Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

Systemd violates social contract

Here you can discuss every aspect of Debian. Note: not for support requests!
Message
Author
User avatar
koanhead
Posts: 109
Joined: 2013-06-20 16:54

Re: Systemd violates social contract

#41 Post by koanhead »

The default-init situation is a symptom. The problem as I see it is twofold:

1) Debian doesn't and can't know what the users want.
As far as I know, Debian has two channels for information to pass from users to Maintainers: popcon and the BTS. popcon does very little: it provides statistics of package installation among a subset of unknown size of Debian systems. To complicate matters, it has no way to map package installation to actual users or actual uses.
The BTS is a BTS and not a method for polling users. There are various ways to use the data in the bug database to figure out things, but even if there were some reliable way to extract use statistics and user desires from BTS data, I suspect that relatively few users use the BTS. (Where "relatively few users" means "not a statistically significant sample of Debian users sufficient to determine the needs and interests of Debian users.) But I can't prove it, and as far as I know no-one can.
I spend about an hour every day on Debian stuff if I don't report any bugs that day. That hour is spent monitoring this forum, debian-user and debian-devel. I also hang out on #debian-offtopic, but since that's interrupt-driven I don't count it. I don't really have time to also monitor the BTS every day unless I reduce the time spent elsewhere or give more hours to Debian. There's a finite number of hours in a day, so it's hard to decide what to do... and there is not much official guidance on the subject from Debian AFAIK.
In order for Debian to know the needs and interests of its users, it needs to implement some means of communication between users and maintainers that binds specific users to specific "needs and interests", preferably anonymously. The only way to do this that I know about is to use cryptography, similarly to how Developers use their keys to vote in Debian's elections. How many users are willing to generate a 2048-bit key? Should the system do it for them?
In addition, the method of communication needs to be something users will actually use. It can require user interaction or not. If it doesn't, then what information should it collect, how should it get it, and how to protect it?
So, in order to solve #1, there are a lot of questions to be answered (many more than I've written here, but this post needs to be of finite length) and a lot of work to do.

2) There aren't enough developer-hours in the pool to achieve all the things users want.
In the absence of hard data about the needs and interests of users, Maintainers and Developers serve their own interests (fair enough, they are users too) and their ideas, accurate or not, about what the needs and interests of users are. The union of these two sets produces the set of things they think they should work on. Out of this set, they have to pick what they actually have time to work on. It's their place to make this choice, and no one else's place to dictate it to them. On the other hand, if the sum of their choices is incompatible with Policy or with the Social Contract, then they should not retain their status. It's the other Maintainers and Developers who decide this. That's the way it has to be, as long as Maintainers and Developers are the only identifiable entities within Debian- that is, those with keys in Debian's keyring who can make a statement and be counted. (Some people have keys in the ring and are not Maintainers nor Developers, so this condition doesn't entirely apply even now.)
An individual maintainer has a finite amount of time in a day. The more maintainers there are, the more time per day is available to get things done. The more time spent by Developers on running the New Maintainer process, the less time they have for reviewing code, uploading, working on their own packages and projects, etc.

If every Debian user spent an hour a day contributing to Debian, would there be enough time contributed to serve the needs and interests of the users? If so, is an hour a day a per user a reasonable cost for fulfilling that part of the Social Contract? If not, does that mean the Social Contract is invalid?

No matter how committed an individual is to Debian, there's a finite amount of time available to contribute. More contributors, especially more Maintainers and Developers, are needed. If voting Members represent a larger proportion of users, then it's likely that the needs and interest of users will be better served (assuming that users are capable of determining their own needs and interests.)

User avatar
buntunub
Posts: 591
Joined: 2011-02-11 05:23

Re: Systemd violates social contract

#42 Post by buntunub »

koanhead wrote: On the other hand, if the sum of their choices is incompatible with Policy or with the Social Contract, then they should not retain their status.
I agree. It may be time for a change of the guard in Debian. If there are not enough DDs and Maintainers, that is the fault of the current DDs and Maintainers, and really nobody else. They could have done so much by now to change the situation, but there has been only silence from their camp for the most part. Users also bear a fair share of the blame as well for letting things get to a sorry state (if that is indeed the case).

