A philosophical diatribe: How to avoid having users - and how not to

esr · 2018-12-07 01:04:09

My username ain't lying, I am in fact "ESR". If that declaration makes you sit up and pay attention, good - it was supposed to.

Practically, I want to get the hell out of systemd-land before that gigantic crock collapses into rubble. Theoretically, Devuan's premise - holding out for Unix modularity and loose coupling against monolithic monstrosity - appeals to me greatly. I do respect and admire what the Devuan crew has done so far on a purely technical level, something I especially need to make clear because I'm going to use strong language about some things Devuan has *not* done - and needs to.

To set the scene, here's how my house net looks:

I have an experimental install of ASCII running on a fanless compute brick (hereafter "grue") downstairs. Its twin brother, hereafter referred to as "grelber", has been running stable under Ubuntu 16.04 LTS as my mailserver for a couple of years - that's the first machine I'd like to Devuanize. They have a potential triplet, "hurkle", which is just now in pieces on my worktable. All three are a brand called Jetway, if that matters - generic Chinese mini-PCs intended for use as media centers and in light industrial computing.

There are two upstairs machines. The first, my wife's home-office machine, is "minx", an Intel NUC running Mint 19 - I've tried to Devuanize that and failed. The second is snark, aka the Great Beast, my desktop, running Ubuntu 18.10. The network glue is quintaped, a CeroWRT-equipped router which will not make any further appearances in this rant because it Just Works.

In an ideal world, I'd migrate all of these to Devuan. I can't yet.

The near subject of this diatribe is the problems I discovered while attempting to convert my house machines. The far subject is how Devuan can stop shooting itself in the foot. If you want to have more devs, you need to attract a larger userbase. To attract a larger userbase, you have to be friendly - or at least not passively hostile - to new users. Right now, Devuan is passively hostile, not just to non-techies but even to old Unix/Linux hands like me.

Yes, I know you inherited a lot of this passive hostility from Debian; now stop making excuses and face the actual problem. If the fightback against systemd is going to be won (and I'd like to see that) it's not going to be won by outcompeting Debian but by *demonstrating superiority to the mass-market distros*.

The problems start with choosing an install image.

Here'a a clue: if the user has to *choose* an install image, you've lost the plot before he/she even got started. To be fair Mint and others make this mistake too, which is one reason Ubuntu stays stubbornly dominant. But, again, stop making excuses: to beat systemd you have to beat Ubuntu, and that road starts with outcompeting the Ubuntu installer.

No, I'm not talking about choice by hardware - of course you need different installation images for different host processors, that's unavoidable. I'm talking about bullshit, pointless distinctions like "classic" vs. "live" or "desktop" vs. "server" installers.

Sure these are different use cases for the installed system. There is no reason for these cases to have different *installers*, though (large thumbdrives are cheap), and three excellent reasons they shouldn't.

1. User confusion and frustration. A user-friendly experience does not start with "Oh no, which installer image do I pick?" and follow with "Oh, shit, I got to the 43rd prompt and realized I got the wrong installer, now I have to waste time starting from scratch." A user-friendly experience starts with a frictionless install in which the user is guided through a fanout of installation types by questions from the *one* installer.

Anyone who comes back at me with "Oh, no, these choices are easy, you just read this obscure piece of documentation or these buried release notes or this particular web page" is announcing loudly that they have *no clue* what "frictionless" means. If you want to beat back systemd - which means beating Ubuntu and Red Hat - you have to be smarter than this. You have to build your installer as if no documentation outside it exists or ever will.

Anyone who comes back at me with "But, nobody gets this quite right" gets answered with "That's called 'window of opportunity'. Do you want to argue for your limitations and fail, or fight back effectively and win?"

2. Inconsistent behavior. When you have multiple installers you will have genetic drift and inconsistent behavior. I have been personally bitten by this; the "Live" installer offers to install drivers (some of) my Jetways need, but the Classic one does not.

Hello? Hello? Is there intelligent life here? That behavioral divergence just said "We don't care if we frustrate our users; they can deal with our bad choices or fuck off to Ubuntu/Mint/Red Hat."

3. Total defect load. Multiple installers are duplicative. It's simply bad software engineering to repeat yourself - takes more effort, invites more bugs.

So. *One* installer. Which is also the live image - you know, the way Ubuntu does it. There's no reason one installer couldn't cover the whole range of desktop and server installation cases from text-only minimal up. For extra credit, give the ARM live image the capability to flash the SD on an RPi.

