So, its kind of official, Ubuntu is coming to Maemo.
Today, Canonical announced that they are to officially create a ARMv7 version of their most popular software product, Ubuntu Linux, but what does that mean?
There has been a lot of talk lately about a merge of technologies in the Maemo and Ubuntu camps with the 'unofficial' word being that a somewhat arranged, but welcome marriage between the two has been coming for some time. The move today from Canonical signals a clear intention that Canonical, with "demand from device manufacturers" (hello Nokia), are targeting the very popular net book and tablet markets.
The press release is full of thinly veiled statements that hint at the fact that Canonical have had more than a few exchanges with the big boys in the mobile space.
"This is a natural development for Ubuntu, driven by the demand from manufacturers for an ARM technology-based version,” said Jane Silber, COO of Canonical.
Canonical are clearly in collaboration with major manufacturers.
The time frame given for the ARM based Ubuntu release also fits in exactly with Nokia's Maemo 5 plan's. The alpha version of Nokia's new Maemo platform SDK (Software Developer Kit) is due this month, followed by a beta release March/April next year. This will mean a summer release at the earliest for Nokia's flagship Linux device and the statement today firmly aims the Ubuntu release at April 2009, coincidence?
With Nokia's admission that they are making steps to work more upstream, collaborate rather than invent, the writing seems to be on the wall. But is that a such a bad thing?
Nokia will benefit from the huge support that the most popular Linux distribution has gathered over the last few years and having the backing of a major commercial Linux entity such as Canonical can only propel Maemo far more into the mainstream than anything Nokia could do.
"The combination of a commercially supported, optimized Ubuntu distribution for ARM, together with Canonical’s ability to tailor solutions to specific ARM technology-based devices and OEM requirements, ensures that highly-optimized systems can be rapidly deployed into the fast growing mobile computing market".
Jamie & co, it looks like you are making too many assumptions - going really off the track (e.g. Nokia is not involved in this agreement and the Fremantle SDK timing has nothing to do with it). See http://www.internettablettalk.com/forums/showpost.php?p=242071&postcount=25
Quim
Thanks for the comment Quim.
Although the stance from what you have said is one of differentiating from the Ubuntu project I see it as a loss for Maemo. If we continue to duplicate the development effort even in the face of a real viable alternative backed by a huge company, then I see it as a huge waste. There is nothing stopping Maemo being based on Ubuntu's efforts and tailoring it to the needs of the Maemo devices. This is what every other derivative distribution does so why can'y Maemo?
Well, it's not that simple as just trusting that whatever Ubuntu delivers on ARM is going to be exactly what Maemo needs, modify your own baseline in the middle of a release and tell your top management 'don't worry, they always deliver'. Maemo has critical missions precisely in areas perhaps not critical for the "full Ubuntu Desktop": power management, constrained memory, touchscreen support, UI optimized for finger usage, some proprietary technologies and so on.
Later? Well, this is another story. Note though that to solve the equation of "duplication" considering Ubuntu only is not enough. What about Debian or Moblin, for instance. In a perfect world of non-duplication, Ubuntu would be a Debian blend and Moblin would have stayed in the Ubuntu context. However, none of this happened and I guess each player had good reasons to try other way.
Platform development while shipping products while competing with other platforms/products while paying salaries is not an easy equation. I wish the best luck to Canonical/ARM in their first release. All we will know more once the current announcement (and Fremantle, for that matter) turns out in real code.
The "Duplications" comment was written by Quim - and this one too. :)
There are two or even three main problems:
1) Control
2) Time
3) Open vs. Closed
There is currently no way to tell how stable Ubuntu will be and if there are going to be tests Nokia will not have the power to enforce specific versions they tested their own software-stack against.
Then there is the time issue, Ubuntu was just announced (with deals made behind closed doors) but the development of Fremantle is nearing an end already (atleast for some public betas) thus basically switching the OS would cost alot of time.
