Nokia drops MeeGo phone before launch (Reuters) [LWN.net]

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 15:24 UTC (Wed) by teknohog (subscriber, #70891)
In reply to: Nokia drops MeeGo phone before launch (Reuters) by spaetz
Parent article: Nokia drops MeeGo phone before launch (Reuters)

/me hugs his Maemo -- a Debian derivative that works like any distro, with glibc, Xorg, Gtk etc.

IMHO, having a normal Linux distro on a phone is a simple and brilliant idea, so naturally it must be killed by management. The first Maemo tablets were launched in 2005, but it took about five years to get phone capabilities. Perhaps it took that long to realize the competition of other Linux phones. Also the progress of mobile hardware means that Linux no longer needs the most bleeding edge smartphone.

Just as this Linux distro on a phone became reality, Maemo was merged with Moblin into Meego, and nothing good has seemed to come out of the deal.

to post comments

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 15:48 UTC (Wed) by daniel (guest, #3181) [Link] (25 responses)

"/me hugs his Maemo -- a Debian derivative that works like any distro, with glibc, Xorg, Gtk etc"

Basing Meego on RPM instead of Debian seems to me a big mistake that took much of the momentum out of the project by betraying the trust of the Maemo devs. I wonder if the project can survive that decision.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 15:56 UTC (Wed) by drag (guest, #31333) [Link] (23 responses)

They were idiots for making their own Linux distro from scratch.

Whether or not it was RPM vs Deb is really immaterial.

Switching to QT was another stupid move. QT may be nicer, but toolkits matter a hell of a lot less then Gnome vs KDE ra-ra-ra crowd likes to pretend.

They had years of history shipping Linux devices using the same formula (Debian + Mobile version of Gnome). By the time N900 came out they already were well into the second generation of the OS.

They literally threw it all away to start over from scratch. Huge waste of time.

They should of concentrated entirely on development tools and UI environment from _DAY_ONE_ of Meego. That is what people actually care about.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 16:58 UTC (Wed) by xxiao (guest, #9631) [Link] (15 responses)

When meego was out I commented at lwn with similar opinions:
1. rpm is a dead-on-arrival choice, package format matters in real life, and deb-derivative probably owns 90% of the embedded market. Except for Redhat, all rpm followers are either struggling, or dead.
2. QT is just another nail on the coffin.
3. Intel/Nokia does not have a clue to work with open source community,yet.

These combined, Meego can't fly.

On another note, Ubuntu is moving to Qt instead of working/improving Gtk, so I happily returned to the newest Debian6.0.

My 2011 prediction here: Ubuntu will be acquired this year.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 17:03 UTC (Wed) by jspaleta (subscriber, #50639) [Link] (1 responses)

Don't you mean to say to predict that Canonical will be acquired?

-jef

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 19:18 UTC (Wed) by jengelh (subscriber, #33263) [Link]

Just who wanted to have that, now?

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 17:08 UTC (Wed) by drag (guest, #31333) [Link]

>On another note, Ubuntu is moving to Qt instead of working/improving Gtk, so I happily returned to the newest Debian6.0.

So far they seem to be aiming at making QT suck less, rather then worrying to much about abandoning GTK.

Because it's going to take a lot of work on Ubuntu's part to turn KDE into Gnome 4.0 :P

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 17:41 UTC (Wed) by dowdle (subscriber, #659) [Link] (4 responses)

Your comment that "Except for Redhat, all rpm followers are either struggling, or dead" while possibly true... it really doesn't have any relation to rpm as a package format... unless you mean to say that because of rpm they are struggling or dead... which I don't think anyone would agree was the case.

In fact, name the commercial Linux or open source companies that are doing well besides Red Hat? Are there any .deb-based companies doing well? Not that I'm aware of. Of course I'm not saying that they aren't doing well because they use .deb. I think the package format really doesn't matter much... as .rpm and .deb are pretty darn similar and it is more about the quality of the packages and the skills of the packagers than the package format.

Yes, I mean to remind you that the financial status of Canonical and its ability to tread water is still in question. They don't report their financial status because they don't have to... but if they were doing well, I imagine they would want it publicly known. I don't mean that as a bash to Canonical, any of their employees, or any of the volunteers who work on Ubuntu. Just trying to operate with the facts rather than biases.

The Tools

Posted Feb 9, 2011 18:04 UTC (Wed) by pboddie (guest, #50784) [Link] (3 responses)

One reason for people favouring the Debian tools and packaging system has been the relative convenience of being able to bootstrap a distribution, although I would argue that the tools don't really go far enough to permit cross-bootstrapping without messing around with things like QEMU, even though projects like Emdebian are presumably meant to do so. Stuff like debootstrap, although now complemented by febootstrap (for Fedora), at least points to a more sane solution for getting people started than some of the alternatives (OpenEmbedded springs to mind).

