Combatting revisionist history

[n_hologram]

The English teacher in me would have loved to see this editorial laden with references for further reading, but obviously I can do the Googling myself. 10/10 post.

Up next: "Deconstructing systemd: a collection of analyses" by j_derrida_420

[golinux]

FYI, you are getting some well-deserved love on the Devuan mail list.

https://lists.dyne.org/lurker/message/2 ... 93.en.html

[emariz]

The vast majority of Debian's developers are actually packagers. I am not minimising their excellent work, but stating a fact that has been ignored in this conversation.

While we could get into a long discussion about Red Hat's motives behind the development and implementation of systemd, nowadays, giant sets of tools, like Gnome, depend on systemd, and packagers must decide whether to use these tools and adapt them into their projects, or reject them and create alternatives to supply that functionality.

I am confident that Debian's developers could create brilliant alternatives to all these tools afftected by systemd, but I am not sure that they want to invest their time and effort doing so. They are already very efficient, and very busy, at putting together thousands of pieces from thousands of sources into one successful project, Debian.

I do not know who will pay for the implementation of systemd in the years to come, nor do I know who was going to pay for the effort of Debian's developers to provide us with an alternative.

[dasein]

The English teacher in me would have loved to see this editorial laden with references for further reading..."

And the psychologist in me says, "'Incidental' learning is a good thing."

That is to say: I actively encourage folks to check my facts, because it is in the act of inquiry that true learning happens. But there is so much background knowledge involved that there is simply no way to anticipate all possible combinations of prerequisites. Watch:

Do we start with the notion that it's almost always demonstrably stupid (and way more expensive) to do a wholesale replacement of existing working code with new code? (Particularly when we're talking about devs who act as if "We're crossing our fingers and hoping this never happens" is an effective substitute for Defensive Programming best practices.)

Or do we start with maintainability? This one is gonna be killer, particularly given Poettering's known history of dumping half-finished code into userland and leaving it to everyone else to sort out and fix. Pop quiz: Initial development is/is not more than half of any system's true cost.

Or let's talk code complexity, but let's do it like we're in the NFL. Frame the discussion in a way that eliminates folks who think that the word data is the plural of the word opinion; if someone doesn't know a Halstead metric from a hole in the ground, then they don't get to play.

Or perhaps we start with a short course in software engineering economics? Because that's the bottom of bottom lines; it doesn't farqing matter if systemd has an orgasmatron feature if it never pays back the cost for its implementation. I want NASA spending my tax money exploring space, not rewriting system maintenance procedures because of systemd.

Indeed, if you watch the historical "debate" carefully, every time anyone raises this point in one way or another, the systemd fanbois' response seems always to consist of equal parts hand-waving and name-calling. (Something like, "Because... because better, you Luddite pedophile!") There is no facepalm adequate to the task.

(Edit: NSFW wording removed)

[dasein]

FYI, you are getting some well-deserved love on the Devuan mail list.

I suspected something along those lines when I saw that it got something like 300 views in just a few hours. But thanks for sharing the love.

Speaking of love (and hate), I take notice of the fact that this account has been de-purple-ized since my OP, which pretty much eliminates the need for a "non-staff" account. "Haters gonna hate" and all that, but at least now they can ignore me if they choose. (Besides, this way I don't have to transfer all ~100 names on my existing ignore list to the other account.)

I still won't be around much, because once Jessie becomes Stable, a huge part of what I can contribute technically (not to mention large portions of what I have already contributed) becomes obsolete. And the devs' cowardly response to Ian Jackson's GR makes Debian Just Another Distro as far as I am concerned. I mean, Debian has now publicly declared that the decision to adopt a platform whose explicit goals include making it easier to insert binary blobs into a system is unworthy of even a moment's somber reflection. At that point, DFSG might as well stand for Diabetic Foot Study Group. (My sincere thanks to saulgoode/alansmithee for taking the time to educate me on this point.)