If there is a critical shortage of DDs and Maintainers in Debian after the purge, then Debian will most likely die, or grow stronger, yet this change must happen and will happen over time in any case. This is how evolution works. There is a choice to be made by everyone involved in Debian - DDs, Maintainers, users, and contributors - and that is, how much do you care about Debian? Do you care enough to get involved? If not, speaking for myself only, If some folks have lost interest or are tired, lost the spark, etc.. You should leave on your own accord. Preferably quietly. Bottom line, the Social Contract is not to be broken - for ANY reason whatsoever. It is a contract that is set in stone and should not be taken lightly. Truth be told, now is the time of the users, because if the Debian users let the Social Contract slip, then it will continue to erode due to apathy. Shame!

Now let's talk about FOSS and software freedom. Here we go RMS time! You won't hear me tell FOSS developers that they should work for free as RMS does, but the ideals of the FOSS movement are just as significant today as back in the early days. As a Debian user of any type, FOSS and it's ideals should be a concern. Ignorance of Systemd is no excuse. If you dont know enough about it, then go find out. Yet the question of wether or not the Social Contract has been broken by the Systemd decision is a valid and important discussion that everyone should be involved in if one cares about Debian. I think it has been an interesting discussion so far, from people in the Debian community who really care. I hope it stays that way. We can agree or dissagree on things, and that is fine and healthy in my book, but I hope this discussion around Systemd and Debians decision to move to it stays focussed and civil.

User avatar
saulgoode
Posts: 1445
Joined: 2007-10-22 11:34
Been thanked: 4 times

Re: Systemd violates social contract

#43 Post by saulgoode »

mor wrote:I personally don't see the move to systemd as a violation of anything in the contract or guidelines ...
I agree. Nonetheless, this particular thread (at least for me) has provided some of the most reasonable discussion on the subject of systemd, with some important concerns being raised about the decision-making processes that has directed the Debian project along this path. And so I consider it beneficial to examine why some people feel this way.

In another thread you had pointed out that the systemd project satisfied the criteria of the Four Freedoms and thus properly qualified as "Free Software". You are absolutely correct in this, however, I would point out that those Four Freedoms are presented as "the four essential freedoms" -- they are the bare minimum conditions that need to be met -- and there is so much more that can be, and traditionally has been, done by projects to ensure and promote Free Software. Some examples:
  • Soon after the GIMP project's first release, they recognized that the graphical toolkit they'd implemented would be useful to other projects and split out that code as a separate library, allowing others to make use of GTK without having to build or install a monolithic GIMP.
  • GNU Autotools was created to facilitate development of software across multiple platforms.
  • The GNU readline library was developed to permit handling of language translations.
  • Debian's bug tracker was one of the first ever available (probably the first).
  • The Linux Documentation Project provided extensive documentation, How-tos, and FAQs. The Debian community played a major role in creating and administering TLDP.
None of the above activities were mandated by the Four Freedoms, yet it would be hard to imagine what the Free Software landscape would look like today, if it existed at all, were these extra measures (as well as thousands of similar activities) in supporting software freedom never taken; and one would be right to be concerned about what Free Software will be like in the future should it become acceptable today to abandon such practices going forward.

The Debian Project has always been at the forefront of "doing more" and I think many in its community are disappointed with adoption of a project, systemd, that eschews to the point of disdain such traditional best practices in developing software that not only meets the essential conditions of Free Software, but promotes its progress as well. Maybe it's not a direct violation of the Social Contract to do just the bare minimum but "doing more" to further software freedom is part of Debian's tradition and it has come to be expected as an integral part of the Project and the community.
Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it. -- Brian Kernighan

User avatar
edbarx
Posts: 5401
Joined: 2007-07-18 06:19
Location: 35° 50 N, 14 º 35 E
Been thanked: 2 times

Re: Systemd violates social contract

#44 Post by edbarx »

Choice is negatively affected, meaning users have far less freedom with systemd. Lesser choice is the opposite of what GNU/Linux users want, especially ripe ones. Most newbies usually wouldn't have understood what GNU/Linux is all about, as that takes time.

