Hacker News Reader

Comment by gshulegaard | Hacker News Reader

I may be a minority, but I am very saddened by this. Not because I have any particular love for Unity, but rather I share Mark's conviction that convergence is the future.

Love or hate it but Unity was IMO the best shot we had at getting an open source unified phone, tablet and desktop experience...and now this is effectively Canonical not only shutting down Unity, but refocusing efforts away from convergence and towards more traditional market segments. I mourn the death of this innovative path.

That said, hopefully this convergence with GNOME will eventually lead back to convergence...but for now that dream is dead it would seem.

View on HN

data convergence is the future, not interface convergence. I think they made a serious mistake conflating the two.

You should be able to shift your view of a document from a desktop to a laptop, but that doesn't mean the fundamental interface from one should be shoe-horned into the other.

That could mean a phone that is docked, but when you have 10x the screen real estate, a keyboard and a mouse the interface should be different when docked vs when not docked.

Agreed.

Incidentally, what you describe as data convergence is part of a theoretical computing model I call PAO: http://tiamat.tsotech.com/pao

> That could mean a phone that is docked, but when you have 10x the screen real estate, a keyboard and a mouse the interface should be different when docked vs when not docked.

At least in the demos I saw, the mobile UI would turn into a traditional Ubuntu Unity look, with floating windows and keyboard+mouse interaction.

So it sounds like they were working in the direction you describe.

The dumbest move, IMO, was not using Wayland and working with that dev community to improve it's functionality on mobile (one of the reasons given for Mir). Shuttleworth seems to point this out in the post.

Turns out rolling yet another display server rankles noses and takes a lot of dev cycles.

We have a solution for data convergence, SD cards.

You mean dropbox.

Yeah, that's good for a lot of people too. Github/Gitlab is great for software projects. I have rsync for my music collection, etc.

You mean Nextcloud.

Agreed. I used to bevery enthusiastic about continuum in Windows. But now, I've come to realize that the interface better be separate for different usecases - plus, no one want to plugin their phone and use it for work. What I want is seamless takeover from one device to another - something like Apple's continuity but way better. In today world this is all scattered into different components*, the first company that does everything can lock in a huge revenue stream.

a) Files - Dropbox b) Activity - Various browser sync for bookmarks, history (and passwords) c) Handoff - Mac is pretty good here. Windows has no equivalent.

> I used to bevery enthusiastic about continuum in Windows. But now, I've come to realize that the interface better be separate for different usecases - plus, no one want to plugin their phone and use it for work.

You seem to have a misunderstanding of how continuum works in Windows. The interface is separate when you hook up a phone into a dock (it's a desktop interface) and the newer devices can also do it wirelessly.

Windows went with a "similar UI language" for all platforms but still maintained huge differences between each platform because that made sense. It was never the same UI.

I have to say that I find the concepts of "convergence" presented so far by various companies a bit much though.

As you said, unified interfaces aren't that important. I'd be perfectly happy with a phone that runs a Linux desktop while docked and gives me an Android GUI/Stack while in "portable mode". The only thing really needed then would be the possibility to open the Android part in a window in the "Desktop mode", so one could still do messaging etc. while docked.

Personally, I think that would already get us 80% there. The much bigger hurdle (and afaik the one Canonical had trouble with) is then establishing a market position.

I disagree - data convergence is already here if you want it. Device/UI convergence is the pinnacle of everything to me. Meaning my phone should be able to power everything from a laptop shell to a monitor. That was the exciting part, not just an interface. No one cares if context on UI changes per device implementation - we already deal with that daily.

I agree. I want my phone to have a phone interface and my computer to have a computer interface. I also want to switch seamlessly between the two.

Who cares if you have the same interface on your phone and PC. I mean seriously, there is absolutely no point and you end up wasting time designing some sort of interface that doesn't suck on either platform.

This was what was being attempted. The UI would be optimal for a phone, until you docked it at which point it would adjust to take advantage of the new peripherals and larger screen. Developers could write an interface that morphed, or even completely different UIs, for the different environments.

And why developer would in an ecosystem of mostly free app?

This is a ton of added complexity to developers without any real return but for ubuntu.

And ties programs to ubuntu flavored linux, unless obe is willing to do even more work for free.

And maybe multiple non-touch monitors, where the comfortable mobile interface feels cartoonish.

I'm with you. I'd like a fully free software stack across my personal devices, and I also thought that Ubuntu had the best chance of achieving this. I think the issue is in a business model that would work for Canonical.

OTOH, most of the hard work is done. My Ubuntu Phone mostly works. The stack is complete. I'd love to see interested people taking that code and keeping it going. I'd spend spare time on contributing if this were to happen.

(I'm a Canonical employee and Ubuntu developer, but not on the convergence/desktop/phone side, and my opinions here are my own and not that of my employer)

A number of canonical employees here I note - OP included. So what's the mood like 'a zillion lines of code and four years later' internally? Does juju look like any better bet? Go on - we won't tell... ;)

It looks like you also have a charisma at telling dad jokes.

personally always preferred fvwm to unity...

And Wayland.. Wasn't it all because of phones and tablets?

Wayland solves a number of other problems than phones and laptops. It modernizes code left over from the dumb terminal world. It also should be pointed out that the dream is not dead. Another project (hopefully Gnome) can pick this up, so the effort isn't lost.

Which dream is not dead? Wayland? It's definitely not dead, as Wayland has both GNOME and Red Hat behind it. Ubuntu decided to roll their own (Mir) instead of also going with Wayland.

I think we both mixed up Wayland with Mir.

Most likely Ubuntu goes wit Wayland now instead of Mir.

But how can I get a phone that can actually run the code first?

As someone who applauds this pivot, I never saw much value in the whole convergence idea. Why would I want my laptop and phone to be the same device? This seems to follow the misguided idea in software that if two ideas are similar, just make a single generic idea that solves both problems. But then you get into leaky abstractions and have to make lots of sacrifices to get it working well for both use cases.

Application convergence already happens via the browser; this takes much wind out of the sails of OS convergence. However, it also means that browser focused OSes, ChromeBook and FireFoxOS, will probably be the ones to win, in the end.

I thought they cancelled FFOS, didn't they?

They did, though of course it's open source and somebody else could pick it up and continue development if they so chose.

Or WebView-based cross-platform toolkits will win, in the end. (Electron, Cordova, React Native)

Given FirefoxOS is also retired and ChromeOS seems to be in the process of being replaced by or merged with Android, I figure the cross-platform toolkits are more likely to win than the browser OSes.

They share the browser limited performance and bring a load of issues with them, they're at best a way to work around browser -> sensor access and bypass the appstores barrier.

For now, at least. The thing that will happen is that as they continue to "win", they find deeper ways into the platform. Cordova apps on Windows already run "first class" in the UWP app platform's JS stack. Electron apps don't currently, but there are bridges being built there too.

React Native and NativeScript are exploring different tactics beyond the most limited webviews on Android and iOS systems. Crosswalk and others explore yet another approach for bringing better webviews to such platforms.

At some point too, performance doesn't matter so long as the users are happy. A lot of technical folks notice performance, but most consumers do not. WebView-based application toolkits don't need performance to "win", they just need that sweet spot of developer productivity and user engagement that is much easier to do than you think, with or without "performance".

If "winning" means "same operating system on all devices" then I don't think Electron, etc are competing in that contest. ChromeOS (which I understand to roughly consist of a Linux kernel and Chrome, and not much more) is by far the closest in concept.

Imagine, a process per browser tab for all the things! Why should this not be the case? It would be a revolutionary improvement in process visibility, in general. moreover I don't see why there would be any limitations over what you get with ps and dtrace.

Indeed, I suspect there is a quite nice visualization of the kernel itself that would fit nicely in a browser tab. Maybe not a real-time picture, but surely you could do simple simulations at the very least, and perhaps set configuration options if you wanted to change your own kernel.

