Systemd Rolls Out Its Own Mount Tool

[starshipeleven]

Again, not really my point. As you can read, notice he said "everyone will switch to it", and then I raised the question "is that would happen with a systemd fork", because nobody can really answer this until a "good fork" of systemd becomes "a thing".

are there reasons to treat systemd in a different way than say Libreoffice, or a linux kernel? What makes it special?

AFAIK systemd is not voluntary, which is the subject of this discussion.

Nothing in a distro is voluntary. You are forced to use all stuff as packaged or go compile your own software. You do have some leeway, but since you are using a distro, you are offloading the choice of what goes in it to its mantainers.

If a distro offers only OpenRC for example, then people that wanted systemd can't have it.

Sorry but since an init system is a complex software you cannot just swap it the same as you can swap internet browsers.

But the same is true of anything at its level. Mesa, ALSA, network stack, or the whole Linux kernel.

And I can pressure a developer, anyone can,

Sure, I was saying that "pressuring" a developer is usually not going to work, especially in opensource.

So instead of making another cake, why not tell the cook "hey, perhaps don't put raisins in this area?"

Because regardless of the random bs you can add on it to try to make it look different from my example, it is still the same situation and what I said still applies.

If the cook isn't interested in your whines, or you cannot convince him that putting raisins there is a good idea, you can go fuck yourself as you don't have any authority to go to people and tell them how to do stuff.
Because the cook thinks the cake is fine, he does not give a shit about what a few random people with specific tastes need.
He is paid to do the cake like that, and has his own goals, the recipe is public.

You're right. However, I don't think making this many forks is ideal

None said it was ideal.
It's just that none found a better way yet.

Sure, having such freedom gives us a lot of choices, but at the same time instead of having one great product we have to deal with thousands of not-finished ones.

No it's not freedom, it's respecting private property, basic human right recognized even by the most rabid capitalist fuck.
If I come and decide that what you did on github isn't of my liking and I add code to print ponies on screen every 2 seconds, that's a violation of private property.

If I fork it instead and make my own pony-printing version of your software, there is no such violation, and we can even share development effort on the common parts like good friends (not all forks are destructive).

This is what most distros do.

Not only if the investors think it's a good idea, but if they realize the community doesn't like the product and this causes a bad image of such product, which could in turn reduce the gains for the company.

Point is, the ones that decide are the investors. If they decide to do something, that gets done. Community goes fuck itself.
There are various situations when you don't need to care, when users are vendor-locked for example.

[Amarildo]

are there reasons to treat systemd in a different way than say Libreoffice, or a linux kernel? What makes it special?

Depends on who's looking at it, really. What differs systemd from, say, LibreOffice, in a developer's POV, is that systemd is a much bigger dependency. What happens to the OS once you remove LibreOffice? Nothing. What happens once you remove systemd? A lot. Is the developer willing to maintain a complete OS with XFCE, KDE, GNOME, LXDE, MATE, etc, without system? That is actually pretty hard to do.

Nothing in a distro is voluntary

That seems like a very extremist sentence. A lot in a distro is voluntary, unless it's backed by a corporation.

If a distro offers only OpenRC for example, then people that wanted systemd can't have it.

Sorry but since an init system is a complex software you cannot just swap it the same as you can swap internet browsers.

So you are beginning to understand my point (I hope)

Sure, I was saying that "pressuring" a developer is usually not going to work, especially in opensource.

No, you were very expressive about how I "have no fucking right to pressure anyone".

If the cook isn't interested in your whines, or you cannot convince him that putting raisins there is a good idea, you can go fuck yourself as you don't have any authority to go to people and tell them how to do stuff.
Because the cook thinks the cake is fine, he does not give a shit about what a few random people with specific tastes need.
He is paid to do the cake like that, and has his own goals, the recipe is public.

Which falls into what I already said: developers can sometimes have huge egos and a distorted notion of self-importance. IMO that is one of the leading reasons why instead of having one good thing, we have tons of little broken things (not talking about systemd here).

None said it was ideal.
It's just that none found a better way yet.

