An interview with Ken VanDine, Ubuntu desktop lead at Canonical

To be fair, I've had continued problems with PulseAudio that were entirely resolved by purging it from the system and relying on ALSA more directly, or using apulse as PA proxy. These were not ALSA driver problems, these were "no sound, no possibility to get sound, pulse was seemingly doing whatever the hell it wanted regardless of configuration or attempts to restart the service". My favourite was ignoring the mono sink and outputting to speakers that didn't exist.

While your problems very much could be related to pulse (I do not want to convey the idea here that it's the most bug free software out there [somewhat due to the complex nature of the problem it is trying to solve which pipewire have demonstrated since some people for whom pulse worked perfectly well pipewire doesn't at all]) even problems like this could be due to errors on ALSA; the whole architecture in pulse depends upon the rewind functionality of ALSA, something that no other software uses and it turned out that a lot of drivers therefore also had massive problems with it (aka completely untested feature in the drivers). Granted one can blame pulse for using that feature when it was so badly supported.

And true enough that internal QA cannot hope to cover all the myriad configurations and setups out there, there was still a level of rolling out that could have been handled better. A response to buggy software by saying what is akin to "deal with it" will never sit well with me.

If that indeed was the answer then yes that is both a horrible and idiotic answer. Just like you I would be without customers if I had that attitude (basically also why we gain customers from competitors, but I digress).

I am, however, rather concerned about the code complexity and just how susceptible it is to security flaws, especially given the level of control over the rest of the system.

And I would argue that the old way with lines upon lines of bash scripts was a way more complex and error prone world. But yes there will always be a tradeoff between functionality and more chance of security problems due to more lines of and more complex interactions of code. And there are no objective truth here, myself I subjectively like the features more than I'm afraid over the added complexity, others are free to do a different tradeoff analysis and that is of course fine.

Sure, security flaws are elsewhere too, but I'm one who suscribes to the old unix philosophy of keeping everything as simple as possible, do one thing and that one thing right, and systemd does away with that. I'm also against the kind of spread of scope - there's software that now relies on systemd (init) that ideally shouldn't have to know about the init system at all.

Does those softwares really rely on the systemd as init or are they relying on some D-BUS interface provided by various applications in the systemd project? In any case this is of course due to the project providing functionality and services that other software writes find useful, I get that it sucks for people not wanting to run systemd or not wanting to write software providing similar functionality using the same D-BUS namespace but at the same time should be forbid developers from providing great features?

I've also personally suffered systems that refused to boot because systemd wasn't recovering properly. Root cause was not systemd, but it was clobbered far too easily and locked up everything else. On non-init, I've had problems with logging, I don't even know what was going on with timesync and ended disabling the entire damn thing until a full OS reinstall was done, and a bunch of other problems that probably bias me against the whole project.

I can see how such a situation can be frustrating enough to be mad at the entire project, I've had similar experience with various pieces of software (hello macOS) that made me hope that they would freeze in hel. The irony of it all though is that if something, systemd brought excellent logging to Linux with the journal (I'm not a friend with how it handles tab-completion though) so unfortunate that we don't have a time machine so we could get back to your machine with the systemd-timesyncd so figure out what it was, perhaps sitting behind the machine the both of us would have showed you exactly why I like systemd and how I see the journal as a life save. Or the result would have been the same, who knows when it comes to software :)

In any case I asked for your reasons and I thank you for providing them. I fully understand that you don't want to make this into a debate and I see no reason for it to be that either. As I said I was simply interested in your experience since you sounds like a sane person (you don't want to know how many insane anti-systemd people I have talked to on Slashdot or Phoronix over the years) and I'm glad that I was right :).

And learn to use your system, because probably (about 99%) you are the problem, not your machine, nor the software.

Oh no no no, never say that. Take any device (a computer, a door, a hose, whatever), and if most users have trouble using it, that's definitely a design flaw of the device.

Here, see this: https://m.youtube.com/watch?v=yY96hTb8WgI

At work we've got, on a single walk through a single floor, doors that
* are just opened manually,
* must be opened manually, but only after you show your entrance card to some device,
* open automatically after you show your entrance card to some device, and must not be moved by the handle they've got, because otherwise, most probably, world explodes.

Please, if you constructed a door so stupidly that it must not be moved by its handle, AT LEAST DON'T GIVE IT A FU**ING HANDLE!