Portability and duplication. I don't want to have to lug a phone and laptop everywhere. They both can do the same thing and increasing have the same amount of power. You have different interfaces for how you interact with them: on the go or at a desk. But you can just pick up and go somewhere else and pick up without missing a beat.

And who brings the dock in that somewhere else you're going? How's leaving a dock everywhere better than leaving a laptop or even a chromebook everywhere?

I was always under the impression that convergence for the end user meant having the same 'computer' spread across many devices. That your phone has access to the same files as your laptop as your tablet. That identity and rights management was shared and unified across all devices. That your devices natively formed a network, a sandboxed cloud.

I never understood it all looking the same, having a shared UI. The screen real-estate and input are too different to make a unified presentation. My biggest complaint about Windows 10 is that the power user on a laptop suffers for the imagined mobile user.

The funny part is, focusing on shared data and identity brings you closer to convergence than an elegant UI does.

I have to agree that sharing the UI seems like a dead-end.

We saw this in the leap from WinCE to the iPhone, and from tablets that were essentially laptops with the monitor facing outwards, to iOS & android tablets. Besides the hardware finally getting to the right place, designing the UI for the experience was a huge leap.

"Convergence" feels far too much like trying to compromise one to suit the other, right when we've finally learnt not to do that.

I totally disagree actually. With relative position pointing (except for the onscreen keyboard) desktop UIs actually are significantly more usable (not to mention, often more responsive) on my android than the phone centric android GUIs.

The only problem with this is that SDL X11 eats the battery.

My phone is a computer. My chromebook is a screen attached to a computer. My desktop is a monitor attached to a computer. My work laptop is a screen attached to a computer. My tablet is a screen attached to a computer. Why do I need so many separate computers when I have my phone on me at all times?

Because your phone will never have the raw power that you can pack into a larger form factor? Granted, after a certain point this may only be relevant for specific tasks.

Pshhhh 768k ram ought to be enough

But what if I could plug my phone into a dock that gave me substantially more power, but I didn't have to make a massive context switch i.e. wait for syncing over a network for files etc, or log in again and load up what I already had open?

Because my work computer is owned by work, my phonr by me, and the desktop by my family. They don't want an unused monitor and docking station when my phone isn't there.

Don't they all have their own phones?

I don't really like the one device to rule them all philosophy, a single device would be too compromised for the different use cases I have for computers, but I don't think this is a valid objection in this vision of the future.

Computers aren't special, rarified objects. Computers are dirt cheap, and it's simply more convenient to have lots of different computers lying around than to have to move one around.

No one has to worry if the compute core of the family room computer is available, for example- they just use the family room computer, which is its own separate device, and already affordable for nearly everyone.

It's not that convergence is useless per se - indeed, the fact that both Windows and Android keep trying indicates that it's desirable.

But I think that any hope that this could come from the desktop Linux world, and be successful, was misguided, for the same reason why "Linux on the desktop" never happened at large scale (to all the people who are about to respond "it happened for me": sure, and I'm well aware of that - but that's not what the phrase actually ever meant, and we all know that). This is an even bigger problem with phones, because with the PC ecosystem, at least you can go around shopping for an alternative OS relatively easily, and common hardware is relatively open spec-wise. So something that has 1% market share on the desktop, will probably have an order of magnitude less on mobile.

On top of that, it's not enough to implement the base system. You also have to maintain it long-term, to port existing apps to it or write new ones, and to maintain those. Within the OSS model, maintainers are generally users. So the maintenance burden is high enough for desktop Linux as it is (because of relatively fewer devs/users, compared to OS X and Windows) - and then you throw mobile into the mix, which is an even smaller group. I just don't think it can work at that small of a scale.

Spot on, these were the major challenges that always faced the idea.

Interesting to see the latest Samsung phones doing the same thing, but running Android and having Samsung behind it that attempt is much more likely to bear fruit.