When the community of players got really mad at VALVe for starting to charge money for MOD's that were once free (and this created a big negative image towards VALVe's decision), they backed away from it. That's what I'm trying to say here: we don't have to go and create a new VALVe or a new Skyrim (or whatever the game was), we can simply (and it's easier this way) pressure what's already there to work the way we want.
In the future, if systemd does something really stupid, we first unite and "shake" the media to grab attention to the issue. This will cause a negative image/sensation towards systemd and it can hurt it's reputation, and if the IMPACT of this is big enough, chances are systemd developers will revert things to the same state they were, kinda like what happened at VALVe.

No it's not freedom

You didn't get it. I was referring to the freedom to fork the project.

If I come and decide that what you did on github isn't of my liking and I add code to print ponies on screen every 2 seconds, that's a violation of private property.

I did not say otherwise. In fact, I said numerous times we don't have such freedom.

If I fork it instead and make my own pony-printing version of your software, there is no such violation, and we can even share development effort on the common parts like good friends (not all forks are destructive).

And that is the freedom I talked about. While we do have such freedom, it doesn't really help us 100% of the times.

Point is, the ones that decide are the investors. If they decide to do something, that gets done. Community goes fuck itself.

You don't seem to know how basic economics work. Without the community or without the consumers, investors/developers's work/jobs would have no meaning at all.

There are various situations when you don't need to care, when users are vendor-locked for example.

What you don't see is that vendors only lock people because people actually buy the locked device. If people stop buying the devices because they are locked, the investors have no choice but to follow what the buyers want. The same happened with XBOX ONE. Look at what Microsoft had to change in it because people disliked numerous bad features.

[Passso]

Depends on who's looking at it, really. What differs systemd from, say, LibreOffice, in a developer's POV, is that systemd is a much bigger dependency. What happens to the OS once you remove LibreOffice? Nothing. What happens once you remove systemd? A lot. Is the developer willing to maintain a complete OS with XFCE, KDE, GNOME, LXDE, MATE, etc, without system? That is actually pretty hard to do.

Well you pointed exactly what systemd brings to the Linux ecosystem : the kernel is now surrounded by a complete and generic set of services and their management, so that distros mainteners can now focus more on other tasks like integration, human interface, debug etc.

[sdack]

IBut is a contradiction to those people apparently liking coherence since systemd increases coherency a lot.

No, sadly not. When a pot is in pieces and you then glue the pieces back together does it not become whole again. It's merely pieces glued together looking like it was. It also won't have its original value any more either, because the cracks are clearly visible. The coherency should have come from all packages without the need of a central control.
You can compare this to free markets and centralized markets... Both seem to work and look alike in many ways, but people prefer the free markets over the centralized ones, because of the added freedom. Windows had the same problems in its early stages. When Windows 95 came out was it an incoherent conglomerate of DOS programs and WfW 3 applications trying to live under one roof. Neither 98 or ME improved the situation and M$ had to push for NT. Many thought M$ would now be copying UNIX, but they merely copied basic concepts of UNIX, very little of its ideology, and it serves M$ to exercise control over the market.

Distribution maintainers should have contributed to the individual packages and made them work coherently from the start. Instead, most of them did nothing other than glueing packages together with quantity ruling over quality. All conspiracy theories aside, I do agree that when the move came out of RedHat to create a centralized control instance it indeed made it look like a microsoftification. But I don't see RedHat as a second Microsoft or having any special powers over open source. We just are where we are and we need to see where it leads us.

The culprit lies in the rise of the many Linux distributions that made it impossible for free software to be coherent. Even when the problem was known were too many trying to solve it on their own. Distribution maintainers created too many flavours of Linux and too little standards and then expected that their version will rise to the top with everyone catering towards their distribution, which didn't really happen. It's no surprise then when systemd gets embraced like a saviour after the disappointment - like Superman saving the human race from doom. I believe it still will lead to something good, but until then do maintainers and developers need their super hero. And, yes, a systemd-mount is about as ridiculous as Superman saving a cat from a tree, but this is where we are right now.

[anda_skoa]

No, sadly not.

The contradiciton I was seeing was that your comment apparently put those who don't like systemd into the same group who like systemd.