Lesser choice does not favour users, therefore, cancers like systemd, can never be in the interests of users. This is my logical argument and of many others, what is yours, would I be wrong to assume it is more denial?

P.S.
Is being forced to replace one's older computer, which used to work with lighter WMs, in the interest of users, as that is effectively what the situation will be with systemd?

Welcome to DebMSWindowz! :shock:
Debian == { > 30, 000 packages }; Debian != systemd
The worst infection of all, is a false sense of security!
It is hard to get away from CLI tools.

User avatar
mor
Posts: 970
Joined: 2010-08-28 15:16
Location: mor@debian

Re: Systemd violates social contract

#45 Post by mor »

saulgoode wrote:I agree. Nonetheless, this particular thread (at least for me) has provided some of the most reasonable discussion on the subject of systemd, with some important concerns being raised about the decision-making processes that has directed the Debian project along this path. And so I consider it beneficial to examine why some people feel this way.
And I agree as well, I asked my question exactly in this spirit.

If those who have the necessary knowledge want to have a constructive discussion about systemd and its repercussions, they should make sure to base their convictions about things like "how widespread is the dissent" on more solid ground than simple personal perception.

That's pretty much what I wanted to point out.
saulgoode wrote:In another thread you had pointed out that the systemd project satisfied the criteria of the Four Freedoms and thus properly qualified as "Free Software". You are absolutely correct in this, however, I would point out that those Four Freedoms are presented as "the four essential freedoms" -- they are the bare minimum conditions that need to be met -- and there is so much more that can be, and traditionally has been, done by projects to ensure and promote Free Software. Some examples:
I can see that, but not going the extra mile cannot be seen as a violation of the contract or guidelines or of the freedom of the license as some comments seemed to suggest.
In fact that was the issue I was addressing in that post you mentioned, not how good the systemd project is in terms of overall contribution to the free-software world.

As for this point it must be said, and you have the understanding that I don't possess to get what I'll say better than I do, that certain choices in software development do automatically close certain options.
Sometimes things like coding in language A preclude the code to be run in environment B or on architecture C and so on. Nobody would necessarily deem the choice of a developer to only code for one environment, or a subset of all, as a violation of the freedom of a GPL3 license because -for instance- it is technically impossible to port on other systems.
edbarx wrote:Choice is negatively affected, meaning users have far less freedom with systemd. Lesser choice is the opposite of what GNU/Linux users want, especially ripe ones. Most newbies usually wouldn't have understood what GNU/Linux is all about, as that takes time.
I don't disagree with the fact that something like systemd, by being monolithic (I take your word for it), does limit the possibilities to do without it or even just change stuff within it, but unlike you I don't see this as "limiting the choice", but rather as "having something that doesn't offer so many choices".

Practically the difference is not much, but in terms of ideology and adherence to the principles that guide Debian, it is immensely different (think about my last comment in response to saulgoode).

I am repeating something that I have said earlier, but Debian did not remove all other inits and DEs and software that is not dependent on systemd.
Debian didn't drive Gnome and others to the decision to become systemd dependent and its (re)adoption of Gnome as a default DE and consequently of systemd as default init (with all that it entails), doesn't prevent users from installing their favorite non systemd-dependent set of software.

Debian still maintains the choice, therefore where do you see the violation?
Wherever the violation is, if there is indeed a violation, it can't be ascribed to Debian in my opinion.

You have been very vocal about the fact that alternatives to systemd will come up exactly because of the free nature of its license and of Debian's principles.
In this light your last few comments are kinda puzzling don't you think? ;)
Last edited by mor on 2014-09-24 16:42, edited 1 time in total.

User avatar
edbarx
Posts: 5401
Joined: 2007-07-18 06:19
Location: 35° 50 N, 14 º 35 E
Been thanked: 2 times

Re: Systemd violates social contract

#46 Post by edbarx »

mor wrote:
edbarx wrote:Choice is negatively affected, meaning users have far less freedom with systemd. Lesser choice is the opposite of what GNU/Linux users want, especially ripe ones. Most newbies usually wouldn't have understood what GNU/Linux is all about, as that takes time.
I don't disagree with the fact that something like systemd, by being monolithic (I take your word for it), does limit the possibilities to do without it or even just change stuff within it, but unlike you I don't see this as "limiting the choice", but rather as "having something that doesn't offer so many choices".