Probably designed by the same geniuses that designs the self scanning check out stations that have a touch screen and then either a IR reader for a card (drivers license in this case) or VISA/Mastercard touchless reader and if you accidentally scan your card before touching the screen it locks with a "you must touch the screen to begin before you scan your card". Ok so why is the damn scanner even active before I touch the screen if this is so important and secondly why is it so important?

Not to mention how completely AI-less they are, here I am using a machine that requires me to have a visa card (which the banks over here only hands out if you are 18+) and have the shop member card (which again they only hand out if you are 18+) and I have bought stuff for say €200 but I no I also have bought a few energy drinks (which the store on it's own have decided to have a 15 age limit on) so now it have to blink red until some underpaid clerk 15 minutes later can come to my aid and logon, click "age verified" so I can continue.

Because the modus operandi of some pre-teen wanting to buy an energy drink (which is not illegal. again the age limit is set by the store) is to steal some one else visa card, their store membership card and then also spend €200 on other stuff...

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it.
Also you don't understand basic package management and yet somehow you qualified to say that snaps are wrong. HOW would you know exactly? Also there are plenty of google finds if you really want to know how to remove it. You know..instead of speak of "forced" snaps.


Last edited by Lycurgus87 on 25 May 2022 at 6:43 am UTC

Probably designed by the same geniuses that designs the self scanning check out stations that have a touch screen and then either a IR reader for a card (drivers license in this case) or VISA/Mastercard touchless reader and if you accidentally scan your card before touching the screen it locks with a "you must touch the screen to begin before you scan your card". Ok so why is the damn scanner even active before I touch the screen if this is so important and secondly why is it so important?

Right, we've got the same with our attendance recorder! (Yeah, for some weird reason, this is a thing where I'm developing software...) If you try to use the touch screen with the hand you're holding the card, card is registered first and device is confused...

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it.

Also you don't understand basic package management and yet somehow you qualified to say that snaps are wrong. HOW would you know exactly? Also there are plenty of google finds if you really want to know how to remove it. You know..instead of speak of "forced" snaps.

If I get a car, I want it to drive, decently, and without me needing to know how to tune the motor. And I'd still feel qualified to say e.g. that the acceleration is not as it should be.

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it. ...

I used to think like you, opposing simplifications to computer software (or as I would have called it, stupidification) when I was young, because "you need to know exactly how computers work, otherwise you lose your smartness". Some bs like that.

I was wrong, and hope you can take this chance to learn this well.

Just because you have to use it, doesn't mean you need to understand it. You in particular may be smart and curious and capable and willing to understand, which is a great thing. Most people however are either incapable or unwilling to learn about everything.

And please understand that this sort of talk also drives people away (don't get defensive please, and appreciate the feedback), so be mindful of the effect of your statements on people who are showing an interest in Linux.

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it.
Also you don't understand basic package management and yet somehow you qualified to say that snaps are wrong. HOW would you know exactly? Also there are plenty of google finds if you really want to know how to remove it. You know..instead of speak of "forced" snaps.

OK, so first, go back through the thread and figure out that I didn't actually say any of what you're now saying I said, and that those things were in fact said by other people. Then if you've got a lot of moxie you'll come back and admit your mistake.

And, yes, you need to learn to USE it. So for instance, to use a washing machine I need to know the difference between delicates and heavy load, cold water and hot on the dials and buttons. If I'm good I'll also know something about the impact those settings have on different clothes. I don't actually need to know how agitators work, where the fan belts are in the guts and so on, in order to either use it or conclude there's something wrong, my clothes aren't getting washed, or acquire an opinion about whether side-load or top-load are better, or conclude that washing machines last far less long than they used to and realize that this is actually a planned thing so they can force me to buy a replacement.

Similarly, in order to use a computer I need to know how to use a mouse and generally navigate a WIMP interface, ideally something about how files and folders work, have a feel for how software tends to lay out its menus and commands, that kind of thing. As a bonus it's nice to know some things about how computers and operating systems and so on work at an abstract level. There are fuzzy edges, but this is largely distinct from what I need to know to be a proper computer enthusiast, or programmer, or sysadmin (In point of fact I happen to know a couple of things beyond what I've itemized, but those are not things I would need to know to use a computer). But this is all I need in order to, say, conclude that my system gets buggered up when I install certain software. It's more than I need in order to evaluate political arguments about computers. And so on and so forth, as per the washing machine analogy. Your claims about what opinions people should be allowed to have are false and ignorant. Nobody even makes these arguments about most other machinery and technology we use and interact with every day, it's almost entirely a computer snob thing (sometimes a car snob thing, and in both cases it usually has male chauvinist undertones), although I've never before met anyone going so far as to say that if I don't want to be a computer enthusiast I've lost my qualification to be human ("nearest tree").