Anyone who says "If you think this is such a good idea, do it yourself" is missing my highest-level point. To attract and retain the devs who can do these things the Devuan crew needs to demonstrate that it has a clue about what better user experience looks like and is willing to make the strategic commitment required to get there.

Anyone who says "Well, if you don't accept Devuan's present style and design objectives, what are you doing here?" gets answered with "Accept limiting design objectives as now defined and remain an almost-irrelevant niche distro, or play to win. Choose one." And if you choose to lose, there's certainly no reason for *me* to stick around.

Second major topic: proprietary firmware blobs.

Look, I don't like those any better than the next guy. But you've inherited from Debian a set of practices about them that is counterproductive and insane, and you need to fix it.

Debian practice: Don't install those blobs by default. Require user to explictly opt in or out to proprietary things as a class *before he/she knows whether any are actually needed*. (At least I think this is Debian behavior. I've never done a Debian install; it could be worse.) Make a big ideological deal about being "pure".

"Worse" is what the Devuan Classic installer does: never offers to install them at all. You end up where I was, with higher display resolutions on one of my my Jeways not working and me having no clue why. (Radeon chip; enough said.)

Someone actually said to me on #devuan IRC "Well, you have to read the release notes to find out how to install the proprietary firmware blobs." Which is pretty much the same as saying "We don't care how painful the install is, users who aren't old Linux hands can just get stuffed."

What is anyone involved in this clusterfuck actually *thinking*?

If being a software-purity refusenik about proprietary firmware blobs put any pressure at all on hardware vendors, there might be some shred of justification for this nonsense. Clue: it doesn't. They don't care and will never need to care. Your virtue-signaling is pointless.

The only way you can drive these blobs out of existence is by outcompeting them - doing what they do *better* in open source. Until then your "purity" only harms Linux users without having any positive effect at all.

Since hackers seem to be *particularly* dense about this, I'll make some logic explicit. When I bought the Jetways I wasn't thinking about avoiding hardware that needs proprietary firmware blobs - because it's not practical to do that; the hardware blacklist is too big and complicated to apply when you're shopping on Amazon or Newegg. And I'm an *expert* at this compared to J. Random Newbie.

Often you can't even *get* to the relevant facts about the hardware. One of my Jetways (grue) needs the Radeon blob. The other (grelber) doesn't. If you think the Amazon listings for these puppies gave me any clue either way, you're dreaming.

When I got grue and found out after a long, tortuous series of diagnostics that I need to manually install firmware blobs, it's wasn't the vendor that had screwed me over and wasted my time; it's the distro that could have installed them automatically, no muss, no fuss - but didn't.

The right way to handle this would be with install-time dialogues that say things like: "Your graphics chip is a a Radeon card that requires a proprietary firmware blob to run at resolutions above SVGA. Install?" And then default that to "Yes", because your job is to *enable* your users.

If that's too complicated, install all blobs rather than risking that a missing one will degrade performance. Because your job is to enable your users, not make an ideological point about which the vendors will never, *ever* care.

This is the person who discovered and wrote down the economics of open- vs. closed-source business models telling you to stop being stupid about it.

Third topic: stale kernel.

Two weeks ago, when I upgraded my wife to an Intel NUC from an aging tower system running Ubuntu that had developed flakiness in its graphics card, I wanted to upgrade her to Devuan. Couldn't do it. I had to fall back to Mint 19, and my window is closed now. If I destabilize minx when I don't absolutely need to Cathy gets upset - and with good reason, these migrations are *always* pains in the ass.

The problem turned out to be that the Devuan ASCII kernel is too old. The list of vendor-ID/subtype pairs for the e1000 didn't include the subtype the NIC on the NUC has.

The e1000 is not weird hardware, and the NUC series is pretty popular. This is the universe's way of telling Devuan its release frequency needs to increase.

I think this is a minor problem compared to the others I've pointed out, but worth a mention.

What I'm willing to do:

At the moment, I can't justify putting more effort into improving Devuan myself. That's because none of my production machines are running it yet. When that changes (and I want it to) budgeting some of my time to pitch in could happen.

There is no way I'd even consider destabilizing the Great Beast before I have successful experience with multiple Devuan installs.

I am much further away from the point where I can shout from the rooftops: "Hey everybody! Devuan is a good thing and you should totally switch!" Because right now, considered as a polished, user-friendly distribution - a competitor to Ubuntu or Mint - Devuan still sucks pretty badly.