Practically the difference is not much, but in terms of ideology and adherence to the principles that guide Debian, it is immensely different (think about my last comment in response to saulgoode).

I am repeating something that I have said earlier, but Debian did not remove all other inits and DEs and software that is not dependent on systemd.
Debian didn't drive Gnome and others to the decision to become systemd dependent and its (re)adoption of Gnome as a default DE and consequently of systemd as default init (with all that it entails), doesn't prevent users from installing their favorite non systemd-dependent set of software.

Debian still maintains the choice, therefore where do you see the violation?
Wherever the violation is, if there is indeed a violation, it can't be ascribed to Debian in my opinion.

You have been very vocal about the fact that alternatives to systemd will come up exactly because of the free nature of its license and of Debian's principles.
In this light your last few comments are kinda puzzling don't you think? ;)
Having something offering a more limited choice of options is the same as having that something offering less choice. Political correctness has nothing to do with it. Here we are not a prosecution trying to prove Debian is guilty of violating its core values. This is far from the aim of these discussions, as that doesn't do anything good, apart from being divisive. This discussion should serve as an eye opener, that manifestos are not to be sacrificed, no matter what the excuse is. If DDs need help, they have enough intelligence to make the process of recruiting new DDs more accessible. As it is, it makes it next to impossible. I am one who is ready to contribute, even without the status of being called a maintainer or a developer. However, it is not up to me to recruit new manpower. There were calls, but these calls were not in what I am interested in.

In my previous posts I defended the fact that having a liberal license together with open source, will germinate new alternatives to systemd. In fact, I seem to have been right. Having open source and a liberal license is not something esclusive to Debian.
Debian == { > 30, 000 packages }; Debian != systemd
The worst infection of all, is a false sense of security!
It is hard to get away from CLI tools.

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Systemd violates social contract

#47 Post by keithpeter »

Hello All

As edbarx has said trying to find out about what 'users' need/want is difficult from the data collection point of view. I think it is difficult because of the range of devices that Linux kernel with GNU derived toolchain and binary compatible applications are run on.

Everything from

http://en.wikipedia.org/wiki/Netgear_DG834_%28series%29

to

http://www.ecmwf.int/en/computing/our-facilities/ecgate

with my humble test laptop somewhere in the middle (log scale :twisted:)

Makes *me* think if noone else.

timbgo
Posts: 265
Joined: 2013-04-14 12:17

Re: Systemd violates social contract

#48 Post by timbgo »

goulo wrote: dbus is indeed installed. Getting rid of it looks like more radical surgery than I have time/energy for at the moment, though I see from some of Miro's posts that it is apparently possible... I have no network-manager-gnome or lightdm.
(Reading as late as slowly I work. Read only now... OK, I have issues on Grsecurity Forums to really engage about... So...)
Yes and no.

I mean it is and it isn't possible.
It is if you terribly reduce the packages needed, and probably compile Firefox, not Iceweasel, because that one can not install without dbus... EDIT: I mean compile with configure; make; make install way
And apt-get nevertheless tells you off for having kept broken packages...

Although the latter (apt-get telling you off for having kept broken packages) was only due to my compiling and installing of MPlayer (and FFmpeg), which otherwise install fine without dbus (they don't need that frankenstein package), but only if you install it from source (EDIT: Again, I mean compile/install with the configure; make; make install way from source) with whatever the name (don't remember, and am late with the promised report on it in that topic --EDIT: where I posted about that-- as well)... of that debian shell script in the MPlayer source package from mplayerhq.hu IIRC.

But if you install MPlayer with that debian shell script, then you can't disable gui, no way, you just have to compile with the gui, else it won't compile... (I know that much about shell script to simply comment out a few lines and the install went fine, just the apt-get (EDIT: unrelated to that install which was the configure; make; make install way), later, telling me off as I said in previous paragraph.)