That was one of the perplexing things about OpenMoko: there were a bunch of different approaches, but they were all suboptimal. There was the MokoMakefile which did manage to build things, but then complained loudly if run a second time (if, for example, it didn't work the first time); then there was some kind of OpenEmbedded affair which probably needed gigabytes of disk space and days of build time. For many people, what was really required was just a sandbox running natively (not under emulation) on the development machine (not on the device) which gave a basic sense of the constraints of the device, populated with the libraries and tools that would be present on the device (because the sandbox would be populated in the same way). Bringing up a nested X server would provide enough "emulation" without any need for fancy boilerplate graphics showing the actual device casing, LEDs, and so on.

I regard it as a positive sign that there is a Debian Lenny root image for GTA04.

The Tools

Posted Feb 9, 2011 20:20 UTC (Wed) by xxiao (guest, #9631) [Link]

This is going off-topic a bit, but yes OE takes too long to compile and requires too much space, it's the Gentoo for embedded, but even Gentoo has stageN these days to speed things up.

they could prebuild libraries/compilers by default the way Android did under prebuilt, that could save at least half of the time/space.

The Tools

Posted Feb 10, 2011 23:07 UTC (Thu) by wookey (guest, #5501) [Link] (1 responses)

Bootstrapping an exsting debian arch is dead easy, as you point out. Cross-bootstrapping a new one or a variation, or a new-compiler rebuild is still very difficult, but linaro is doing good work on that front. Emdebian did useful work on this front, but that work showed that without suffient buy-in upstream it was a massive job to maintain. The linaro/ubuntu tie-up has provided that upstream buy-in, so people actually care about cross-buildability as well as native building, and painless developer tools to use that functionality. This stuff, combined with multiarch to remove a whole boatload of cross-path-pain and general bodgery should mean that cross-bootstraping actually becomes something that works, as does cross-building when necessary/useful.

In the meantime of course powerful new arm boxes and convenient form-factors make native-building dead easy too, just like it is on your x86 desktop.

Expect to see some interesting stuff in this area over the next year. It should all be rather spiffy :-)

The Tools

Posted Feb 11, 2011 21:28 UTC (Fri) by xxiao (guest, #9631) [Link]

I tried bootstrap debian from source to no avail, i.e. build debian from source for a new chip, just those perl and texinfo related stuff for setting up the initial native rootfs alone are very difficult...
of course if you want to debootstrap a binary distribution for supported chip it's really easy.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 18:28 UTC (Wed) by RCL (guest, #63264) [Link] (5 responses)

What's wrong with Qt? It's the sanest C++ widget toolkit available, and probably the most widely used nowadays.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 19:59 UTC (Wed) by drag (guest, #31333) [Link] (4 responses)

Because it's stupid to take something that works and is shipping to customers and destroy it by starting development over from scratch for the sake of a toolkit with better C++ support.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 17:38 UTC (Thu) by jldugger (guest, #57576) [Link] (3 responses)

In my book, when you make short term sacrifices for long term gains, it's an investment.

Qt is a much broader and better lib than GTK, and will likely stay that way. And with QML, it should be possible to write apps quickly in Python and then switch to C++ to get a perf boost if needed without touching the UI. Conversely, getting Qt on Symbian, Meego, Ubuntu OSX and Windows makes it far easier to depend on Qt and customize the UI to each platform.

This logic isn't very Qt dependent, but you have to look at what's more likely to be the common denominator, and Qt seems to be it. Fortunately, there's a lot that can remain, like the Sharing stuff.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 10:22 UTC (Fri) by drag (guest, #31333) [Link] (2 responses)

> In my book, when you make short term sacrifices for long term gains, it's an investment.

Everything is short term. Long term success is nothing but a long string of short term successes.

This 'short term sacrifice' means that Meego is now dead.

I hope that people seriously learn the lesson here. Potential technical excellence in the future is nothing compared to having something that works right now.

QT4 being better then QT3 didn't stop KDE 4 from sucking and QT4 being better then GTK did not save Meego.

Nokia has moved officially to Windows phone 7 and will continue to develop Symbian. Looks like bad decision making is going to kill Nokia, too.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 10:22 UTC (Fri) by drag (guest, #31333) [Link]

I am NOT happy about this.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 15:43 UTC (Fri) by albertoafn (guest, #64225) [Link]

I dont think there is a lesson to learn here. Everybody but nokia chairmen thought this was an awful move (even people working for nokia) and they had to have continued with the maemo branch.