What will be next? The cars?? oh wait thats already happening, and its the same problem because humans getting so entitled and stupid they don't even bother to learn how to drive properly anymore..and companies adjust cars for ppl who not willing to learn to drive.

Cars might be a good comparison: Yes, you should know how to "drive" your computer. But you shouldn't need to be able to fix its "motor". Yes, that's of course a nice thing to know, but most won't, and that's all fine.

b) I wondered if it was just Poettering-hate.

That was my impression as well.

And while I can't say much about this person, in general I'm of the opinion that social competence does matter, not only technical one, as open source should be about working together.

While social competence does matter, having a certain quantity of out-of-the-box / potentially slightly authoritarian (within reason) persons can be useful. Those personalities tend to be much better at going against the stream, and push more ambitious projects (such as pulseaudio). It's all a matter of balance though, the person can not be a pure asshole either, but being socially awkward, very (too) sure of themselves and slightly authoritarian can actually be useful in some scenario. You also can not only have persons like that, just a few, with persons used to them as buffers usually. At least that's from my experience.


Last edited by 3zekiel on 26 May 2022 at 3:11 pm UTC

Nice interview, but I really dislike how Canonical is forcing Snap. Also, it's really sad that Ubuntu is one of the few distros that have really bad Flatpak support with outdated Flatpak/XDG portal packages, no Flatpak support in their GNOME Software fork etc.

So not only are they forcing snaps on their users, but also gimping flatpak? Wow....

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it. ...

I used to think like you, opposing simplifications to computer software (or as I would have called it, stupidification) when I was young, because "you need to know exactly how computers work, otherwise you lose your smartness". Some bs like that.

I was wrong, and hope you can take this chance to learn this well.

Just because you have to use it, doesn't mean you need to understand it. You in particular may be smart and curious and capable and willing to understand, which is a great thing. Most people however are either incapable or unwilling to learn about everything.

And please understand that this sort of talk also drives people away (don't get defensive please, and appreciate the feedback), so be mindful of the effect of your statements on people who are showing an interest in Linux.

BUT (a big capitalized but), people should at least have a knowledge of how a tool they use works, right? Like you want to give some preliminary rules on how to use a soldering iron "Don't touch these parts when it's heated, or you will get nicely burned" or firearms, "don't point a gun at a person ever, even if you're sure it's unloaded." etc. Sure a computer generally won't burn you or cost someone their lives, but it never hurts and only helps to learn something about a tool before you use it.

Now while some people concentrate on the wrong things 'You need to know how to compile your own kernel!' isn't the depth we're talking about, but 'this is how you launch your Sudoku game' level of learning, and so many people don't even know that. I think this is the level some need to know about. And in this day and age, there is some actual harm that can befall people who are too ignorant to do simple things like 'run updates', or 'don't click on shady things that we warn you about!' as there can be financial, reputational, etc damage.

So yeah, for sure people should learn something about a tool before trying to use it.

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it. ...

I used to think like you, opposing simplifications to computer software (or as I would have called it, stupidification) when I was young, because "you need to know exactly how computers work, otherwise you lose your smartness". Some bs like that.

I was wrong, and hope you can take this chance to learn this well.

Just because you have to use it, doesn't mean you need to understand it. You in particular may be smart and curious and capable and willing to understand, which is a great thing. Most people however are either incapable or unwilling to learn about everything.

And please understand that this sort of talk also drives people away (don't get defensive please, and appreciate the feedback), so be mindful of the effect of your statements on people who are showing an interest in Linux.

BUT (a big capitalized but), people should at least have a knowledge of how a tool they use works, right? Like you want to give some preliminary rules on how to use a soldering iron "Don't touch these parts when it's heated, or you will get nicely burned" or firearms, "don't point a gun at a person ever, even if you're sure it's unloaded." etc. Sure a computer generally won't burn you or cost someone their lives, but it never hurts and only helps to learn something about a tool before you use it.

Now while some people concentrate on the wrong things 'You need to know how to compile your own kernel!' isn't the depth we're talking about, but 'this is how you launch your Sudoku game' level of learning, and so many people don't even know that. I think this is the level some need to know about. And in this day and age, there is some actual harm that can befall people who are too ignorant to do simple things like 'run updates', or 'don't click on shady things that we warn you about!' as there can be financial, reputational, etc damage.

