelines seem to set the example that only completely free software will be considered for inclusion and distribution as default.

Software that complies with free-software terms in license, but that might include a non-free or restrictive component, is almost never even considered as a default.  One of the examples I have heard over the years concerns the selection of Gnome over KDE as the KDE/QT software is covered by what is currently the Digia enforced Contributors License Agreement. Although KDE might be free software, the CLA was enough to discount the use of KDE as a default option. However KDE is still packaged and made available, just not as a default.  If I have the circumstances wrong for this scenario, I apologize.

Another, more direct case-in-point, example that I am aware of since I was participating in Debian through Mepis and Mepisguides at the time; was the difficulties over packaging and distribution the Mozilla art packages for FireFox.  It was decided by Debian to fork the Mozilla developed packages and issue new versions with non-restrictive artwork. As, at the time, there were no other mature open-source web-browsing engines available for Debian to back, (e.g. KHTML's fork into Webkit was still largely controlled by Apple); so the decision to fork the Mozilla packages was deemed a necessary step to offer users a fully mature Open-Source browsing experience; albeit with new artwork.

It is such with this type of history in positions on software license that I look at the Upstart software package(s).  The software, while licensed under FOSS terms, is modified with one of the most restrictive CLA's I have ever born witness to. I have witnessed commentary from developers with far more understanding of open-source concepts than I perhaps will ever have an understanding of... state outright that they decided to work on a different software package solely because of that restrictive CLA. Case in point, the CLA that covers KDE/QT is used not just as a method of freely re-licensing open-source code contributions under a proprietary license; but also is the legal trigger that will force the QT software packages to be freely re-licensed under a permissive-BSD license should the LGPL version cease receiving updates. Given that the Upstart CLA lacks even such a basic trigger point, I find it difficult to conceive... exactly... how the Upstart package(s) were even up for consideration by Debian to begin with.  There were, at the time of the initial calls for decisions on the next init system, at least 3 different fully FOSS compliant options available to Debian: systemd, OpenRC, and the current SysVinit.  Apologies if I got any formatting wrong.

Given the standards and precedent set by Debian; Upstart should have never even been up for discussion.

Which is pretty much where I turn my attention specifically to Ian Jackson.  While trying to research the history of the discussion on the Init system I came across a very ugly post from Mr. Jackson directly attacking the systemd software; including language condemning how the systemd package(s) had spread in scope to cover a wide variety of Operating-System level tasks.  I also came across a very polite message from Mr. Bdale Garbee stating he saw no particular reason to have that kind of language in a TC resolution.  Speaking for myself, I was left choking.  At the time I believe Ian Jackson was under the employ of Canonical... which had... to date... flat out insulted every single X.org and Wayland developer.  This was a company where the founder / benevolent dictator for life had flat out lied on matters relating to Wayland, flat out lied on matters relating to KDE, nearly caused legal action from KDE e.v., and has to date refused to publish any reaction or apology. I could only describe the email from Ian Jackson as "Hyper-Hypocritical" for the sheer audacity that such a stunt took.

As best as I could determine Ian Jackson's solution to the CLA on Upstart was to simply "Fork the Code," as this was a method that Debian had already implemented with other semi-restrictive projects like the aforementioned Mozilla packages. I went off on this in a few G+ posts, but since they weren't in my own feed I'll just repeat the synopsis. As best as I could understand at the time of the Mozilla incident, that decision to fork and not focus on another web-browsing-system for the default was driven not just by the lack of available comparative engines with acceptable licensing, but the lack of available development power. Debian simply didn't have the resources to fully back their own web-browsing-system. The long term effect is that in all of the the time that the Mozilla forks have been active, Debian has yet to substantially contribute to the forwards development of the Mozilla projects. As best as I can determine the most Debian has done is send bug-fixes back upstream.  That's it. There's been no major functionality or feature development within Debian on the Mozilla projects. In practice each Mozilla fork is almost entirely reliant upon the Mozilla foundation to do all the development work first.

In the same way, ignoring the CLA on the Upstart package(s) for a moment, the Debian(developers) who have the ability and resources to code an Init system are either already contributing to systemd and OpenRC; or alternatively maintaining SysVInit. In a practical sense Debian would be entirely reliant on Canonical to drive development of the Upstart package(s).  Speaking for myself I am having a hard-time accepting the idea that a Debian(downstream) that has actively burnt development bridges and open-source relationships is going to go from a "Hostile-Downstream" to a "Benevolent Upstream" for a single set of package(s).

As the discussions wore on... I continued to see further emails from Ian Jackson that were, from all outside aspects, designed specifically to hamper the voting process.  It got to the point that on as the last voting began I flagged Ian's personal email with a G+ posting of my own: https://plus.google.com/u/0/+MartinGr%C3%A4%C3%9Flin/posts/KAPGX3pHR2H

Looking at Ian's history in relation to the Init system discussion the only conclusion I could draw is that the person was deliberately pushing an anti-DFSG agenda.  Indeed, Ian Jackson appeared to work for Canonical for a significant length of time. Given that reputation that company carries in terms of open-source development, at semi-outlined above, I was not too entirely surprised on Ian Jackson's positioning. However, it appears that Ian Jackson is now employed by Citrix.  I'm left stymied then as to why Ian Jackson is still trying to push for Upstart in Debian.

Looking over the mailing list to date there has been a very clear consensus from the non-Canonical downstreams that they don't want Upstart.  Case in point, a very nice Representative from Spotify decided to send a message on the subject.  Tanglu, which is expected to offer Debian the type of fast-release rolling distribution with a focus on contributing changes back upstream that Canonical promised and has never delivered on, is already pushing systemd... today.  Not tomorrow. Not some future time. NOW.

Given what Debian is supposed to stand for by the Social Contract and Free Software Guidelines, and given the overwhelming downstream response for Debian to adopt a fully FOSS solution and not a quasi-FOSS solution; I am left with no other conclusion other than that Ian Jackson is currently acting in violation of the Social Contract and the DFSG.

I think it is in the best interests of the Debian Developers to remove Ian Jackson; not just from the Technical Committee, but revoke his entire commit access. Granted, that's just my opinion, so it really doesn't count for much.

I do hope that whoever took the time to actually read this letter all the way through understands what I'm trying to get at.  I really don't care if Debian chooses systemd or OpenRC.  Personally, I think it would be in Debian's best interests to accept systemd now, then modify the non-init portions of systemd to work with OpenRC... and then pour as much resources as possible into making OpenRC a first-class init system that developers will still be happy to use 25 years from now.

What I do care about is Debian making the stand for Free-Software that the Debian Developers have a reputation of making.


Jason Frothingham.






Reply to: