Hacker News Reader

Comment by erikb | Hacker News Reader

It is not surprising if one follows their development, though. Neither Unity, nor mobile, nor Amazon integration was a move that anybody liked. And many of these were just announced to be dropped. That this strategical shift also means a shift in leadership is normal.

And honestly, I have to say: Finally! Nothing of these things were really doing any good to the community as a whole, and therefore also can't be in the long term financial interest of Canonical.

The only thing I wonder is if this Jane Silber, who I've never heard about before, is really responsible for all these failures or if she's just the goat that gets slaughtered for the oracle of public opinion.

View on HN

Plenty of people like Unity, it's just that the naysayers are super-vocal about it. Mir was an attempt to meld mobile and desktop at a time when mobile still hadn't really settled down - it was a bold experiment that didn't work. Amazon integration was another experiment that didn't work. They have tried a lot of things, some work, some don't.

But if ubuntu listened to the naysayers, there wouldn't be an ubuntu, and desktop linux would have nowhere near the user count it has today. Ubuntu drove a lot of development towards user-friendliness (directly and indirectly). Another of their success stories was Upstart simplifying service management, which was only unseated by Systemd (which started development nearly half a decade later with the benefit of hindsight).

It's so tiring hearing the constant bitchfest aimed at Ubuntu when they've done so much for the community, including significantly growing it. Yes, they've made some missteps, but so have all of the big players.

I do not understand how Ubuntu brought users to the Linux desktop. When Ubuntu started, it was just a crappy version of Debian. It was exactly as user friendly (or user unfriendly) as Debian. It ran the exact same software! Everything that made Ubuntu user friendly (or not) was there on Debian.

Then they made Unity, but that was after it was established in popular culture that Ubuntu is "user friendly". But now they are abandoning Unity anyway, so it must not have been that important after all.

I do not understand why Ubuntu became more popular than Debian. I really don't. Something to do with those free CDs.

I think one of the major driving factors initially was that Ubuntu didn't follow the long package tree freezes that Debian stable/testing was subjected to after a period of time. Debian stable still tends to contain older packages with security patches applied rather than the latest and greatest(tm), which most standard users tend to prefer over stability

More frequent releases. Mattered a lot for hardware compatibility.

Does Debian come with a graphical installer yet?

Debian has had a graphical installer since at least 2007.

Yes, of course.

But ... I've never understood this fixation with graphic installers, let alone why they are considered so important in assessing the quality of a distribution.

You install once, and it's fifteen minutes of your life. During installation you need to inform the installer your hostname, domain (maybe), IP address information (maybe), location and timezone, keyboard layout, and disk layout (most people pick the 'sort it out for me' option anyway) ... none of which really benefits from, let alone requires, a mouse.

After installation it's all about the maintenance, and that's where Debian shines. Install once, upgrade for the next decade or more. You can use a GUI for management, if you want, though I'm not convinced GUI distro / package management tools are better than the CLI.

Perversely colleagues and clients that have chosen to rely on Ubuntu typically end up having to do full re-installs every time a new LTS comes out, as the migration process often simply doesn't work for them.

A mouse is useful if you don't know how to tab around boxes. The disk layout option is super-scary for beginners, and full of warnings.

I typically don't know what my 'domain' is (I've never known to this day what I should type in that box, and nothing I type ever seems to make any difference).

> A mouse is useful if you don't know how to tab around boxes.

I've heard it said that the only intuitive interface is the nipple - after that, everything is learned. Tab (to move) and space or return (to select) is not that hard to get a handle on - and, really, if that's the limiting factor for a computer user that's got their hands on a Debian installation (from 8+ years ago when GUI install wasn't an option) and managed to convince their computer to boot from CDROM / floppy / ZIP disk etc ... they almost definitely have some bigger problems in their near future.

> The disk layout option is super-scary for beginners, and full of warnings.

Disk layout options are comparable in GUI and text-based installs. Ditto the warnings.

I note that in both (and it's always been in the text-based) there's been a 'pick an automatic layout' option. Again, if you're dual-booting, and don't know what you're doing, then a GUI isn't going to be the thing that saves you from snotting your Windows partition during this process.

> I typically don't know what my 'domain' is (I've never known to this day what I should type in that box, and nothing I type ever seems to make any difference).

Same question is posed in every GNU/Linux distro installation, text or GUI, and in every instance of either that I've seen, the option exists to ignore it. Again, the problem of a fqdn isn't a function of whether you're using a text based installer, or a GUI one.

It makes things look friendlier to a user lacking in confidence, and less like "your computer is dying, look it's dropped back to text mode"; makes them less likely to balk, or at the least start their experience with unpleasant feelings. Particularly when doing, e.g., dual boot, i.e. risking your Windows stuff.