So yeah, for sure people should learn something about a tool before trying to use it.

Misunderstanding here. The original comment that instigated this discussion claimed that users should know how to disable snaps, switch repos and some such things.

Definitely not something I would ever teach to or expect my parents to have to learn for instance.

Mate you are tiny bit condescending
(snip)
Or the other options again..the tree over there.

You're condescending, says the guy who told me to climb back to the nearest tree.
But I wasn't. I am a tiny bit now: The point went over your head. The point was, there are lots of branches of knowledge that can make claims that if you don't have them you are inferior. It is rare for the claims to be valid; claims about the arts or political economy or DIY home maintenance are no more/less valid than your claim that if I don't want to learn computers I should return to the apes. Do you get it now?

Yes there are lot of branch of knowledge and you don't need to know all but...and this is a big but, usually you don't need to know all that because you don't want to use it.

BUT if you want to use something you have to learn it. ...

I used to think like you, opposing simplifications to computer software (or as I would have called it, stupidification) when I was young, because "you need to know exactly how computers work, otherwise you lose your smartness". Some bs like that.

I was wrong, and hope you can take this chance to learn this well.

Just because you have to use it, doesn't mean you need to understand it. You in particular may be smart and curious and capable and willing to understand, which is a great thing. Most people however are either incapable or unwilling to learn about everything.

And please understand that this sort of talk also drives people away (don't get defensive please, and appreciate the feedback), so be mindful of the effect of your statements on people who are showing an interest in Linux.

BUT (a big capitalized but), people should at least have a knowledge of how a tool they use works, right? Like you want to give some preliminary rules on how to use a soldering iron "Don't touch these parts when it's heated, or you will get nicely burned" or firearms, "don't point a gun at a person ever, even if you're sure it's unloaded." etc. Sure a computer generally won't burn you or cost someone their lives, but it never hurts and only helps to learn something about a tool before you use it.

Now while some people concentrate on the wrong things 'You need to know how to compile your own kernel!' isn't the depth we're talking about, but 'this is how you launch your Sudoku game' level of learning, and so many people don't even know that. I think this is the level some need to know about. And in this day and age, there is some actual harm that can befall people who are too ignorant to do simple things like 'run updates', or 'don't click on shady things that we warn you about!' as there can be financial, reputational, etc damage.

So yeah, for sure people should learn something about a tool before trying to use it.

Misunderstanding here. The original comment that instigated this discussion claimed that users should know how to disable snaps, switch repos and some such things.

Definitely not something I would ever teach to or expect my parents to have to learn for instance.

Ha, disable snaps... or 'install Mint/pop_os that uses the same base, but removes snaps for you...' I know what I would choose.

So I had more complaints about Ubuntu's use of Snaps then I ever had of Mir.

Ubuntu put some early work into Wayland if I remember correctly before Wayland was on anyones radar, before they decided that Mir was more suited to their purposes.

Would have been nice to see more support for Mir on the desktop from the open source community when Mir and Wayland looked like they were competing against similar goals. Good things have come out of a bit of competition, even in the open source community. Just look at gnome vs kde.

As far as Ubuntu being the source of the NIH philosophy, wasn't SystemD a bit of NIH from the Fedora / Redhat team, and now it is incorporated throughout most of Linux?

So I had more complaints about Ubuntu's use of Snaps then I ever had of Mir.

Ubuntu put some early work into Wayland if I remember correctly before Wayland was on anyones radar, before they decided that Mir was more suited to their purposes.

Would have been nice to see more support for Mir on the desktop from the open source community when Mir and Wayland looked like they were competing against similar goals. Good things have come out of a bit of competition, even in the open source community. Just look at gnome vs kde.

As far as Ubuntu being the source of the NIH philosophy, wasn't SystemD a bit of NIH from the Fedora / Redhat team, and now it is incorporated throughout most of Linux?

Depends on how you look at it... upstart was a thing, but wasn't something that seemed either good enough, or widely tested enough for all the other distributions to adopt. Plus it literally only covered init, whereas systemd covers much more. Systemd for example has some watchdog like elements to it.
Canonical only has a small team of developers, IBM/Redhat have tons more, and if there is a core project out there for Linux stuff, it seems like redhat's hands are in it at some level.
I knid of think of the story of the rabbit and the turtle. Canonical is the rabbit in this, trying to hurry up and get to the end without all the planning and development, so they can get something out quicker, but is a mess underneath. Where Redhat is slow and methodical and ends up winning in the end.