That can change. But the existing Devuan community has to *want* it to change, first. In this diatribe I've tried to make the case why you should.

I can't make that choice for you. What you decide about this issue of direction will, more than anything else, determine whether I stick around.

Last edited by esr (2018-12-07 18:12:59)

golinux · 2018-12-07 09:22:34

We are doing our thing and quite happy doing it.

I can't remember anyone ever whining about lack of "user friendliness" or expressing a desire to vanquish the dominance of Ubuntu/Mint. In fact, if the VUAs had decided to fork Mint, I would not have spent the last four years committed to Devuan.

Different strokes. Devuan may not be to your liking. If you feel changes to Devuan are needed, offer them for consideration. Not just talk but tangible alternatives. Whether you stick around or not is a decision you will have to make for yourself.

MiyoLinux · 2018-12-07 10:31:51

The following...in no way...represents the views of Devuan nor the administrators of this forum.

I'm responding, because I'm a bit perplexed. I'm not sure what to think...is this a complaint or an actual offer of advice?

I'll be the first to admit...and others here will agree...that I don't get out much. I don't have a clue as to the importance of the username "esr". Sorry. What is the significance?

Extended Security Release?
Endoplasmic Subdural Reticulum?
Eat Some Rice?

If the poster has never installed Debian (as he/she acknowledged), then that may explain his/her problem with the installer? That may also explain that the poster may not be familiar with Devuan's background?

I'm not trying to stir up conflict, but I truly don't understand why a user would come here and threaten not to hang around unless their "suggestions" were met. Does esr have that much pull and influence? (that's a serious question)

In closing, I will leave you with this thought...

cynwulf · 2018-12-07 10:33:58

esr wrote:

Second major topic: proprietary firmware blobs.
Look, I don't like those any better than the next guy. But you've inherited from Debian a set of practices about them that is counterproductive and insane, and you need to fix it.

I'm glad you brought this up. I have been banging on about this for years and only come up against blinkered ideologues - essentially people who don't want the firmware because Debian (or GNU) said it's "non-free". Debian only introduced this silly policy in the Debian 6 squeeze release (2011'ish?), before that you got the firmware as standard and your hardware worked, for the most part.

esr wrote:

Debian practice: Don't install those blobs by default. Require user to explictly opt in or out to proprietary things as a class *before he/she knows whether any are actually needed*. (At least I think this is Debian behavior. I've never done a Debian install; it could be worse.) Make a big ideological deal about being "pure".

I understood recently, from another thread here, that Devuan may just install all the firmware - and one user was not particularly happy about that.

Debian treats the binary only firmware images as non DFSG compliant software - the same way it treats Linux binary non free packages (i.e. the "we'll just leave this non-free repository here and pretend it's not part of Debian... even though we have tons of the developers busily packaging non-free stuff so as to be indistinguishable from free stuff" silliness).

The big difference of course is that the firmware images are not Linux binaries and are never touched if you don't own the hardware. If you do own the hardware, then presumably you want to use that hardware?! That hardware is probably already loaded with proprietary firmware in the shape of the BIOS/(U)EFI, video BIOS, CPU microcode, etc.

ralph.ronnquist · 2018-12-07 11:14:29

@MiyoLinux: I gather esr's tax identity is Eric S. Raymond, and his CV, would he present one, would or at least could be jam packed with notes about contributions to open source software.

Of course I would agree that the style of "barging in" and presenting an authoritarian opinion is ill suited to this community. But it's certainly worth to hear him out. I think his points are more directed to end-user distribution aims, than to the core Devuan project of maintaining a systemd free platform. Even so, they can well be discussed on this thread.

ianbruene · 2018-12-07 16:55:26

Unlike ESR I am a relative noob. I haven't used anything but Ubuntu for more than a few hours, and I haven't even tried to install Devuan yet. I have just been watching from the sidelines as ESR and others poke at Devuan and other systemd removal methods. I don't have the experience to say systemd is bad on my own knowledge, but I do want something to replace Ubuntu that has less hidden magic and yet works with things such as the latest Mesa drivers. I am part of the market that would adopt Devuan shortly before mass appeal, and help to push it over the hump.

The right way to handle this would be with install-time dialogues that say things like: "Your graphics chip is a a Radeon card that requires a proprietary firmware blob to run at resolutions above SVGA. Install?" And then default that to "Yes", because your job is to *enable* your users.

Better: include a "What does this mean?" button. And then actually put a useful explanation there.