(Though I loved the one ML comment by the guy who said I was "biased," as if I hadn't admitted (even highlighted) my bias from the very start, or as if being biased in favor of empirical answers to empirical questions is something to be ashamed of. )

(Apologies for all the edits. There was one sentence I just couldn't quite get right. And I'm still not happy with it, damnit.)

[mardybear]

Thanks for posting. Good read.

dasein wrote:
As noted above, the only reason “systemd is a bad idea” wasn't the outcome of the GR vote is because the DPL offered a “political cover” amendment that allowed “This conversation is superfluous” to be the correct answer. Which it isn't.

Certainly a disheartening outcome. Thought i would be using Debian forever...the OS search continues.

“The Only Thing That Is Constant Is Change -”. ― Heraclitus

[dasein]

One last thing and then I'll shut up, I promise. (I think.)

Systemd advocates love to twist words, notable among those being the word monolithic. By their logic, the giant rocks at Stonehenge don't qualify as monoliths (literally: single stone), because after all, each one is made up of zillions of constituent atoms.

Not only is a farkload of binaries masquerading as a single process "monolithic," it's the friggin' dictionary definition of monolithic!

Ok. Sorry. It's just that that's been bugging me for a while. I feel better now. Sorry.

(Edit: typo)

[golinux]

“The Only Thing That Is Constant Is Change -”. ― Heraclitus

Heraclitus and also the Buddha

Sabbe sankhara anicca

[friartek]

--- cut ---

--- end cut ---
The GR was not a mandate for systemd
I have no idea where Debian's systemd fanbois get the idea that a victory lap is appropriate.

The results of the GR vote were diluted and obfuscated by two non-resolution outcomes. Of the three technologically-relevant resolutions to the GR, one was unequivocally pro-systemd, the other two were contra-systemd, differing primarily in phrasing (essentially the difference between “must not” and “should not”).

(Aside: I'm a “must not” guy at heart, but I grudgingly admit that those who contend that an absolute prohibition might prove unnecessarily inflexible or self-limiting might have a valid point. Maybe. But I don't have to like doing it. )

But here's the thing, and there is just no getting around it. Once you eliminate the ass-covering amendment offered by Nussbaum, “systemd is a bad idea, the only real question is how bad” didn't place third.

It placed first. By a substantial margin.

Conversely, “systemd is a good idea” didn't place first.

It placed last.

If this surprises you, even a little, then by all means, go look it up. And for the love of whatever you hold Sacred, refrain from uttering a single word of GR-related drivel until you do..

Speaking of the GR...


The GR was too necessary
As noted above, the only reason “systemd is a bad idea” wasn't the outcome of the GR vote is because the DPL offered a “political cover” amendment that allowed “This conversation is superfluous” to be the correct answer. Which it isn't.

And here I thought they were voting on the question of what init software would be the default.
Actually when I read what they were voting on it made me sick. Talk about a loaded question. I hate politics.

--- cut ---

--- end cut ---
Speaking of costs...

Changes cost, and big changes cost big
There's a reason why sysadmins in large organizations are routinely among systemd's biggest detractors.

Downtime is expensive in terms of both time and money. So is re-training. So is rewriting gigabytes of artificially-obsoleted documentation. Add them all up, multiply by hundreds of thousands/millions of installs, factor in the associated opportunity costs, and before you know it, the costs to the global economy associated with systemd deployment reach into the billions (or thousands of millions, if you prefer) of dollars/euros.

I'm sure Redhat would like a big piece of the pie.

--- cut ---

--- end cut ---
So long, and thanks for all the fish...

If I could I would go with them. I guess in my own way, by moving to another distro, I am.

Thanks for you insights.

[dasein]

And here I thought they were voting on the question of what init software would be the default.

No, that decision was made long ago, by a similarly sharply divided small committee.

(See, this is why it's important to do one's own research. )

I'm sure Redhat would like a big piece of the pie.