> Everything that made Ubuntu user friendly (or not) was there on Debian.

It was there in Debian (mostly) but it lacked that final bit of polish for the average user. Debian has always been very conservative as well (which is a good thing for some situations, but not necessarily for a desktop) and its documentation and guides on the terse side.

Ubuntu just clicked. It was, simply put, Linux for human beings. The user community at that time reflected this as well.

It's not just the free CDs that did it.

> But now they are abandoning Unity anyway, so it must not have been that important after all.

Unity filled a void left by the end-of-life of Gnome 2. Gnome Shell at that time was looking okay, but it had way too many bugs — and this is to be expected in a new product with new UI paradigms and libraries.

Unity was Canonical's attempt at providing a stable desktop environment that wasn't quite as revolutionary as Gnome Shell, and that could succeed Gnome 2 in stable releases of the distro. They pretty much succeeded there in my opinion. Today Gnome Shell likely is ready for mass use with all the early bugs ironed out (I hope), so Unity may not be needed as much now, but it did what it needed to do (and very well at that) for years.

> It was there in Debian (mostly) but it lacked that final bit of polish for the average user.

I've heard this before, but I don't buy it. I have been using Linux for 20 years now; back in the day, and back when Ubuntu came out I was using it on my desktop (today, not so much) and I simply not believe this claim to be true. It's often repeated, but never with concrete specific examples on what was more polished and how it was more polished.

In fact I believe the opposite to be true. Back then I was using Debian on my desktop. I naturally tried Ubuntu and found out it was less polished because it contained more bugs.

> never with concrete specific examples on what was more polished and how it was more polished.

Network drivers for anything other than an intel card. Still, to this day. They're not in Debian for philosophical reasons, but they're a pain-point to non-expert users.

Not just those, Nvidia drivers!

Ubuntu was the first distro I was able to successfully deploy to non-tech members of my family, friends, and clients. For people who just wanted to browse the web, watch videos, etc and already had old hardware lying around, it was (and is) perfect.

(If someone wants cheap and simple nowadays, I recommend Chromebooks. But if they need free, as in I can re-use some of the flow of decommissioned machines from work, Ubuntu it is!)

I found the standard "put the disk in, press 'yes,yes,yes' " with Ubuntu got me into a X server, with nice Gnome configuration, with a working Network.

Debian did none of these (at the time), it was a text installer which asked loads of questions I didn't really understand, and didn't install X or gnome or install wireless drivers by default.

Of course, I suspect the things you had to do to Debian to make Ubuntu wasn't very much -- just hide lots of installer options behind an 'advanced' option, and set a bigger set of default packages to install. But it was enough to help me get into Linux.

concrete specific examples on what was more polished and how it was more polished.

Sound and video playback worked out of the box for far more cases than with Debian. Networking just worked for far more network cards. X just worked the way you expected it to in many more cases. And while it certainly was buggier than Debian Stable, I don't recall it being buggier than Debian Unstable on which it was based.

Another aspect was the community surrounding Ubuntu. If you asked a really 'dumb' newbie question on the Debian mailing lists chance are you'd get a lot of condescension and RTFMs and probably not so many useful answers. Ask the same question on the Ubuntu forums and the chances where much better that you'd get a clear simple answer that actually solved the problem you where having.

ubuntu-restricted-extras was one of the packages that was a deal-breaker for me, that solved a common pain point at the time.

I think you're 100% wrong there. Ubuntu supplied, and supplies, polish that Debian just doesn't bother with. You're grossly underestimating the It Just Works factor.

(I remember trying to move back to Debian. Apart from Xfce compared to Xubuntu looking like it's had a two-day session with the ugly stick, getting my wifi to work felt like I was being lectured by a small RMS on my shoulder.)

I do like Debian a whole lot, but for my purposes Ubuntu is functionally an acceptable Debian. (I use Debian on my personal server, Ubuntu on work servers and Xubuntu for my desktops.)

> Ubuntu supplied, and supplies, polish that Debian just doesn't bother with.

This is often repeated but never with any concrete and specific evidence.

You want concrete and specific evidence for a subjective judgement call based on peoples own experiences?

Where's your concrete and specific evidence that nearly everybody else's subjective personal experience is somehow objectively 'wrong' and needs justifying?

The 100 Papercuts projects, for example.

Debian has bothered more in the decade or so since Ubuntu started. And, of course, fixes go upstream where applicable - Ubuntu deals with a lot less platforms than Debian has to account for - and Canonical employees work directly on Debian in a few areas.

But dude, I remember how much pain Breezy Badger wasn't compared to Debian at the time, in so many ways.