To the people asking variants of "Why should we care about dethroning Ubuntu?" here is why: you are the back-pressure in the ecosystem. If systemd is as bad as people say - and if not why does Devuan exist? - then we can define a failure condition for linux as systemd being an absolute requirement for anything to work. Conversely we can define a success condition as systemd being unneeded, fundamentally changing its nature, or everything that wants to use systemd being happy to use some other system.

If everyone is happy to include and run systemd then the back-pressure against it will consist of people writing angry comments on the internet. There are few activities that are less effective. This will result in reaching the systemd failure condition.

If some people try to avoid systemd but have a culture and systems that don't care about or actively oppose wide adoption then the rest of linux will ignore them as "those weirdos over there", and their back-pressure will only be marginally more effective then the angry comments. This will result in reaching the systemd failure condition.

If Devuan out-competes Ubuntu then it acquires a large user base, and by extension a large developer base. With a large base it cannot be dismissed as "those weirdos", and developers risk their code being unusable across a sizable chunk of the linux world if they insist on having a required systemd dependency. Bonus if you get the corporate users on the train. Supporters of systemd simultaneously lose one of their largest sources of being the "inevitable path" that linux must travel. This can result in the non-systemd win condition.

If you want to actually fight systemd you cannot stay in the ghetto. If you want to stay in the ghetto then you will be reduced to fighting a Glorious Last Stand followed by slipping into oblivion. Yes, a Glorious Last Stand can be awe inspiring and the best way to fail; but it is still failure.

Perhaps Devuan is not the mass market anti-systemd distro. But if it isn't then we need one fast.

EDIT: spelling

Last edited by ianbruene (2018-12-07 17:03:10)

sgage · 2018-12-07 17:20:14

ianbruene,

Even if it were the goal, Devuan will never out-compete Ubuntu, because $$$ and marketing if nothing else.

The goal of the Devuan project, as far as I can tell, is to keep the choice open of using the init system of your preference - it's not 'conquering', and 'winning', and 'out-competing'. It may not be the mass-market anti-systemd distro - not the goal. In fact, a stated goal is that others take the Devuan infrastructure and build other distros based on it. You know, the way Ubuntu is based on Debian. Maybe one of those will end up being the 'mass market anti-systemd distro'.

But all the apocalyptic verbiage about Glorious Last Stands, and closely defining your success and failure criteria (your criteria, btw), etc. betrays, in my opinion, a lack of understanding of what the Devuan people are trying to accomplish. They are doing their best with limited resources, growing slowly and carefully (as far as I can tell - I am not in any kind of inner circle). Your comment, along with ESR's original posting, are casting this as a battle royale, good vs. evil, etc. Meanwhile, Devuan devs are doing their jobs, talking with Debian people, and otherwise pursuing their goals, the main one of which is choice, and not allowing systemd to just foreclose all other options.

ianbruene · 2018-12-07 19:02:34

Several parts of your response are correct, or could be argued back and forth. But those would be a distraction from this which I cannot understand:

sgage wrote:

(your criteria, btw)

My criteria is "systemd is bad". Everything I have ever seen or heard about Devuan is that its purpose for existence is based on the idea that "systemd is bad". The Devuan front page starts its sales pitch with "Devuan GNU+Linux is a fork of Debian without systemd.". The statement of intention to fork is about the failure of Debian to avoid systemd.

If Devuan isn't about preventing the complete takeover of systemd (even if just in a small corner of the linux world) then it is the greatest snowjob since the invention of politics.

If it is about avoiding systemd then how could the scenario of systemd being completely unavoidable be anything but a failure? How could the scenario of systemd being mothballed and replaced by something better be anything but a win? These are not "my criteria"; they are the necessary boundary conditions based on the stated goals of the project.

KatolaZ · 2018-12-07 19:53:56

Hi,

I should start by saying that in any technical discussion I am never impressed at all by names, or nicknames, or CVs, or titles. I just look at facts and actions.

I thank esr for his post, but I really don't see much point into it. The main mission of Devuan is to provide a Debian without systemd, not to dominate the world or to outcompete Ubuntu or any other distribution. The mission of Devuan is to provide choice.

Now, there are people that have worked and are working to transform the Devuan's mission into reality, which translates to providing packages, releases, infrastructure, documentation, and services. These people work on Devuan on a voluntary basis, and it is not a mistery that the number of active Devuan developers is far smaller than the number of paid Ubuntu developers (which apparenty amounts to at least several dozens). As a result, those Devuan developers have had to focus on one main objective: make Devuan exist, against all odds and despite the well-informed have predicted its death many times.