It was already there, it was nicer. It was working already.

I have a nexus one, but Id rather have a n900 that I would had bought if they didn't do that weird movement. n900 was one of the greates phones ive seen. Better than the nexus one. And its a couple of years old. N1 is a yeard old.

From my pov, at this point, the most sensible thing for nokia to do is accept the new king and start shipping nice android phones... As this is not gonna happen, I hope eventually meego goes somewhere, even if it copes with the same share of the market as linux have vs windows... ;)

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 17:27 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

Funny, I see things the exact opposite.

Ok, RPM vs Deb doesn't matter at all
GTK failed so they had to move to something which actually has a shot at appealing to developers who are used to nice stuff on android and iOS (Qt & QML)
Nokia and Intel might not have a clue about working with communities but they are doing well by simply hiring companies which DO have a clue - very smart move.

Now Qt (and qml) are finally mature they can write gui's that actually are on par with android and iOS - and they dump it all for a MS silverlight based platform... Of course, it makes sense. They can stop a lot of R&D and development as well as building their own services platform, saving lots of money - making the company more profitable in the short term. Big bonus for the bosses. Longer term - Nokia dies because it has no unique capabilities and added value anymore, being just another hardware manufacturer. But the current bosses have left the company by that time so why would they care? Their bonus is paid :D

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 21:20 UTC (Wed) by aleXXX (guest, #2742) [Link] (2 responses)

> They should of concentrated entirely on development tools and UI
> environment from _DAY_ONE_ of Meego. That is what people actually
> care about.

That's what they are doing with Qt, much improved development tools and UI environment.

I don't see your point.

Actually, with using Qt the operating system below doesn't matter that much anymore.

Alex

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 0:19 UTC (Thu) by ewan (guest, #5533) [Link]

Actually, with using Qt the operating system below doesn't matter that much anymore.

Actually, it matters quite a lot, unless you restrict yourself to only working through Qt, which is what Nokia have been encouraging people to do in a bizarre and desperate attempt to keep Symbian alive. If you do that, you lose all the advantages that could have come from having a full GNU/Linux/etc. stack because you don't use any of it, and just stick to the platform API, exactly like you would on Android or iOS.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 10:32 UTC (Thu) by svkelley (guest, #37299) [Link]

>Actually, with using Qt the operating system below doesn't matter that much anymore.
>Alex

Only if you are wanting to write simple apps for an established platform. Qt in an embedded context is entirely insufficient. It tends to over-simplify APIs in a rather large way, a legacy of their Qtopia branch in some aspects. If you want to make a platform that vendors can leverage to actually make a device, you need the ability to leverage platform APIs and to rely on them.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 8:22 UTC (Thu) by cmccabe (guest, #60281) [Link] (3 responses)

> They literally threw it all away to start over from scratch.
> Huge waste of time.

Yeah, Nokia seems to have a potentially fatal case of not-invented-here syndrome. It's the same illness that made the OLPC project such a disaster.

And in the few cases they did try to reuse things, they were always things that didn't quite make sense to reuse. Like what is QT doing on a cell phone? Don't get me wrong, QT is a great toolkit-- for developing desktop applications. But the wrong set of assumptions are baked in-- stuff like "the user will have a mouse." Meanwhile, it was deemed very important to write a brand new package manager rather than using dpkg or rpm, because obviously those latter two projects were totally insufficient for the needs of installing Nokia software.

Based on the memo, it sounds like they've more or less abandoned the Symbian platform to its fate. Now they're "burning" MeeGo too. Seems like a death spiral to me.

If the CEO of Nokia believes he can escape the incoming tide of commoditization by moving to Windows Phone 7, he's living in a dream world. Microsoft has its own cloud services that it wants to push, and a huge team of talent software engineers to push them with. Ironically, if they had gone with the BSD-licensed Android, Nokia would have been able to modify the OS and probably somehow force it to use Ovi rather than Gmail, Google Calendar, and so forth. If they are indeed moving to WinPhone 7, Ovi is already dead.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 14:10 UTC (Thu) by pboddie (guest, #50784) [Link] (2 responses)

Yeah, Nokia seems to have a potentially fatal case of not-invented-here syndrome. It's the same illness that made the OLPC project such a disaster.

One would argue that OLPC has been anything but a disaster in the larger scheme of things: the concept of the netbook became more pervasive, despite the best efforts of Microsoft to tax/suppress that kind of product (and of Apple to belittle the concept whilst offering something similar for ten times the price); the screen technology is a lot better than stuff that the LCD cartel is still forcing people to have. And I'd rather people tried out new interfaces than copy stuff like Vista, which is what KDE 4 seemed to be all about.