Finally Ubuntu is more about a real free OS where as Maemo is more about being an open Platform to build on. Integration of the power saving and charging logic (BME etc.) and other little things need some low level work and changes. Here again control is an important issue and Nokia not being the contractor but ARM.
You see there are main aspects which atleast make it harder to "just use" Ubuntu. If Ubuntu was further, with a solid basis when Fremantle was planned (or atleast not as late as now) - there where probably ways to encourage collaboration in some ways and use a "release" of it as basis for the Fremantle Mameo. The dynamics which mondern Linux distributions imply are often not easy to cope with for a big and more or less rigid company as Nokia when they want to provide stable (supportable) end-user products.
First of all it's important to remember the facts:
1) The port of Ubuntu is sponsored by ARM not Nokia
2) Nokia always seemed to struggle with busybox in the upcoming releases
3) It is ONLY for ARMv7
1) and 2) let me doubt that Nokia will use Ubuntu for Fremantle. It would mean a drastic change in direction on a rather short notice, given that not long ago there where discussions about future plans of busybox. Also that the port is sponsored by ARM not Nokia makes it rather unlikely that Nokia will use it (differentiation is still the most important Nokia "assest").
3) Means that there will be no official way to get it running on gPhone, nor the Freerunner nor the N8x0 as ARMv7 is only available in the new Cortex series (as used in OMAP3). Thus it seems more likely to me that it will be a reference platform for the Coretex processor. ARM already provided such things, just not that "popular". And Canonical already started to work in the embedded sphere for the Atom MIDs (until Intel switched to RedHat) and thus has some competence in that area (plus there where already some ARMv5/ARMv6 Ubuntu packages build). So Canonical could reuse quite a few things which probably made the contract "easier".
In the end it is very likely that we will see Ubuntu on the next Maemo device as a community project. That's nice and hopefully Nokia will make it easier this time to get a full system running without fighting against closed source components.
differentiation is still the most important Nokia "assest"
The idea that Nokia software apps and interface are valuable differentiation is a Nokia delusion.
The most important "assets" of the Nokia tablets in terms are differentiation are (1) the quality hardware, (2) the relative openness of the device. I believe most users would prefer a traditional distro stack (e.g. Ubuntu) (with main apps tweaked to suit the small/touch screen) over the mediocre (though tightly integrated) Nokia components (e..g media player, media streamer and chat).
> (e..g media player, media streamer and chat).
As for me, player is just a piece of ... ! I will never buy any Nokia device if there will be no option to replace their player. Fortunately, there is mplayer so n8x0 could be treated as player more seriously. Chat on other side is well-implemented. Quite decent Jabber implementation with group chat and avatars and SIP also works. Not bad for default client. And you have options to use any other client instead. For example, Pidgin (there is also Psi being ported afaik).
Differentiation is one of several generic reasons to have a package closed: http://wiki.maemo.org/Why_the_closed_packages
Look at the components explained and you will find that differentiation doesn't top the list.
The "Please update your records" comment was written by Quim - and this one too. :)
That is for you. Nokia still lives under the assumption (if it's really right is not the point here), that differentation in the software-stack has a very high value (as you see with the many, many little closed parts in the UI).
What Ubuntu does is to push Moblin (even if it is tightly related to Intel AND Maemo) as the main UI with it's own end-user ready interfaces. Nokia on the other hand goes in a slightly different direction with Fremantle with the completely clutter based interface that makes much more use of the 3D Chip (that is btw. NOT part of the Cortex platform but of the OMAP3 platform) and thus get's its differentiation again - hopefully fremantle will bring also a bit more openness in the components.
Thus my deductions and Quim atleast confirms the conclusion itself (Nokia not using Ubuntu).
Still I am sure we will see a community based Ubuntu on the devices with nice optimized packages for the ARMv7 processor in the upcoming Mameo device.
As for me they are overestimating value of differentiation. Some of their "differentiation" is actually plays bad joke with them. Yeah, slow and dumb players are surely different from decent opensource players like mplayer. And such crappy players is what really distinguishes Nokia here and there.