On the more technical side:

-the install images that are offered are only those that are actually used by Devuan users;

- the comment about firmware is a bit out of scope as well: Devuan is already installing automatically all the firmware (free or non-free) needed by the underlying hardware. This does not include, so far, proprietary video devices. If there are people with an interest in that, they are welcome to join in, propose patches, write workarounds.

I agree that usability is important, and in fact the effort to provide a usable session-management that did not depend on systemd was the main reason why Devuan ASCII release was delayed by several months. A lot of Devuaners helped to let it become a reality, by installing, testing, reporting problems, and the merit of that achievement is only theirs. But Devuan does not have the energy to outcompete Ubuntu on usability. This might put off entire classes of users, but we can't do much about it. We just don't have the energy to do that. We need to focus the human power we have on ensuring that Devuan keeps existing and providing choice.

In general, my opinion is that writing rants pointing towards what does not work in other's people efforts is relatively easy, and telling others what they should do and what they should not do is quite cheap. Anybody can do that. Literally, anybody. However, if four years ago the VUAs had just written a rant on a Debian forum, pointing to all the odds and mishaps of adopting systemd, and promising that they would have contributed to Debian if and only if Debian had removed systemd, well, if they had just stopped after having written a post on a Debian forum then Devuan would most probably not exist at all today, and this thread (and this very forum) would not exist as well.

Devuan came to existence because a bunch of people that everybody labelled as "mad" actually sat in front of a keyboard and hacked a way around the systemd madness, while the rest of the community was still flaming on forums and mailing lists. This was made possible by dozens times many more "mad" people who believed that the work of those hackers was worth the effort, and put energy, encouragement, support, and money to help them in their quest. This effort was most probably inspired by the words of the wise Guru in "The Loginataka" http://www.catb.org/~esr//faqs/loginataka.html:

Hear, O Nobly Born: The center of the mystery is the act of coding. You have a keyboard before you; pursue the Way through work.

These words are often (but maybe imprecisely, since the wise Guru was much wiser than any other sage or prophet) translated with the more prosaic:

Be the change you would like to see in the world.

Devuan is an example that it is possible to live up to that standard. Literally anybody is welcome to join the quest, if they are ready to live up to that standard as well. Devuan does not need phylosophers, or prophets to show "The right way forward". The predicament of the wise Guru is enough: Devuan needs hackers to get the job done. There is plenty of room at the bottom.

HND

KatolaZ

sgage · 2018-12-07 20:18:20

ianbruene wrote:

Several parts of your response are correct, or could be argued back and forth. But those would be a distraction from this which I cannot understand:
sgage wrote:
(your criteria, btw)
My criteria is "systemd is bad". Everything I have ever seen or heard about Devuan is that its purpose for existence is based on the idea that "systemd is bad". The Devuan front page starts its sales pitch with "Devuan GNU+Linux is a fork of Debian without systemd.". The statement of intention to fork is about the failure of Debian to avoid systemd.
If Devuan isn't about preventing the complete takeover of systemd (even if just in a small corner of the linux world) then it is the greatest snowjob since the invention of politics.
If it is about avoiding systemd then how could the scenario of systemd being completely unavoidable be anything but a failure? How could the scenario of systemd being mothballed and replaced by something better be anything but a win? These are not "my criteria"; they are the necessary boundary conditions based on the stated goals of the project.

You simply can't seem to express yourself without hyperbole and totalizing language.

Devuan is of course about preventing the complete takeover of systemd, in the sense of preserving the possibility of non-systemd distros. By 'your criteria' I was referring among other things to your predicating that on Devuan out-competing Ubuntu. And you go off about 'greatest snowjob' yadda yadda.

If systemd were to go away, of course that would be a win. Is it the only win? I don't consider it a 'necessary boundary conditions based on the stated goals of the project.' That is your criterion. That's what I mean.

golinux · 2018-12-07 20:37:08

The subtitle of this thread - "How to avoid having users - and how not to" - begs the question - "what users?". IOW . . . users have different preferences and needs. It's unlikely that one size will fit all. Choice and diversity in the Linux eco-system is a good thing and Devuan offers a choice. There are users who will like and support what we are doing and others who won't. We are here for those who resonate with what we offer and invite them to add their talents to improve what we are doing either directly or by creating a unique derivative.