And in the few cases they did try to reuse things, they were always things that didn't quite make sense to reuse. Like what is QT doing on a cell phone? Don't get me wrong, QT is a great toolkit-- for developing desktop applications. But the wrong set of assumptions are baked in-- stuff like "the user will have a mouse."

I don't know what irrevocable assumptions Qt has, although I believe it had stuff like gesture support a while ago, but from my limited experience playing with PyQt every now and again, the biggest obstacle with Qt is how much stuff the libraries contain and how much pruning has to be done to get the footprint down to something that us "old timers" would feel comfortable putting on an embedded system.

There's also the argument that the various paradigms don't really translate to effective non-desktop interfaces. I played around a bit with the model-view stuff in order to mock up a small-screen interface (which was actually quite convenient given the Qt VNC backend, and as I noted elsewhere, convenient testing during development is crucial), and it got to the point where I'd need to develop a widget to convey the information I needed to, and as with many GUI toolkits, I started to feel that the amount of work required was about to increase significantly (and started to wish I was doing the thing in something ultra-simple like PyGame).

I agree with much of your assessment except...

Ironically, if they had gone with the BSD-licensed Android, Nokia would have been able to modify the OS and probably somehow force it to use Ovi rather than Gmail, Google Calendar, and so forth.

Android is GPL- and Apache 2.0-licensed, depending on which bits you're looking at. And modifying the OS is something they could do under either licence, although I imagine that what you mean is that they could modify things and keep it proprietary under the BSD or, in reality, Apache 2.0 licences.

Of course, if Nokia paid more than lip-service to openness, rather than saying "These bits are closed, do not ask why! Join us!" they'd have their developer ecosystem.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 9:13 UTC (Fri) by cmccabe (guest, #60281) [Link] (1 responses)

> One would argue that OLPC has been anything but a disaster in the larger
> scheme of things: the concept of the netbook became more pervasive,
> despite the best efforts of Microsoft to tax/suppress that kind of product
> (and of Apple to belittle the concept whilst offering something similar
> for ten times the price); the screen technology is a lot better than stuff
> that the LCD cartel is still forcing people to have. And I'd rather people
> tried out new interfaces than copy stuff like Vista, which is what KDE 4
> seemed to be all about.

Around 2008 a lot of things started to converge: Intel released the Atom processor, which they had had in development for several years, Asus started to get interested in creating small laptops, and the term netbook started to gain credence. OLPC (which was arm-based if I remember correctly) probably led people to take the idea of small laptops more seriously, but it's kind of hard to prove cause and effect here.

Unfortunately, what netbooks turned out to be for most people is rather far from OLPC's original vision. Most people run Windows XP on their netbooks and use an x86-based chipset. "Windows in a small form factor" seems to be what netbooks are all about these days. The true innovation in interfaces is taking place in products like Andrdoid's Honeycomb and Apple's iPad.

> There's also the argument that the various paradigms don't really
> translate to effective non-desktop interfaces. I played around a bit with
> the model-view stuff in order to mock up a small-screen interface (which
> was actually quite convenient given the Qt VNC backend, and as I noted
> elsewhere, convenient testing during development is crucial), and it got
> to the point where I'd need to develop a widget to convey the information
> I needed to, and as with many GUI toolkits, I started to feel that the
> amount of work required was about to increase significantly (and started
> to wish I was doing the thing in something ultra-simple like PyGame).

Well, Android also requires you to do a fair amount of up-front work to get things running smoothly and well-integrated with the rest of the system. Proper GUI design is harder than most people think.

The key insight is really that the touch interface changes the game. Nokia really should have known this already. The whole reason why they ever had a viable platform, as opposed to being yet another Windows OEM, was because the interface requirements for a cell phone were so different from those of a desktop computer.

> Android is GPL- and Apache 2.0-licensed, depending on which bits
> you're looking at.

Thanks, you're right. Still, as you also noted, the practical effect seems of BSD vs. Apache 2.0 seems negligible in this case.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 11, 2011 12:28 UTC (Fri) by pboddie (guest, #50784) [Link]

Around 2008 a lot of things started to converge: Intel released the Atom processor, which they had had in development for several years, Asus started to get interested in creating small laptops, and the term netbook started to gain credence. OLPC (which was arm-based if I remember correctly) probably led people to take the idea of small laptops more seriously, but it's kind of hard to prove cause and effect here.