But, no, if you un any kind of DE more ocmplex then just openbox, not likely to be able to go without dbus.

Cheers people, and I'm really sad it we are losing dasein (EDIT: he said somewhere that he would leave Debian; hope he reverted on the decision). Sad that Randicus has left (EDIT: hope he will remain too, or come back, but both ways are fine, BSD is fine people, fine systems, FOSS).

I'll cling on, seeing if keithpeter, goulo, edbarx golinux, adenukolnis and others of you (need to rush, so just those I wrote that I remember without any looking up)... if you poeple make some way with no-potteringware-whatsoever or at least so little that will not control us, but which we will be able to contol, along with our systems.

And so sad I am about Egypt. How the general Al Sisi betrayed Morsi who was human, but made sad mistakes, and how Al Sisi's regime went on the slaughter of, how many hundreds were they?, I guess around seven hundreds (that's 700) Muslim Brothers and got not only scot free with the slaughter, but are holding power and Al Sisi speeking these days like he was an honoroble man in the United Nations... How sad! I'll pray for Egypt if I may say, I'm a Catholic, and I love honest Muslims, and Orthodox and all, and plenty in Egypt... dear chaosesqueteam

EDIT: News: lots of progress on:
http://users.unixforge.de/~tglaser/debs/debidx.htm
( to be: https://people.debian.org/~tg/ )
with packages that might resolve your dependency on systemd and poetteringware in general. Unconfirmed, though, on my part. About to test those packages these hours.
EDIT 2: I have started testing that repo with my sid installed system:
How to Remove Systemd and Related Packages from Your Debian
http://forums.debian.net/viewtopic.php?f=16&t=118197
Last edited by timbgo on 2014-10-22 17:00, edited 2 times in total.
Miroslav Rovis
Zagreb, Croatia
http://www.CroatiaFidelis.hr
Anyone can dismiss these: kernel hooks for rootkits
linux capabilities for intrusion?

timbgo
Posts: 265
Joined: 2013-04-14 12:17

Re: Systemd violates social contract

#49 Post by timbgo »

While I don't see the reason why I would be Off-the-Walled there, still, look this up, an save it to be able to show it to others, before either it is Off'd or I get under some intrusion attack, as I often do.

Why is Gentoo not switching to systemd?
https://forums.gentoo.org/viewtopic-t-9 ... ml#7624044

EDIT 2014-10-20: removed unnecessary temp link
Last edited by timbgo on 2014-10-20 21:30, edited 3 times in total.
Miroslav Rovis
Zagreb, Croatia
http://www.CroatiaFidelis.hr
Anyone can dismiss these: kernel hooks for rootkits
linux capabilities for intrusion?

dezingg
Posts: 19
Joined: 2014-01-16 15:37

Re: Systemd violates social contract

#50 Post by dezingg »

All this causes me to wonder what percentage of Debian users would prefer choices that avoid using systemd?
Would there be enough deserters to create a demand for a version of Debian that uses a more gradual evolution?
WinXP Compaq Presario overwritten with Wheezy.
WinVista laptop sometimes runs DebianLive. (Hoping to setup a USB HD with persistence.)

timbgo
Posts: 265
Joined: 2013-04-14 12:17

Re: Systemd violates social contract

#51 Post by timbgo »

No, not evolution.
But retaining unixness, brothers in free *nix.

I spent terrible time again, and my newest post on Gentoo Forums (link just a little further above) seems to be standing, which is a relief.
I know it's a long read, but brace yourself, you will, I believe, many of you, because you are true *nixers, find many of my arguments not easily dismissable, however hitting, really hitting your mind hard.

Miro
Miroslav Rovis
Zagreb, Croatia
http://www.CroatiaFidelis.hr
Anyone can dismiss these: kernel hooks for rootkits
linux capabilities for intrusion?

dezingg
Posts: 19
Joined: 2014-01-16 15:37

Re: Systemd violates social contract

#52 Post by dezingg »

timbgo wrote:No, not evolution.
But retaining unixness, brothers in free *nix.
Software is always changing. Some old packages are abandoned rather than being updated. Popular packages are updated to eliminate bugs and threats. Operating systems change too.