Devuan is not just about avoiding systemd but also about providing a variety of init options. In fact, Devuan and Debian devs are now working together to keep init diversity alive in Debian. Collaboration like that is a good thing.

Devuan never has been (and is unlikely to ever be) about self-promotion or marketing or out-competing other distributions. We do what we do and offer it as a choice to the Linux community. It will not be everyone's choice. But it will become what its users contribute to its improvement and growth. Cliche alert . . . actions speak louder than words.

chris2be8 · 2018-12-07 21:12:04

ESR is one of the better known figures in the UNIX world. You can read most of his writings at http://www.catb.org/~esr/ (several are well worth reading).

He made several good points about how to make Devuan easier to install. In particular having one and only one installer per architecture that then asks what sort of install you want to do and if you want to install non-free code. If you must keep old versions available then clearly label them as deprecated (and don't put them on the home page).

He could have been a bit more tactful but that's no reason to reject everything he said.

A magazine cover disk may need to be more specialized since space is limited (they may want to have more than 1 distro on a single DVD) and it should be usable to install on a system that does not have internet access. But the people building the cover disk should be more experienced than the average user.

Chris

golinux · 2018-12-07 21:22:35

chris2be8 wrote:

He could have been a bit more tactful but that's no reason to reject everything he said.

I don't think we rejected everything he said. I think we invited him (or someone with the itch to so) to provide the functionality that he envisions.

ChuangTzu · 2018-12-07 22:05:19

If he is the real ESR, and his writing style appears to confirm this, then his thoughts/observations/suggestions should be considered. Even though I do not agree with all of them.

I do think Devuan has the potential to catapult past the systemd distros, but is this what the dev's want? If not, should they want this? Was this ever considered? How big of a project is Devuan prepared to become, etc...? Has ESR reached out to Devuan via IRC or other methods?

PS: I am usually suspect of offers to help only if you do this first though.....

golinux · 2018-12-07 22:25:25

ChuangTzu wrote:

I do think Devuan has the potential to catapult past the systemd distros, but is this what the dev's want? If not, should they want this? Was this ever considered? How big of a project is Devuan prepared to become, etc...?

Thinking of the future is a waste of time and energy. Wanting also. We are quite practical and do what needs to be done NOW. The future will manifest as it will depending on where contributors take it. If someone wants to write an installer as esr describes, they are welcome to do so. We at Devuan are collaborators and will work with anyone who cares to provide additional functionality/features.

Has ESR reached out to Devuan via IRC or other methods?

Yes.

KatolaZ · 2018-12-07 23:05:21

ChuangTzu wrote:

If he is the real ESR, and his writing style appears to confirm this, then his thoughts/observations/suggestions should be considered. Even though I do not agree with all of them.
I do think Devuan has the potential to catapult past the systemd distros, but is this what the dev's want? If not, should they want this? Was this ever considered? How big of a project is Devuan prepared to become, etc...? Has ESR reached out to Devuan via IRC or other methods?
PS: I am usually suspect of offers to help only if you do this first though.....

I respect ESR as I respect many other people in the free software community (and most of them are unknown to the masses). Anyway, good ideas ain't any better or more effective if they are not turned into actions, irrespective of how they originated. At the moment the current group of people who is working to deliver Devuan does not have time to put into the massive improvement of usability that is envisaged in some of the posts of this thread. This does not mean that Devuan has given up on becoming the most used distro in the world :-) This just means that we have the obligation to ensure that Devuan survives, first, and then that it improves. If there is no Devuan, there is no Devuan to improve.

How many of you would be ready to delay the release of Beowulf by two or three years to improve usability, and to get something that in the end -let's be honest- will not be able to compete with Ubuntu or Mint of Whatnottix in terms of usability and point-and-clicky stuff, nevertheless? Yeah, many would say "me! me!", but then we have lots of people complaining on IRC everyday about the current stable becoming obsolete, and their preferred software not being updated to the latest version. We have lost thousands of users due to the delay in releasing Jessie and ASCII.

Anything that did not existed before was created. Anything that does not exist yet must be created. And creation requires energy, in this universe. Devuan has sufficient energy to survive and thrive for the foreseeable future. Adding more stuff will require more energy, that must come out of somewhere/somebody. I am a positive person, and I am sure that this energy will become available in some form, because Devuan offers something unique to many, many users.

