This is what Stallman had to say about systemd. Really?!?

[somebodyelse]

I'd be interested in Jacob Applebaum's views on systemd. He mentions it in a specific instance close to the beginning of the following talk but I don't really understand his point or whether he's in favour or against https://www.youtube.com/watch?v=b0w36GAyZIA

[WeLoveDebian]

In other words, he may be a good guardian of software freedom in terms of the parameters he has set. But he is no longer a young man and is probably not the best person to fully comprehend how the freedom of the users of computers/devices can be compromised in the age of accelerating changes we are living in.

Therefore, discredit him as his ideas are obsolete...

It is an unfounded assumption the claim that being old also implies one's ideas are necessarily obsolete. It is also a taken for granted, the conviction that old people cannot understand the world around them, and cannot therefore, properly respond to its newer necessities.

Younger people necessarily have less experience which makes them more prone to repeat the mistakes of their predecessors. Therefore, the optimum, if it exists, doesn't belong to being young or old, but it rather belongs to a position, where one is determined to learn from the past mistakes of others, while at the same time, evaluates the situation to change strategy accordingly.

Both mental aptitude and experience are necessary, other than that, it would be a path full of failures.

You, my dear, just nailed it. Bravo!

As we age, our minds become less open to new things and less flexible. This is known and scientifically established. I meant no disrespect to him. Only that he is very good at defending the problem in terms of the parameters he has set. He may be less able to reimagine those parameters as the issue changes over time.

But we still have the freedoms he set. We can look at the code, run it as we wish, modify it and distribute our modified copies. This is essentially still being able to produce a new system that doesn't use systemd.

Our freedoms aren't at risk as you make they seem. Problem is, we (as a community) can't make smart decisions for the vast majority times when there's a vote, and this is IMO what happened with systemd's addoption. Is anyone willing to create another init system (and whatever elese systemd does) just to avoid systemd? Just as systemd's developers did? Because this is how things go on Linux: "I don't like this one little thing so I'll create an entire new big thing". This is effort wasting, and this is part of why we haven't taken off as a desktop sytstem, because there are thousands of choices, dozens of package formats and package managers, dozens of sound systems, display managers, etc. Thousands of people repackaging things that should work on a standards: one and improved/fast pacakge format and package manager (say, pacman); one audio system; one Xorg or Wayland or Mir, not the three; one (and the best) init system; one module controler; etc. As long as we ALWAYS create new things because we don't like the existing ones (instead of improving them), we're never going to stop having "systemds" all around us. We still have the freedom to create new things, nobody needs to stick with systemd. But are we going to improve it or create a whole new "systemd" because we don't like what's out there?

Probably. Again, I really don't think Stallman cares about a software project so long as its free with an LGPL license attached. As much as the systemd haters and conspiracy theorists like to think that Red Hat/Poettering have some ideas to subvert GNU/Linux, I find it hard to find any truth to it so long as the code is open source and can be freely modified by anyone who cares to do that. What you should worry about is what seems to be the blind trust that major distros have in it though. When all is said and done, systemd is a software project, just like any other LGPL software project, but Debian shoving beta quality software on Stable users is quite another issue entirely, and that is where I take issue.

I agree with you almost completely.

The LGPG license, as per gnu.org, should be used only in a few cases.

https://www.gnu.org/licenses/license-li ... reLicenses

GNU Lesser General Public License (LGPL) version 3 (#LGPL) (#LGPLv3)

This is the latest version of the LGPL: a free software license, but not a strong copyleft license, because it permits linking with nonfree modules. It is compatible with GPLv3. We recommend it for special circumstances only.

Stallman is all about freedom, yet not aware of some of the freedoms being taken away by using systemd.

Like which ones?

I smell both contradiction & BS in Stallman's recent statements.

Please explain these thoughts

[Randicus]

Is anyone willing to create another init system (and whatever elese systemd does) just to avoid systemd?