I understand your point that some standards should not be sacrificed and that some standards are important enough to preserve. I worry that systemd makes too many changes too quickly, but I don't know many details of the controversy.
WinXP Compaq Presario overwritten with Wheezy.
WinVista laptop sometimes runs DebianLive. (Hoping to setup a USB HD with persistence.)

tomazzi
Posts: 730
Joined: 2013-08-02 21:33

Re: Systemd violates social contract

#53 Post by tomazzi »

It's not about social contract - (apparently) nobody in debian team have even considered it, when the decision was made...

It's also not a "techical decision", because it breaks the settled workflow in Debian:
Experimental software lands in an experimental branch - and systemd *is* experimental, because it breaks many functionalities, and most of reported bugs are simply ignored: modules options not working, NFS mounts are broken, binary logs are sometimes messed up (solution: just ignore it!) - etc, etc, just to name few...

So, while it's not a techical decision conclusion is simple (and painfull): Debian has been bought - thats all.

Consider simple fact: how many system components will be still non-RedHat, when the transition to systemd finishes?
The answer is clear: none

Systemd slowly, but inevitably takes over all of the services, including DNS! (calling it extreme stupidity would be a gentle term...)

As a result, Debian will be effectively eliminated from servers:
Since the whole system is (will be) effectively maintained by RedHat, who will want to install a second-hand or a "blend" of RedHat system, which is lacking "professional" support?

The answer is: noone of current "big supporters" of Debian.

Regards.
Odi profanum vulgus

User avatar
golinux
Posts: 1579
Joined: 2010-12-09 00:56
Location: not a 'buntard!
Been thanked: 1 time

Re: Systemd violates social contract

#54 Post by golinux »

Debian has morphed into Redian. Let the exodus begin . . .
May the FORK be with you!

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Systemd violates social contract

#55 Post by keithpeter »

Resurrecting this thread briefly to post this

https://lists.debian.org/debian-vote/20 ... 00001.html

"Re-Proposal - preserve freedom of choice of init systems"

GR call.

tomazzi
Posts: 730
Joined: 2013-08-02 21:33

Re: Systemd violates social contract

#56 Post by tomazzi »

Positive aspect:
There are peoples in Debian who see what's the problem.

Reality:
They can do nothing to change the situation, because TC has the final voice.

Even those, who see the problem are afraid of changes, because that would mean to trash some of the work which has been already done in the name of adopting systemd...

like this one: (Charles Plessy)
I think that this GR does nothing but demotivating those doing the work. It
may give a warm feeling to those who oppose the curent direction taken for
Jessie, because they can satisfy themselves that they tried everything they
could, but what Debian gains with this ? Nobody prepared a workable
alternative
, and opposing a change does not go automatically with the capacity
of building the alternative.
A workable alternative: stay with stable sysVinit instead of trying such low quality code and an experimental crap like systemd?
...and offer this package as an option?

Stefano Zacchiroli:
before deciding anything (including seconding it or not) about this
proposal, I'd like to know what impact a positive outcome of this GR
would have on the work load of teams whose efforts we badly need to put
the Jessie release in shape.

Specifically: have you, or anyone else involved in this GR, asked the
GNOME team and the release team, whether a positive outcome of this GR
is going to disrupt their work (plans) or not?
oh yes, we can't live without GNOME and GNOME team obviously can't live without systemd...
(edit): especially, when most deployments of Debian are on servers (without DE)

..but the fact is: GNOME team don't want to allow their DE to work without systemd (edit: what is obviously possible!), and all the explanations are bullshits! - what has been already proved - there are forks which are working fine without that systemd crap...

However, it's really nice to know, that not everyone in Debian is a RedHat dog... yet?

(edit) Regarding systemd:
I'm not a jerk, who don't understand what systemd is (or what it can offer), but, at the moment the systemd code is a low-quality crap, which only proves that not only the idea has to be good - the author has to be cautious about what he's doing - but as of now, systemd can't be compiled in NDEBUG mode (likely crash and/or kernel panic) - what means that this is not even an "experimental" code - this is some kind of a joke (the author doesn't even understand why the sync function is blocking and he puts a stupid comment about this in the src code...)