I knid of think of the story of the rabbit and the turtle. Canonical is the rabbit in this, trying to hurry up and get to the end without all the planning and development, so they can get something out quicker, but is a mess underneath. Where Redhat is slow and methodical and ends up winning in the end.

Wow, more casual Ubuntu-bashing, eh? Upstart replaced the SystemV init system in 2006 and was used by various distributions (including Redhat, ffs) for about 7 or 8 years. It was instrumental in pushing the start up times of Linux from the 30sec+ norm in those days to the sub-10s we have now. Iteratively, it was surpassed by SystemD and that's fine. That's good, that's what makes Linux better, over time. Upstart wasn't "a mess underneath", nor was Mir, nor Unity, nor Snap, nor many of the other truly exceptional, innovative things Canonical have contributed over the decades they've been around.

Occasionally they did indeed put out duds. That's also fine. As long as we're all running Linux, we're all in this together.

Although from the comments on this thread, you'd hardly fucking think it.

I knid of think of the story of the rabbit and the turtle. Canonical is the rabbit in this, trying to hurry up and get to the end without all the planning and development, so they can get something out quicker, but is a mess underneath. Where Redhat is slow and methodical and ends up winning in the end.

Wow, more casual Ubuntu-bashing, eh? Upstart replaced the SystemV init system in 2006 and was used by various distributions (including Redhat, ffs) for about 7 or 8 years. It was instrumental in pushing the start up times of Linux from the 30sec+ norm in those days to the sub-10s we have now. Iteratively, it was surpassed by SystemD and that's fine. That's good, that's what makes Linux better, over time. Upstart wasn't "a mess underneath", nor was Mir, nor Unity, nor Snap, nor many of the other truly exceptional, innovative things Canonical have contributed over the decades they've been around.

Occasionally they did indeed put out duds. That's also fine. As long as we're all running Linux, we're all in this together.

Although from the comments on this thread, you'd hardly fucking think it.

The problem with Ubuntu is they think of people as Ubuntu users, not as Linux users. Upstart was used briefly in Fedora, It was never used in RHEL. Snap is a mess. Unity was... meh.

Snap is a mess.

Oh, snap!

I knid of think of the story of the rabbit and the turtle. Canonical is the rabbit in this, trying to hurry up and get to the end without all the planning and development, so they can get something out quicker, but is a mess underneath. Where Redhat is slow and methodical and ends up winning in the end.

Wow, more casual Ubuntu-bashing, eh? Upstart replaced the SystemV init system in 2006 and was used by various distributions (including Redhat, ffs) for about 7 or 8 years. It was instrumental in pushing the start up times of Linux from the 30sec+ norm in those days to the sub-10s we have now. Iteratively, it was surpassed by SystemD and that's fine. That's good, that's what makes Linux better, over time. Upstart wasn't "a mess underneath", nor was Mir, nor Unity, nor Snap, nor many of the other truly exceptional, innovative things Canonical have contributed over the decades they've been around.

Occasionally they did indeed put out duds. That's also fine. As long as we're all running Linux, we're all in this together.

Although from the comments on this thread, you'd hardly fucking think it.

The problem with Ubuntu is they think of people as Ubuntu users, not as Linux users. Upstart was used briefly in Fedora, It was never used in RHEL. Snap is a mess. Unity was... meh.

Briefly? Like 5 releases briefly? I've never thought of five years as a brief time, I have to admit. It was also a beta option in RHEL for a similar time, but you're right, it was never adopted, since start-up time on a server O/S is pretty insignificant compared to the POST checks in most Enterprise environments. I vaguely remember Debian considering it too, for quite a while, but I can't remember their reasons for not adopting it. Back then, their focus was and still to a large is, servers.

But anyway, my point is that Upstart was very well thought out - a fully asynchronous service management/init system with a simple interface, massively extendable, yet completely backwards compatible with SysV.

Snap is a mess, is it? Based on what?

Unity was meh? Based on... ah, okay. Your opinion, I suppose, which is fair enough. Opinions differ. Mine was/is that Unity is a great example of Canonical's innovation - reacting to the options in front of them at the time and, crucially, doing something about it. Back then (around 2010), Gnome Shell - now that was a mess. A wholly forced, non-configurable change in workflow. Canonical had to react to that, and Unity was born. It was a brilliant half-way house between gnome 2 and shell. It was, itself, also a bit shitty until the 2012 releases, but that's iterative development for you.