First; systemd is not an initialisation system. Initialisation is one of many functions that make up the systemd package.
Second; Why create something to replace systemd when systemd is replacing things that worked perfectly well? It makes no sense to replace a replacement that is unneeded. That logic does not make sense. Those who want to avoid systemd can use OSes that do not use it. Avoiding systemd by creating something that does the same thing?

[WeLoveDebian]

Why create something to replace systemd when systemd is replacing things that worked perfectly well?

That actually is the logic behind my question and that post

[Randicus]

Really? Your logic is so twisted, even I cannot follow it.

[WeLoveDebian]

hehehehe.

Well, sysvinit worked, right? Despite this, systemd's developers thought they should create something new. Nowadays a ton of people dislike systemd, so what's the choice for them? Either stay with it, or create a new thing. We DO have the freedom to create a new product that is systemd-like, even with systemd on the scene. We DO have that freedom, and some pointed out we don't, that's the core of my post. But is that really the best option out there, to simply create a new thing?

This is what actually happens on Linux: "I don't like this, so I'll create yet another almost-useless thing that fits what I think should be the gold-standard out there, and that works 98% the same way as the previous thing I dislike. And when people dislike my new gold-thing? I don't change it, because I'm egocentric as sh1t, a developer who only cares about his own issues. And then people will create another thing, even though mine works just too". <<- Linux all over.

There's no need to create a new systemd-like thing, as there was no need to create systemd in the first place. The same with Mir and Wayland, Cinnamon, .deb packages, Ubuntu, pulseaudio, 99.99% of the distros out there, etc. We, as a community, should embrace systemd and improve it, making pressure so that systemd's developers make systemd work as the community wants it to.

Even though we have the freedom to create a whole new init system (and whatever else systemd does) that is not always the best thing to do. We need standards that work, that makes things easier to developers.

Imagine the thousands of people who work on the openSUSE development team, simply re-packaging packages. They take a package, put it out of it's container, than re-package the same thing, only this time it probably needs a few modifications so this package functions on openSUSE's ecosystem. Imagine if all these developers didn't need to do all this, and instead of doing work that is useless they could actually improve what's out there, like improving Xorg, or creating new GPL drivers (thus making the Kernel actually GPL, without blob firmwares), creating games, new programs, etc.

The end user shouldn't care if the packages of his/her system are a .deb or a .rpm format. And IMO the coordination teams of all major distros out there should unite and agree on making actual standards: one package format, one GOOD audio system, one good display manager, one good and working set of drivers, etc. It should benefit everyone.

[dilberts_left_nut]

We need standards that work, that makes things easier to developers.

This was one of the major driving forces behind the development and adoption of systemd.

[WeLoveDebian]

We need standards that work, that makes things easier to developers.

This was one of the major driving forces behind the development and adoption of systemd.

Exactly.

[somebodyelse]

At the end of the day, with systemd Red Hat has a solved a problem it - and presumably similar companies like Novell/Attachmate/SUSE - had. The community is perhaps too dependent on what comes from upstream. The problem is not systemd as such but the inability for those who don't want it to sidestep it. So what if it swallows x, y and z? Just fork them.

There is no point in mentioning that the LGPL is for certain circumstances unless mentioning what those circumstances are and considering whether it is legitimate that systemd is licensed under it. It could be dodging the GPL. It could also just be using the LGPL as it is meant to be used.

Anyway, the ship has sailed. Systemd was introduced years ago. Since then everyone has left Arch Linux, openSUSE and now Debian and is currently using BSD

Given its use in surveillance and data collection/analysis of all kinds, I am starting to wonder if Linux is the great force for freedom I thought it was. Still, I suppose if it hadn't existed we'd all be being watched and weighed up with UNIX, BSD, Windows etc. or ... BeOS.

Imagine the dozens of people who work on the openSUSE development team

There, fixed that for you.

[Head_on_a_Stick]

We need standards that work, that makes things easier to developers.

[WeLoveDebian]

There is no point in mentioning that the LGPL is for certain circumstances unless mentioning what those circumstances are and considering whether it is legitimate that systemd is licensed under it. It could be dodging the GPL. It could also just be using the LGPL as it is meant to be used.