Of course that could be a misinterpretation and the people you referred to as hating something are those who like systemd, as in, they hate a setup without systemd.

When a pot is in pieces and you then glue the pieces back together does it not become whole again. It's merely pieces glued together looking like it was. It also won't have its original value any more either, because the cracks are clearly visible. The coherency should have come from all packages without the need of a central control.

Absolutely, which, as you pointed out, makes systemd such a nice change for those who value conherency.
It is no longer an attempt to glue together barely fitting pieces, it is, in your analogy, a new, undamaged, pot.

As I wrote before, I agree with your insightful assessment that some peope prefer the simplicity of gluing together imperfect pieces and thus don't like the newly required precision of the more coherence solution put forward by systemd.

Unlike others they don't need the pot in your analogy to be perfect and would rather remain gluing pieces together as that allows them more flexibility to put random pieces where they prefer them.

It had never occured to me that a major contention point could be relaxed vs. required precision, so your comment was quite a revelation.

Distribution maintainers should have contributed to the individual packages and made them work coherently from the start.

I guess part of that, maybe even the main one, needs to come from the upstream developers.

But I agree with you that systemd will help a lot in that regard, as it allows both upstream and distribution maintainers to have a common focus.
Upstream developers will still often have options for other systems, but the systemd option will fit particularily well with other packages on such systems.

The culprit lies in the rise of the many Linux distributions that made it impossible for free software to be coherent.

I don't think the distributions are at fault here.
One of the strengths of the ecosystem is that even individual developers can start creating software for their own needs and then later adjust it to cover the needs of others.

Support of standards often first comes in that adjustment phase because it wasn't a priority for the "in-house" solution.
Sometimes there are no standards for a certain area and developers of similar products first need to get together, agree on something and then have another adjustment phase to actually implement the agreed upon things.

Distributions can of course help in identifying which products appear to be doing similar things and provide a test bed for standardization efforts.

Cheers,
_

[sdack]

The contradiciton I was seeing was that your comment apparently put those who don't like systemd into the same group who like systemd.
...
I don't think the distributions are at fault here. ...

You don't create unity by dividing people into groups. Freedom is great, but it is nice when we all come together once in a while and agree on the same things, or what we have is an anarchy. The distributions still haven't manage to create a single, unified package format for example, and I doubt they ever will. So I don't see why not to blame the distributions for the things they neglected to do.

[Amarildo]

Well you pointed exactly what systemd brings to the Linux ecosystem : the kernel is now surrounded by a complete and generic set of services and their management, so that distros mainteners can now focus more on other tasks like integration, human interface, debug etc.

Exactly, I don't know why people are always pointing out the obvious, I'm repeating myself without need.
Having the free time to do all those things is one good side of systemd and standardization in general; and forking it is duplicate of effort. I've said this numerous times already.

[duby229]

say no more. I was a hardcore gentoo user (from gentoo's day one). still having few gentoo machines doing specific tasks..

wanting to use logind on non-systemd distro is an edge case, iirc we already agreed to not disagree on this

That's fair enough. But I still don't think it's an edge case though.

[duby229]

Gentoo is usually regarded as the "distro for edge-case enthusiasts", even Arch is.
Most people don't feel like all the stuff Gentoo offers is necessary, or worth it.
Not that this makes Gentoo any better or worse. Just stating that the whole point of Gentoo is to cater edge-case enthusiasts.

Lets just say I couldn't possibly disagree with you any harder than I do. Gentoo is what they like to call a meta-distribution. I can build a fully functional desktop linux with it and push binary updates to the end user and that person would never have a problem with it. That's not an edge, it's the whole point. I do use it that way to a limited extent on some workstation I am obligated to maintain. But I choose to push entire preconfigured filesystem images instead.

[pal666]

I'm sure there could be tons of good forks that don't make systemd such a hard dependency

where are they? i see only whiners

, yet it would be very hard to convince distros to switch in the first place

there was even example of forked distro in this thread. of course nobody will use sucking distros with sucking ports and that is what systemd-less ports will turn into

pressuring the developers to make the software work as we want.

lol, nobody will make shitty software because of your pressure