Also back then, remember, Canonical were pushing the idea of a converged mobile & desktop experience. The "Edge" phone was coming, powered by Unity. Incredible vision, when you realise this was twelve years ago. I remember being blown away by the possibility of using my phone as a dashboard screen on my desktop, and resizing, mobile-aware apps - something we take a little more for granted these days, but absolutely mind-blowing at the time.

But I doubt any of this is convincing you, slaapliedje. Pretty sure you've made up your mind about Canonical for reasons I certainly can't fathom. But there's some context for why I'm repeatedly impressed by Canonical and what they've brought to the Linux desktop.

I knid of think of the story of the rabbit and the turtle. Canonical is the rabbit in this, trying to hurry up and get to the end without all the planning and development, so they can get something out quicker, but is a mess underneath. Where Redhat is slow and methodical and ends up winning in the end.

Wow, more casual Ubuntu-bashing, eh? Upstart replaced the SystemV init system in 2006 and was used by various distributions (including Redhat, ffs) for about 7 or 8 years. It was instrumental in pushing the start up times of Linux from the 30sec+ norm in those days to the sub-10s we have now. Iteratively, it was surpassed by SystemD and that's fine. That's good, that's what makes Linux better, over time. Upstart wasn't "a mess underneath", nor was Mir, nor Unity, nor Snap, nor many of the other truly exceptional, innovative things Canonical have contributed over the decades they've been around.

Occasionally they did indeed put out duds. That's also fine. As long as we're all running Linux, we're all in this together.

Although from the comments on this thread, you'd hardly fucking think it.

The problem with Ubuntu is they think of people as Ubuntu users, not as Linux users. Upstart was used briefly in Fedora, It was never used in RHEL. Snap is a mess. Unity was... meh.

Briefly? Like 5 releases briefly? I've never thought of five years as a brief time, I have to admit. It was also a beta option in RHEL for a similar time, but you're right, it was never adopted, since start-up time on a server O/S is pretty insignificant compared to the POST checks in most Enterprise environments. I vaguely remember Debian considering it too, for quite a while, but I can't remember their reasons for not adopting it. Back then, their focus was and still to a large is, servers.

But anyway, my point is that Upstart was very well thought out - a fully asynchronous service management/init system with a simple interface, massively extendable, yet completely backwards compatible with SysV.

Snap is a mess, is it? Based on what?

Unity was meh? Based on... ah, okay. Your opinion, I suppose, which is fair enough. Opinions differ. Mine was/is that Unity is a great example of Canonical's innovation - reacting to the options in front of them at the time and, crucially, doing something about it. Back then (around 2010), Gnome Shell - now that was a mess. A wholly forced, non-configurable change in workflow. Canonical had to react to that, and Unity was born. It was a brilliant half-way house between gnome 2 and shell. It was, itself, also a bit shitty until the 2012 releases, but that's iterative development for you.

Also back then, remember, Canonical were pushing the idea of a converged mobile & desktop experience. The "Edge" phone was coming, powered by Unity. Incredible vision, when you realise this was twelve years ago. I remember being blown away by the possibility of using my phone as a dashboard screen on my desktop, and resizing, mobile-aware apps - something we take a little more for granted these days, but absolutely mind-blowing at the time.

But I doubt any of this is convincing you, slaapliedje. Pretty sure you've made up your mind about Canonical for reasons I certainly can't fathom. But there's some context for why I'm repeatedly impressed by Canonical and what they've brought to the Linux desktop.

5 releases of Fedora is 2.5 years, as they do about 6-ish months. And yeah when you're my age, 2.5 years is nothing.

Evidence of snaps being a mess? Just look at what everyone says about their slow performance, their singular store source, the fact they've had crypto mining bits hidden in programs uploaded.

Anyhow, I, nor anyone else, needs to justify to random people online why we don't use Ubuntu, just as much as you don't need to justify why you'd use it. It's my own personal belief that Ubuntu was, as their mission statement said (yes I was there at the beginning of it) was to create a 6 month release of Debian based upon the Gnome desktop was later on ditched and became just another 'based on Debian' distro for me. So I moved back to using Debian and have enjoyed not having any proprietary software installed on my system in a method other than the one I choose myself.