HND

KatolaZ

Last edited by KatolaZ (2018-12-07 23:06:38)

ChuangTzu · 2018-12-08 00:47:08

KatolaZ wrote:

I respect ESR as I respect many other people in the free software community (and most of them are unknown to the masses). Anyway, good ideas ain't any better or more effective if they are not turned into actions, irrespective of how they originated. At the moment the current group of people who is working to deliver Devuan does not have time to put into the massive improvement of usability that is envisaged in some of the posts of this thread. This does not mean that Devuan has given up on becoming the most used distro in the world :-) This just means that we have the obligation to ensure that Devuan survives, first, and then that it improves. If there is no Devuan, there is no Devuan to improve.
How many of you would be ready to delay the release of Beowulf by two or three years to improve usability, and to get something that in the end -let's be honest- will not be able to compete with Ubuntu or Mint of Whatnottix in terms of usability and point-and-clicky stuff, nevertheless? Yeah, many would say "me! me!", but then we have lots of people complaining on IRC everyday about the current stable becoming obsolete, and their preferred software not being updated to the latest version. We have lost thousands of users due to the delay in releasing Jessie and ASCII.
Anything that did not existed before was created. Anything that does not exist yet must be created. And creation requires energy, in this universe. Devuan has sufficient energy to survive and thrive for the foreseeable future. Adding more stuff will require more energy, that must come out of somewhere/somebody. I am a positive person, and I am sure that this energy will become available in some form, because Devuan offers something unique to many, many users.
HND
KatolaZ

I don't have a bone in this fight so to speak, I favour the less user friendly. I was brought up on RTFM-man pages etc..., Unix and early days of Linux still had this, some distros still hold onto it, Slackware being one, I would like to see Devuan remain this way as well. Provide the tools, provide the programs, but let the user learn and grow. If something is too user friendly the user learns nothing other then how to click "report issue", whether its *buntu, Microsoft or Apple, you become merely a user of the technology with no understanding of whats going on. Just my 2 cents.

ChuangTzu · 2018-12-08 00:48:11

golinux wrote:

Thinking of the future is a waste of time and energy. Wanting also. We are quite practical and do what needs to be done NOW. The future will manifest as it will depending on where contributors take it. If someone wants to write an installer as esr describes, they are welcome to do so. We at Devuan are collaborators and will work with anyone who cares to provide additional functionality/features.
Has ESR reached out to Devuan via IRC or other methods?
Yes.

Panopticon · 2018-12-08 03:36:51

I just have to point this out.

esr said

"1. User confusion and frustration. A user-friendly experience does not start with "Oh no, which installer image do I pick?" and follow with "Oh, shit, I got to the 43rd prompt and realized I got the wrong installer, now I have to waste time starting from scratch." A user-friendly experience starts with a frictionless install in which the user is guided through a fanout of installation types by questions from the *one* installer."

then says...

"Debian practice: Don't install those blobs by default. Require user to explictly opt in or out to proprietary things as a class *before he/she knows whether any are actually needed*. (At least I think this is Debian behavior. I've never done a Debian install; it could be worse.) Make a big ideological deal about being "pure"."

So if you have never done a Debian install, your frustration it seems is not really based on any facts is it, or am i missing something here?

Last edited by Panopticon (2018-12-08 03:37:19)

Ron · 2018-12-08 17:14:46

Panopticon wrote:

So if you have never done a Debian install, your frustration it seems is not really based on any facts is it, or am i missing something here?

He has tried a Devuan install:

esr wrote:

The first, my wife's home-office machine, is "minx", an Intel NUC running Mint 19 - I've tried to Devuanize that and failed.

golinux · 2018-12-08 17:37:58

Ron wrote:

Panopticon wrote:
So if you have never done a Debian install, your frustration it seems is not really based on any facts is it, or am i missing something here?
He has tried a Devuan install:

Not with the Debian installer, it seems.

esr wrote:

The first, my wife's home-office machine, is "minx", an Intel NUC running Mint 19 - I've tried to Devuanize that and failed.

Mint is NOT Debian/Devuan. Not at all surprising that trying to Devuanize Mint failed.

Ron · 2018-12-08 18:39:11

I took what he wrote to mean that he tried to install Devuan over top of Mint. Any other way wouldn't be possible, right? It seems he is knowledgeable enough to know that.

Last edited by Ron (2018-12-08 18:41:14)

golinux · 2018-12-08 19:37:43

Ron wrote:

I took what he wrote to mean that he tried to install Devuan over top of Mint. Any other way wouldn't be possible, right? It seems he is knowledgeable enough to know that.