Ummmm.... the points are in the link I provided.

Anyway, the ship has sailed. Systemd was introduced years ago. Since then everyone has left Arch Linux, openSUSE and now Debian and is currently using BSD

But what would be the point of that? heheheh. The fear that systemd will introduce closed-source binaries? openBSD does that already with it's firmware blobs - Debian don't. Well, Chromium pulled such binary blob that even spied on the users, and the Debian team didn't move it to Contrib or Non-free. Who knows.

Given its use in surveillance and data collection/analysis of all kinds, I am starting to wonder if Linux is the great force for freedom I thought it was.

It still is. There are very good distros like Parabola, which is based off of Arch, is completely GNU and respects the user freedom. I love Arch and I literally feel at home with Parabola, but since my VGA is kind of new my monitor flickers (off and then on again), I'm working to fix that. I'm still deciding on weather to stay on Debian or Parabola.

Still, I suppose if it hadn't existed we'd all be being watched and weighed up with UNIX, BSD, Windows etc. or ... BeOS.

Either that or Linux would be closed source

Imagine the dozens of people who work on the openSUSE development team

There, fixed that for you.

Really? Speaking of overall developers, the people who work on the stable releases, the people who work on Factory and Tumbleweed... there are less than 100 of them?

[WeLoveDebian]

We need standards that work, that makes things easier to developers.

Although I do like some of xkcd's comics, most of them are wrong when it comes to technology. I mean, since when 14 competing can be called "standards"? hehehe. 14 different package formats aren't 14 standards

Not only this one, but that one about the pass-phrases is wrong too. Waaay too wrong.

I guess the person who creates these tech comics is a software developer

[tomazzi]

I mean, since when 14 competing can be called "standards"? hehehe. 14 different package formats aren't 14 standards

Actually this is a bit overdrawn, but not that far from reality: f.e. there were (or still are - I just don't care) 2 official standards for XML.

But for sure, there are still 2 official standards for measuring capacity of digital memory: One of them defines KiB as 1024 bytes and the second defines KB as 1000 bytes. The official explanation on this matter is, that "kilo" means 1000, so 1KB == 1000 is correct.

The problem is, that LITERALLY NOONE uses KB unit of size 1000 bytes. Since this unit is specific and usable only in context of digital world, it makes completely no sense to force such standard - but it exists. Unfortunately, Linux distros in their blind efforts to stick to the standards at all costs, have adopted this crap.

And this is the single moment in my whole life when I'm going to praise Microsoft - they didn't.

Why?
Everyone, literally everyone of engineers working with any kind of digital circuits just don't give a crap about the existence of KiB unit - seriously.
You wont' find a single technical datasheet which will reference the KiB - because this is stupid, impractical unit, standarised to cheat the customers - or in other words, when You're rich enough then You can just force the IEEE to accept any kind of bullshits.

Proof: Storage devices manufacturers are ubiquitously using MB as 1000000 bytes when it comes to storage capacity, explaining that this is the "standard" unit. But at the same time they're using KB as 1024 bytes when it comes to cache size

------------------
Systemd: Dasein said once, that there are plenty of "talking heads", which are expressing their pseudo-expert opinions about systemd, while in fact they haven't written a single line of code in their lives - and I must admit that I'm fully agree with His opinion.

Regards.

[WeLoveDebian]

I mean, since when 14 competing can be called "standards"? hehehe. 14 different package formats aren't 14 standards

Actually this is a bit overdrawn, but not that far from reality: f.e. there were (or still are - I just don't care) 2 official standards for XML.

But for sure, there are still 2 official standards for measuring capacity of digital memory: One of them defines KiB as 1024 bytes and the second defines KB as 1000 bytes. The official explanation on this matter is, that "kilo" means 1000, so 1KB == 1000 is correct.