It is unquestionably true that RedHat now wields a disproportionate influence over the future of Linux (Lennax?). As I noted in the OP, Debian was the only distro whose sheer size and wide influence would have allowed it to stop (or at least significantly slow) systemd deployment; but they chose not to. Whether RH's motives are nefarious is ultimately a matter of opinion, and thus outside the realm of empirical inquiry. But RH is now in exactly the same position that Microsoft was when they started to deploy "BackOffice," which is what ultimately spelled doom for a lot of amazingly good software. (I still miss SmartSuite.)

Thanks for you insights.

Thanks, and thanks to everyone else who's expressed appreciation. I'm delighted that folks are finding it useful. I very nearly didn't post it.

[RU55EL]

One last thing and then I'll shut up, I promise. (I think.) [...]

I would like to encourage you to post when you feel like it. I find your posts very informative, and I appreciate your opinions.

[1of12]

The vast majority of Debian's developers are actually packagers. I am not minimising their excellent work, but stating a fact that has been ignored in this conversation.

True but lets not let that get in the way of a good ol' unfounded conspiracy theory.

While we could get into a long discussion about Red Hat's motives behind the development and implementation of systemd, nowadays, giant sets of tools, like Gnome, depend on systemd, and packagers must decide whether to use these tools and adapt them into their projects, or reject them and create alternatives to supply that functionality.

I am confident that Debian's developers could create brilliant alternatives to all these tools afftected by systemd, but I am not sure that they want to invest their time and effort doing so. They are already very efficient, and very busy, at putting together thousands of pieces from thousands of sources into one successful project, Debian.

Pragmatism will win through - if systemd allows the big desktop developers to streamline things they'll do it as they have done before.

The fuss about systemd is overblown. For years the average user didn't know or care what init system was used and didn't care about hal, dbus, policykit, consolekit, PAM and similar stuff creeping onto their systems. Most desktop users have used or tolerated parts of gnome and KDE, which made big use of these interfaces, for years, and there was barely a complaint. The big desktops have pretty much been reinventing windowsisms, using binary configs, binary logs, confusing xml config, mono and now we have a vocal minority whining about "UNIX philosophy".

Ubuntu, the biggest distribution, implemented what is in fact the predecessor and one of the main inspirations for systemd - upstart - back in 2006, no one cared, no one complained as they weren't using Ubuntu. Now that systemd is default in Debian, certain users are up in arms... they are talking about the end of Linux, Red Hat taking over and so on.

If you're worried about the "UNIX philosophy", worry about what else is on your system already. Switch to a *BSD perhaps, as that's a lot closer to UNIX than GNU/Linux these days.

Hysterical nonsense.

[mardybear]

dasein wrote:
One last thing and then I'll shut up, I promise. (I think.)

Thanks, and thanks to everyone else who's expressed appreciation. I'm delighted that folks are finding it useful. I very nearly didn't post it.

Glad you did...you've probably been waiting a long time to get this out!

Linus Torvalds apparently uses systemd, which could be viewed as the final nail. Don't recall reading an official opinion from Richard Stallman, but based on this it is doubtful he would use it:
https://stallman.org/stallman-computing.html

Edit:

1of12 wrote:
Hysterical nonsense.

Don't hold back, how do you really feel about it

[dasein]

you've probably been waiting a long time to get this out!

Only since the GR vote.

That is, BTW, why @mor deserves partial credit for any informativeness that may have accidentally found its way into these posts. After the GR vote, I was fuming, due in equal measure to the cowardice of vote results themselves, some obvious early indicators of revisionist history, and what I perceived to be the troubling lack of community-centric moderation during the meltdown. Mor encouraged me to wait until the cacophony of delusional drivel posted by technically clueless morons died down before offering up this particular two cents' worth. (Needless to say, he was much more polite in his phrasing. These are my words, not his.)

Sure enough, he was, as he so often is, exactly correct.

(Sorry for the edits, I kept fumbling for a word and finally found it.)

[Head_on_a_Stick]

Thank you for the very informative post -- I was ignorant of the political details until I read this.

I hope you stick around, your contributions to this forum are extremely valuable.