Fair play to Cannonical & Shuttleworth for giving it a try, but widespread adoption was always a pipe dream I think.

I am sad too, in fact, a scan over the comments suggests we aren't so much in the minority. I think too that his conviction was right, it is merely possible that his strategy of going out on his own on Mir wasn't the right one.

I'm not sure if going with Wayland will hamper the eventual convergence, but I doubt at this point we will see an open source convergence before we see one in proprietary form.

It's here already called the Samsung Dex: https://www.youtube.com/watch?v=GEPegrUMXaA

https://news.samsung.com/global/hands-on-8-galaxy-s8-feature...

Kudos to Samsung for the amazing efforts but DeX is far from open.

Of course but I meant that the proprietary version was here.

Then you can add Windows Continuum and whatever it was Motorola's project was called to the list.

Literally a week ago. Could this be related to Shuttleworth backing down?

Convergence doesn't need to go away. KDE project for example put some effort in making common components[1]. It's just that Canonical's methods of handling it went completely off on a tangent with reinventing the wheel (Mir instead of Wayland and so on). So it's not surprising, they finally realized the huge complexity of that task, and inability to do it alone.

No one stops others contributing to projects which still work on convergence, but use shared stack at the same time.

1. https://community.kde.org/Plasma/Convergence_Overview

While I agree it'd be nice to have Linux running on all the things I don't see the point of trying to make the interfaces the same. Each device is used differently and should have an interface that reflects the way it's used.

While interfaces should not be the same across devices, being able to dock you phone and continue using its applications on normal monitor of your desktop, with keyboard and mouse would be convenient.

I mean when you walk it's a phone, when you at the desk, the same device is connected to monitor, keyboard and mouse.

But then each application is essentially 2 applications in one.

It's the same application, but with 2 different interfaces. Of course, a UI toolkit that can do "write once, interact in any way" would be helpful for the implementation.

Which is essentially two applications as I said. The only difference is that they may share some state, but there are other ways to handle that too.

Just like any responsive website is.

So it's easy with what is essentially a document viewer, it's much harder when it comes to real apps, the desktop interface is dumbed down for the limitations of a phone.

What I mean is that "convergence" benefit is not same UI on different devices out of pure whim, but the ability to use the same application from phone directly or using desktop monitor mouse and keyboard.

I don't mean the interfaces should actually be different. Just that "the same interface" is not the motivation by itself.

Unified interface would be good for that. If the interfaces are actually different, the app could be adopted automatically by its UI framework, or by "window manager", etc.

Convergence already means that GNOME on my (quite un-rotatable) desktop PC has a rotate-lock button and that I have to do an awkward slide-to-unlock gesture with a mouse to get to the login screen.

No thanks.

Just FYI, if you hit enter or just start typing your password the lock screen slides away. Or at least it does on Fedora, but I can't imagine that's anything customized.

Also works in Ubuntu.

I personally really liked Unity, and I was following Unity8 development for a long time so I'm genuinely disappointed to see it go - partly because I really thought it could shape up to something really cool and partly because as you say, the convergence dream is basically dead, as is any realistic future of Ubuntu on touch devices.

I know Gnome has for ages built features with some theoretical touch device in mind, but only Canonical/Ubuntu has ever shipped remotely viable touch based hardware.

Does anybody besides platform developers want a unified phone, tablet, and desktop experience? I certainly don't; it sounds like a very bad idea.

Engineers overengineering things when they have no new ideas. That's the entire convergence story

Phones are already damn powerful, and will get more so.

Why should we all have to buy another completely separate computer if we want to use a mouse/keyboard/monitor, when we already have phones?

You have that backwards: why should your phone present the same UI when it's plugged into a keyboard and mouse and when it's not docked?

Correct

It makes no sense from an usability point of view

It's like insisting on a steering wheel for a plane.

Different means of interaction and limitations and use cases require different interfaces.

Also the "Ubuntu on the phone thing", it seems they were doing it because of the wrong reasons (mainly "We like Linux") rather than focusing on the user and the ecosystem

I'm for device convergence, a small phone as powerful as a current pro laptop. We're still not there and the battery would be a problem too.