Multiple factors: Ubuntu had the Linux for human beings slogan, and at the time that was super important. They promised to create a distro that normal users can use, in contrast to the elitist attitude that prevailed in other communities. That was heavily reflected in the mainstream press Ubuntu reached seemingly effortlessly.

The slogan was also taken seriously. If you asked in my local Ubuntu community (UbuntuUsers.de) there never was a RTFM or a google link. You got help, even if the question was asked already a hundred times that day. We combined that with writing the best linux documentation that exists to this day, but that's specific to the german Ubuntu community.

Also, the installation was easy, it had a good installer. It wasn't even necessary to burn a CD, you could get it shipped. Not everyone had DSL back then nor CD-burners (well, that's true today as well, but now we have USB-boot). That sure helped.

Debian has a lot of rough edges. I still wouldn't recommend it for a naif. You have to understand the Debian philosophy to not get angry at how Debian's default install simply doesn't work (still!) with most network cards. Debian has had a lot of user-friendly polish in the last half-decade, but was pretty gnarly before then.

I came over to the linux side around 2007[1], and Debian was painful with network drivers and xorg configs, whereas ubuntu just gave you the binary blobs (like most distros) and worked on making X config less painful. Debian at the time was also exceptionally slow to bring in software, and that's improved in recent years. The user-friendly reputation of Ubuntu came well before 2010 when Unity came out - basically in its early days, Ubuntu gave you a distro with APT (= bye-bye, dependency hell) and also had friendly UX as its primary goal. Debian was about user-friendliness in terms of philosophy, not in terms of UX.

I now use Debian on my workstation, in the cloud, and on my home laptop. But I would never recommend it to anyone without making them aware of the Debian philosophy first, and it's a philosophy you have to buy into at least a little bit. All the other devs in my workplace are either on macs or ubuntu/unity.

[1] My very first linux experience was running ShowEQ, and dependency hell made me swear off linux for almost a decade... :)

Well put, in my mind Ubuntu will always be a great gateway drug to Debian. Ubuntu community did a fantastic job to popularize deb ecosystem even though it is often criticized for not contributing enough upstream. I can't imagine anyone using Debian as their first Linux distro, Ubuntu (and it's derivatives) on the other hand is the Linux distro. Debian is arguably more stable and technically might be better implemented but Ubuntu has won over the average user because of ease of use and helpful communities around it.

> Ubuntu community did a fantastic job to popularize deb ecosystem

This can't be emphasised enough. I was a Debian user before Ubuntu launched, and it really seemed like Debian was going to only ever remain an unsupported (by anyone else) niche in the face of Redhat.

> I do not understand why Ubuntu became more popular than Debian. I really don't. Something to do with those free CDs.

Certain things just worked, like hardware. For a time, X worked on Ubuntu with my hardware and I had to jump through hoops to configure X with the hopes it would work on Debian. My touchpad and sound card just worked. Proprietary drivers were easier to install. Desktop packages came with sane defaults. There were helpful forums and IRC channels.

The crucial point you missed was it had a really polished installer, easily better than Windows and contemporary Mac OS.

My non-tech friends managed to install it from CDs that came with magazines with almost no guidance.

Of course after you have everything setup properly, it is hard to distinguish between distros (other than ..say.. package manager), but those choices do not drive the vast majority of people.

Amen.

Also, I am not sure the melding of desktop and mobile is such a bad or even a dead idea. It has apparently died at Canonical but I think that may be due to the position Canonical specifically is in. I mean Samsung just entered this space in a big way with Dex...and that is 4 years after the original Ubuntu Edge Indiegogo.

I have lamented elsewhere over the death of the Unity mobile direction...so I won't repeat myself but I still think it would have been really cool.

Had they listened to the naysayers, they would have used gnome shell all along, the development resources would have been more focused, and desktop Linux would be better off.

I recently switched from unity to gnome shell and it's much much better. Unity's bugs were simply embarrassing.

I honestly don't get the 'community' bit. Who gets to decide what is NIH, what is duplication of effort, fragmentation and what is 'good' for the 'community'?

Who do these individuals speak for? Which community? There is no organized overarching community in Linux. Are they speaking for Debian, Redhat, Fedora, Suse, Mint, Ubuntu, Elementary, specific projects? They should clearly articulate who they represent.

This looks like hijacking the 'community' to push personal preferences. I am beginning to feel there is a serious undercurrent of bad actors in the open source ecosystem who are threatened by anything not from that they perceive as 'their community'.

For instance I don't see any knives sharpened by all these 'community' advocates for Flatpak when it is preceded by Ubuntu Snap. In this case no one is harassing Redhat and Fedora about NIH, duplication of effort, fragmentation and pressuring Flatpak developers to join hands with Ubuntu. Why the double standards?