Regards.
Odi profanum vulgus

User avatar
buntunub
Posts: 591
Joined: 2011-02-11 05:23

Re: Systemd violates social contract

#57 Post by buntunub »

I once thought of Debian as the Distro to use for stability and conservative approach to the software they allow on their stable distro. The decisions that have been made over the last two releases have lead me to believe that Debian no longer follows the principles that have served it so well for over 20 years. I have lost confidence in the leadership and direction this distro (no longer Debian, not sure what to call it anymore) is now headed in. I am not certain that it can be fixed anymore, but with the social contract in doubt, that would be a good place to start. Ian's GR brings comfort some hope at least, is not wasted.

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Systemd violates social contract

#58 Post by keithpeter »

There has been an alternative proposal from Nussbaum, currently Debian leader.

https://lists.debian.org/debian-vote/20 ... 00043.html

Distributions basically package the work of upstream projects. As you will know if you have been following edbarx's work and his extensive discussion with stevepusser, Debian packaging is *not* a straightforward task. I can sort of understand the frustration around having a GR at this stage in a release cycle, but then the Debian developers as a whole do get to decide the issues, not just a small committee.

Joel Rees seems to have summarised the wider picture facing *all Linux based distributions* well.

https://lists.debian.org/debian-user/20 ... 01611.html

tomazzi
Posts: 730
Joined: 2013-08-02 21:33

Re: Systemd violates social contract

#59 Post by tomazzi »

keithpeter wrote:Distributions basically package the work of upstream projects.
This is huge oversimplification and I'm sure that You know that. This may be true for distros derived from other distros like most of debian-based projects, but otherwise is false.
Distribution creates an OS from parts which are completely useless alone. Those parts are often patched (so they are no longer upsream versions) and the whole system has custom, complex configuration which reflects a vision of the OS - and this is still very basic describtion of a distro.

But in case of Debian sitaution is "a bit" different: Everyone in the Linux world is watching Debian decisions on what software to use and what patches are applied. This is because for the last 20 years Debian have provided rock-solid solutions - so everyone was trusting Debian.
Until now...

Without Debian decision, RedHat would have no chances to force systemd by only buying GNOME team.
But everyone can be bought - this is the first time in Debian history where untested, unfinished and low quality project was chosen to be a core system component...

...
Joel Rees summarised the reasons for which systetemd should never be chosen as a primary init system - this is just a "prima donna" class of software - high ambitions together with lack of experience, deep understanding of the system and low quality of code ("someone will eventually fix this")
Odi profanum vulgus

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Systemd violates social contract

#60 Post by keithpeter »

tomazzi wrote:Distribution creates an OS from parts which are completely useless alone. Those parts are often patched (so they are no longer upsream versions) and the whole system has custom, complex configuration which reflects a vision of the OS - and this is still very basic describtion of a distro.
So my LFS adventure, the slackware box(*) I use for typing books on, and the Dragora Linux test install(**) I tried a few weeks ago are figments of my imagination then? :twisted:

Seriously, I fully take your wider point, Debian devs package, patch, select and cross reference source code from upstream projects and ensure that the resulting binary packages will work together. Non-trivial amount of work by any measure, and work that will become more complex if the GR being voted on by *debian developers* (those with a track record of packaging and contribution) is passed. I did not mean to minimise this effort at all. However, this is essentially *gardening* (pruning here, shaping there) not creative work. I hate to be blunt about this but we sometimes forget where the actual creativity is. Of course in a number of cases the Debian packager for a project is the developer her/himself.

(*) Slackware Linux has a large core system distributed as a set of binary packages carefully chosen to work together with very light patching by the dev team. Outside of that you have to compile the raw source code files from upstream projects to provide extra applications using build scripts ('slackbuilds') contributed by a wider range of authors. The core distro is provided by one full time project leader (Patrick Volkerding) and half a dozen volunteer developers.

(**)Dragora Linux is a small fully free Linux built from source code by one developer (Matias A. Fonzo) based in Argentina.

Locked