The interfaces for the phone mode and the laptop mode could be very different, no problem with that.

With that in mind I welcome this move from Canonical. Converging on the same stack as everybody else is going to be good for the Linux desktop. I'm not worried anymore about compatibility issues, Mir and Wayland.

When the phone form factor will be able to fit enough computing power to be also a laptop, we'll find a way to design the appropriate UI. However note that with a fast and responsive network connection we could use phones as dumb terminals to remote processing power, so maybe that will be the solution to convergence.

I agree. I think once (if/when) we ever get next gen battery tech, we may start seeing more things changing.

MaruOS might be interesting for you:

http://maruos.com

https://github.com/maruos

Only the Nexus 5 and 7 are fully supported at present though.

Huh. I have a barely usable Nexus 7 2012. Might be worth getting a used Nexus 7 2013 in order to play with this.

I agree, I was really looking forward to it. Imho the mistake was to partner with Meizu and BQ whereas distribiting images (at least to start with) for the Pixel/Nexus line phones would have been more effective. That way people could have put their toe in the Ubuntu Phone water before committing. I donated to the dev that was working on the OnePlus 3 port, I was really looking forward to it, it could have been my next laptop (if only the OP3 would have supported HDMI via the USB3 port or MiraCast.)

the requirements for a good desktop interface is different from the requirements for a good phone interface and arguably different from the requirements for a good tablet interface

size matters

yes like you i believe one OS across all devices is a good idea, but the interface need to adapt to screen size, and provide a different and suitable experience for certain categories of screen sizes and different mode

I believe a good option would have been to have a touch-mode and a desk-mode (desk mode being a more traditional task bar desktop UI)

anyway, it is really weird they didnt see this, with all the talent they have and the resources they had, but, they failed, so they were wrong

Someone else will try and will succeed, because yes, one OS experience across all devices is still a very good idea .. we just need someone to execute it better

I don't really see why there are comments like this in this thread. Obviously they need to have a different UI, but Ubuntu Touch already offered a different UI depending on what it was running on. Or at least that was the idea. See the animation on https://www.ubuntu.com/mobile.

No one was trying to make a one-size-fits-all UI.

Windows? Microsoft are pushing ahead with development of UWP and Continuum, with the creator's update due in 3 weeks - despite no new Lumias.

At least one third party is still keeping the dream alive with a crowd-funded WP10 phone announced this past week:

https://www.indiegogo.com/projects/cerulean-moment-smartphon...

It's an interesting device, but I think it's going to be a dead-end because of ARM.

The moment someone makes a phone that runs one of those Intel mobile chipsets, which can actually run x64 version of Windows, and all the existing apps, is the moment when I'll buy into that "it's a smartphone! it's a desktop! it's both!" line. But not before.

I'd consider windows 10 evidence that proves convergence is a bad idea. It does an absolutely terrible job of being a tablet and a desktop OS. At times it feels like two OS's in one and at other times it feels like you're stuck in the wrong mode (and after a reboot you are).

Half the apps (explorer and vlc are the two I used the most) work horribly in touch mode and the other half work horribly in desktop mode, like the way the pdf reader default to full screen.

It did have a touch mode and desk mode ffs!

I agree about the intuition about convergence, and I'm not at all persuaded about the comments that "we want data convergence but not device convergence". I think that without device convergence, we'll see the desktop continue to become marginalized.

And I agree about Ubuntu, and thus Unity, looking like the best shot for a free software unified OS.

But none of that was gonna make me use Unity on my Linux desktops. shrug

You're not, I don't like Ubuntu as a distro as I believe they've put testing and security in a back seat, but the one thing I really was excited about was the convergence of mobile and desktop, it seems like the future to me and it's something I've wanted since my first smart phone.

No!

Convergence and ramification are the cycling nature of any system's evolution path. Not just convergence.

Reply on news.ycombinator.com