The OLPC uses an AMD Geode CPU which one might argue isn't really as great a choice as ARM for, say, power efficiency, but then the screen is usually a bigger concern in a laptop, hence the interest in a non-backlit viewing mode for the screen (as well as the benefits of the back-reflective screen in bright sunlight).

Most people run Windows XP on their netbooks and use an x86-based chipset.

Most people don't really get a choice about that, however. Indeed, there are various devices out there which would once have been called "subnotebooks" using MIPS processors, such as those made by Lemote and other Taiwanese/Chinese vendors, not to mention stuff from the likes of Sharp which is only available in Japan, but these are kept conveniently outside the mainstream.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 13:56 UTC (Thu) by hitmark (guest, #34609) [Link]

Mostly it was the multiple sudden switchovers that carpet bombed the momentum to death.

First they go N900, complete with a 300 device give away at a conference and a new Maemo version that was barely compatible with the old stuff (the user interfaces did not even line up). Instant community fragmentation. Then came the announcement that they would go Qt, when until then it had been all about GTK. Then came Meego. And finally one get the news that Meego would use RPM, when until that point it had been DEB (Likely pushed by Intel as Moblin went RPM with 2.x, and most of the available Meego releases are basically Moblin with some new graphics. The netbook and IVI versions still use GTK for interface for instance).

This, on top of existing grief related to closed source mixed in with the open stuff (Browser UI, hardware control) that made deep participation from the community difficult, basically alienating the old guard.

And funnily enough, i think having a non-phone device actually helped interest as it allowed people to separate the vital importance of a functioning phone from the inevitable breakage of tinkering. So if one tried to mod something on the non-phone devices and broke something one could reformat and start over, keeping important data on the memory card.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 15:56 UTC (Wed) by ssam (guest, #46587) [Link] (2 responses)

"/me hugs his Maemo -- a Debian derivative that works like any distro, with glibc, Xorg, Gtk etc."

/me hugs my openmoko. a phone open enough that you can install debian, gentoo, or one of several openembedded based systems (or even android) on it. and with GTA04 on the way things are looking up.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 17:45 UTC (Wed) by teknohog (subscriber, #70891) [Link]

You can also install other distros on Maemo devices. I have Gentoo on my N800, though not everything works there. Some drivers are closed, and I agree that Openmoko as a whole is more open. People have also installed other distros on Android devices, for example on the Toshiba AC100.

IMHO, the big point is that Nokia has the only mass-market phone that ships with a regular Linux distro by default.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 18:48 UTC (Wed) by climent (guest, #7232) [Link]

Yep. Only problem? OpenMoko was released in July 2007 and still has only one decide and no market share, which is something that Nokia values more as a company than having a truly open source device. And yet they dont have it either, even though they started with Maemo before 2004.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 9, 2011 19:18 UTC (Wed) by jzbiciak (guest, #5246) [Link] (1 responses)

As a fellow N900 owner (hey, are all 10 of us here now?) I'd have to say Nokia's execution on it sucked. They either needed to truly put the full force of company behind it (in which case it could've been well ahead of Android), or kill it before it became a phone. As a computer geek, I'm grateful that I got one. Nothing about it, however, makes commercial sense to me. I can't recommend it to anyone in my family.

Nokia was too wedded to Symbian. I don't care how much you polish S60. The only people who will appreciate it are existing Symbian users. My last phone was an N82, and quite frankly, it was a far better phone than the N900. Better camera too (despite very similar specs on paper), and it could actually do "modern" phone things such as MMS and voice dial. (Modern by, oh, 2005 standards.) But, being a Symbian S60 device, it had all the quirks and annoyances that came with that, such as each app seeming to have its own private network stack. I don't know how much they might've polished Symbian since I got the N82, but given how far ahead iOS was when it debuted around that same time, it's clear Nokia had lost the plot.

The N900 didn't have to be such a gigantic step back. It's got the same processor as the original Droid, but unrivaled clunkiness as a consumer gadget. Had Nokia put Symbian on life support 4 years ago and put their full force behind Maemo, it could've taken much of the space that Android it. Instead, Nokia got a few geeks interested, hoping they could get us to build their ecosystem for them before it ever had critical mass in the market. Hrm... I don't think that works.

I stick with mine 'cause I like not having to root my phone to get a bash prompt. But unless something seriously awesome happens with MeeGo, I'm going to have to look elsewhere for my next phone.

Nokia drops MeeGo phone before launch (Reuters)

Posted Feb 10, 2011 0:21 UTC (Thu) by ewan (guest, #5533) [Link]

Hear, hear. That sums things up very nicely.