The problem is, that LITERALLY NOONE uses KB unit of size 1000 bytes. Since this unit is specific and usable only in context of digital world, it makes completely no sense to force such standard - but it exists. Unfortunately, Linux distros in their blind efforts to stick to the standards at all costs, have adopted this crap.

And this is the single moment in my whole life when I'm going to praise Microsoft - they didn't.

Why?
Everyone, literally everyone of engineers working with any kind of digital circuits just don't give a crap about the existence of KiB unit - seriously.
You wont' find a single technical datasheet which will reference the KiB - because this is stupid, impractical unit, standarised to cheat the customers - or in other words, when You're rich enough then You can just force the IEEE to accept any kind of bullshits.

Proof: Storage devices manufacturers are ubiquitously using MB as 1000000 bytes when it comes to storage capacity, explaining that this is the "standard" unit. But at the same time they're using KB as 1024 bytes when it comes to cache size

------------------
Systemd: Dasein said once, that there are plenty of "talking heads", which are expressing their pseudo-expert opinions about systemd, while in fact they haven't written a single line of code in their lives - and I must admit that I'm fully agree with His opinion.

Regards.

That bugs the hell out of me with Debian. The good thing is that I create my partitions with fdisk, and fdisk (in all distros I could find) does use 1024 bytes pattern. So if I create a "1G" partition I know I'll get a 1024 MB partition.

A quick tip: To convert harddrive units into actual GiB, just do:

Code: Select all

X-GB x 931.532

Where "X" is the labled size of the HD, according to the manufacturer. Works everytime

40 GB = 40x931.532 = 37.2 GiB actual

[Randicus]

The fear that systemd will introduce closed-source binaries? openBSD does that already with it's firmware blobs - Debian don't.

You seam a bit misinformed. Binary blobs are not allowed in the OpenBSD kernel or base system. Security is one of the main foci of the system's makers, so hidden code is considered a security risk. Lack of binary blobs is a main cause of some hardware not working with the system. Perhaps you are confusing OpenBSD and FreeBSD? (I do not know if FreeBSD has blobs, but I do know they are more lenient to such things.)

[WeLoveDebian]

The fear that systemd will introduce closed-source binaries? openBSD does that already with it's firmware blobs - Debian don't.

You seam a bit misinformed. Binary blobs are not allowed in the OpenBSD kernel or base system. Security is one of the main foci of the system's makers, so hidden code is considered a security risk. Lack of binary blobs is a main cause of some hardware not working with the system. Perhaps you are confusing OpenBSD and FreeBSD? (I do not know if FreeBSD has blobs, but I do know they are more lenient to such things.)

I'm not misinformed. There's an interview with Theo in Youtube (IIRC) that he says "the firmware is merely copied to the device so it's not executed by the CPU".

openBSD's policies state that no binary DRIVER should be present in the Kernel. However, non-free FIRMWARE *IS* present.
I have an AMD R9 270X that is bugged as hell in pure GNU systems (Because these don't have the AMD firmware blob). In Debian/Parabola my monitor goes only at 1280x1024 @ 60 Hz, even though it can reach 1280x1024 @75Hz or 1600x1200 @68 Hz. This happens because there is no closed-source firmware loaded.
In openBSD, Arch Linux, Ubuntu, etc, my monitor functions as expected because these vanilla kernels (Linux included) contains formware blobs. In fact, this makes me think why Linux is still licensed under GPL when it has firmware blobs.

[shevegen]

RMS is now an old man. He no longer knows what is going on.

What else did you expect?

He never used systemd, he does not know it.

[golinux]

RMS is now an old man. He no longer knows what is going on.

That is an absolutely ridiculous statement made by a someone without much life experience. Guess the younger generation never learns . . .

[millpond]

Ridiculous?
Hardly.

RMS is around my age - and I havent a clue as to what the hell is going on these days.

With all the pinheads putzing up perfectly working system architectures.

[golinux]

Ridiculous?
Hardly.

RMS is around my age - and I havent a clue as to what the hell is going on these days.

That is a sad admission . . . (from someone who has a decade on RMS)