He said, "The first, my wife's home-office machine, is "minx", an Intel NUC running Mint 19 - I've tried to Devuanize that and failed." I have no idea what that means because it contains zero detail about what he tried.

Even Debian advises not to use Mint/Ubuntu Packages. This from the Debian Wiki

Don't make a FrankenDebian
Repositories that can create a FrankenDebian if used with Debian Stable:
- Debian testing release (currently buster)
- Debian unstable release (also known as sid)
- Ubuntu, Mint or other derivative repositories are not compatible with Debian!
- Ubuntu PPAs

So how is the failure Devuan's fault?

Then came the rant about how confusing Devuan installers are. Well, how are Devuan isos so different from the selection that Debian offers except that the RPi stuff is at a different location? That indicates a (rather shocking) lack of familiarity with Debian.

You might remember these two threads (in which you participated) on this very forum relating to Mint (there may be more):

Migrating from Mint (includes some discussion of (or whining about) "user friendliness")
https://dev1galaxy.org/viewtopic.php?id=2522

Linux Mint Devuan Edition?
https://dev1galaxy.org/viewtopic.php?id=2317

esr · 2018-12-08 23:14:55

I tried a Devuan install from scratch on the NUC, failed due to the NIC problem, then did a Mint install from scratch.

Last edited by esr (2018-12-08 23:16:29)

esr · 2018-12-08 23:31:34

Ron wrote:

I took what he wrote to mean that he tried to install Devuan over top of Mint. Any other way wouldn't be possible, right? It seems he is knowledgeable enough to know that.

No, this take is incorrect. Attempted a clean install of Devuan using the Classic image, failed, then started over with a full Mint install.

Last edited by esr (2018-12-08 23:31:56)

The officially official Devuan Forum!

#1 2018-12-07 01:04:09

A philosophical diatribe: How to avoid having users - and how not to

#2 2018-12-07 09:22:34

Re: A philosophical diatribe: How to avoid having users - and how not to

#3 2018-12-07 10:31:51

Re: A philosophical diatribe: How to avoid having users - and how not to

#4 2018-12-07 10:33:58

Re: A philosophical diatribe: How to avoid having users - and how not to

#5 2018-12-07 11:14:29

Re: A philosophical diatribe: How to avoid having users - and how not to

#6 2018-12-07 16:55:26

Re: A philosophical diatribe: How to avoid having users - and how not to

#7 2018-12-07 17:20:14

Re: A philosophical diatribe: How to avoid having users - and how not to

#8 2018-12-07 19:02:34

Re: A philosophical diatribe: How to avoid having users - and how not to

#9 2018-12-07 19:53:56

Re: A philosophical diatribe: How to avoid having users - and how not to

#10 2018-12-07 20:18:20

Re: A philosophical diatribe: How to avoid having users - and how not to

#11 2018-12-07 20:37:08

Re: A philosophical diatribe: How to avoid having users - and how not to

#12 2018-12-07 21:12:04

Re: A philosophical diatribe: How to avoid having users - and how not to

#13 2018-12-07 21:22:35

Re: A philosophical diatribe: How to avoid having users - and how not to

#14 2018-12-07 22:05:19

Re: A philosophical diatribe: How to avoid having users - and how not to

#15 2018-12-07 22:25:25

Re: A philosophical diatribe: How to avoid having users - and how not to

#16 2018-12-07 23:05:21

Re: A philosophical diatribe: How to avoid having users - and how not to

#17 2018-12-08 00:47:08

Re: A philosophical diatribe: How to avoid having users - and how not to

#18 2018-12-08 00:48:11

Re: A philosophical diatribe: How to avoid having users - and how not to

#19 2018-12-08 03:36:51

Re: A philosophical diatribe: How to avoid having users - and how not to

#20 2018-12-08 17:14:46

Re: A philosophical diatribe: How to avoid having users - and how not to

#21 2018-12-08 17:37:58

Re: A philosophical diatribe: How to avoid having users - and how not to

#22 2018-12-08 18:39:11

Re: A philosophical diatribe: How to avoid having users - and how not to

#23 2018-12-08 19:37:43

Re: A philosophical diatribe: How to avoid having users - and how not to

#24 2018-12-08 23:14:55

Re: A philosophical diatribe: How to avoid having users - and how not to

#25 2018-12-08 23:31:34

Re: A philosophical diatribe: How to avoid having users - and how not to

Board footer