Wen it comes to Mir or Unity it is completely out of control. And its only when I decided to overlook the bad press to actually try Unity that I realized its by far the most polished desktop on Linux in 20 years. This kind of organized hatchet job is a serious loss for open source.

Its too self serving and political and I suggest anyone presuming to speak for the community should be immediately treated with suspicion.

> What is 'good' for the 'community'?

Decision means someone has to do something. So probably nobody decides that. But everybody is free to have an opinion. Everybody presents what they see as "the community" and feels a fantasy connection to people who he has barely anything to do with, and a distinction from people who he may have more in common with than he assumes. That's how community works. It's mostly in people's heads.

Well if we're speculating, I'll speculate something else. Mark is shopping the company around, trying to find buyers. The buyers told him he needs to axe some projects and fire Jane. These things might not be related outside this context.

It even says in the other article that they are downsizing due to an "external analysis". External analysis happen when you are about to be bought.

It says in this headlined article "We're now entering a new phase of accelerated growth".

I think 4ad has a good point here. It's been a long time since I has the same kind of thinking: Shuttleworth won't spend all his money in Ubuntu indefinitely and/or he wants his firm to be financially sustainable over time. They've tried hard to make a living out of desktop market, and it never worked out. Now they cut it and focus on B2B markets. No longer expect Canonical to make innovation for the end user market in foreseeable future.

The way he is expressing his bitterness to the open source community in its Google+ post is telling.

We may expect Ubuntu to drift slowly into the once-in-a-time-popular distribution graveyard.

I think you've nailed it. Every move and announcement screams that Shuttleworth has given up on his ambitions. Which is a pity, because the ambitions themselves were laudable.

Could be the reason.

I don't understand their strategy; Red Hat has some unique offerings for enterprises, but I've never heard anything from Canonical.

Getting involved inthe systemd and Wayland turf wars is probably not the most effective way to spend developer resources.

>The only thing I wonder is if this Jane Silber, who I've never heard about before, is really responsible for all these failures or if she's just the goat that gets slaughtered for the oracle of public opinion.

The answer is in the question..

(Though I still do like Unity, it's the least memory-hungry DE of the popular ones)

only siths deals in absolutes? Maybe try: "Unity didn't provide enough advantage compared to use an already baked DE compared to the cost of making it". I would be happy also with: "not enough people liked unity or mobile". Plenty of people liked unity and I'd like in this discussions a little bit of respect was given to the people who thought, designed and liked Unity. At least acknowledge their existence now that you "won"... I liked unity and actively dislike kde and gnome 3 (gnome 2 was ok but clearly nowadays is old and I still liked unity more).

Many people liked mobile too: I backed their failed indigogo and plenty of people did, but that was a weird move and it failed being TOO ambitious (I'm not sure they would have been able to deliver in hindsight).

When they did make a mobile, plenty of people would have loved to have a more privacy minded alternative to google (but still open), but they focussed on lower end where google was already present (and where this privacy power users are not) and, I think, made the mistake of discarding the option of going with android that had the apps needed to make the platform viable for users. In a way it's the "unity" move of mobile, but the difference is that on desktop you didn't throw away all linux desktop software.

I left linux on desktop a couple years ago, not because of unity but because

* I suddenly needed to buy a laptop, I wanted it with linux and the dell xps was about to be replaced with the newer. I also didn't bother to understand how to get a lenovo (possibly I wasn't aware you could get it with linux at the time, I don't remember) so I got a mac

* having a mac seemed like a good occasion to try some proprietary software I wanted to try for a while. Both the proprietary stuff and mac os were good and I could develop in virtualized linux so I couldn't see any reason to go back to linux for not linux.

So I think ubuntu failed in the desktop and mobile. Not that others did any better there in the linux camp.

Now for the armchair CEO moment:

If I had to go mobile AND drop unity, a few years ago, I would have rebuilt ubuntu as a multi form factor experience based on android, possibly replacing (or giving the option to opt out from) the google bits.

Now it could be late (but no problem I guess, as they are not doing it :D). This if convergence was the target, otherwise ubuntu + unity on desktop and android based ubuntu on mobile.

I don't know about that android thing. It feels like there are pretty significant differences under the hood that would heavily disrupt the server and desktop areas where ubuntu is already popular

ubuntu server was going to stay the way it was, but considering ubuntu desktop popular seems kind of a stretch though; relatively to other linux distro in the sub 1% of the desktop market? then maybe

> only siths deals in absolutes?

At least we have cookies.

Reply on news.ycombinator.com