Forgive my ignorance, but will the move to ARM also mean that Ubuntu Mobile will move into the handset distro space? i.e. work on platforms like gPhone hardware and FIC's 'neo' handsets?
> work on platforms like gPhone hardware and FIC's 'neo' handsets?
Uhm, at the end of day I like custom UI made for fingers. Are you sure it is a great idea to get Ubuntu running on OpenMoko or n8X0 devices? And if ARMv7 (Cortex A8) devices are powerful enough, older devices with earlier (less efficient) command sets may provide poor performance. Look, Ubuntu is sometimes slow even on full-featured 2 GHz dual-code desktop. I doubt it will perform greatly on 400MHz ARM which have to care about power consumptions so it can't just always run at full speed. At least, Canonical have to prove that they can release something working and that it will have a decent performance.
Nokia has proven that they can create good debian-based system which is well-optimized for fingers, has polished UI and acceptable performance and this OS quite compatible with software. Small adaptations required for software is GOOD since usual desktop software was NOT designed with small screen sizes and high DPI in mind so using software without tweaking may provide poor experience.
The only drawback of Maemo system for me is closed components. For example, I hate how damn File Manager displays Bluetooth file transfer progress. Sucking progress bar turning modern multi-task OS into a single-task brick more than annoying. At very most I can tweak window manager to disable supermodal dialogs so nasty filemanager no longer able to completely lock up UI and if I'm transferring several files to to my PC or from my PC or from my phone, etc I do not have to suck looking on damned progressbar without ability to use device. That's where I can feel power of OPEN technologies - if you don't like something you can fix this (uhm, please someone in Nokia, kick asses of UI designers for nasty supermodal dialogs [ab]use!)
Unfortunately, I also do not like progress bar window at all and it's still sticky over other windows even if does not blocks windowing system completely. And I can't fix this easily - "thank you" Nokia for making your file manager CLOSED. That's really SUCKS!
I'm pretty sure that:
1) It is VERY EASY to change this damned progress dialog to something less annoying form if you have sources.
2) I'm sure this progress bar annoys not just me and at least half of n8x0 users blames it too when it comes to transferring some big files over bluetooth.
In short, everyone dislikes proprietary software because there is annoyances present and you're denied options to easily fix them and have to reinvent wheels and bicycles just to cope with some small annoyance. That's really sucks. Actually, all Nokia devices have questionable UI with some long-standing usability issues. Some ideas in their UI is good but some dumb issues like complete screen locking during file transfers, etc are really nasty. Why there is need to deny to others any options to fix these annoyances?
I hope Nokia will make things more open in future and will rather prefer to collaborate than to cause headache and annoyance to it's users. I'm pretty sure we can both benefit from this: users will be able to fix most annoying problems and Nokia will get better apps in return.
I can't see why a powerful enough handset couldn't run Ubuntu. There is a port of Android in progress to support the beagle board which is what Nokia are using to produce Maemo 5 until their own hardware is available. Also as we know Android also works on the current crop of tablets so I can't see why that can't go both ways.
Ubuntu on the gPhone, now that's another post :)
Surely it can. But Android is just pretty useless toy centered on running Java craplets and hence users do not have to expect serious apps and decent performance from this platform anyway - it's just a TOY, like J2ME devices. This is just another Java-based dumb dualer ( wtf this crap called "smart" phone at all if it can't run fast binary code like Symbian and WinMobile can?). Architecturally, Android uses Linux only to start up JVM and provide very basic things to it. There is no X, no GTK, no Qt, nothing - you can not write app on something familiar, you have to learn Java in Google's flavour and that's all options you have. So, this is probably most useless Linux I ever seen. There is not much difference between Android and proprietary J2ME devices IMHO. Benefits of Linux are simply unused by Android platform at all.
Ubuntu in contrast is a fully-featured desktop system - and even much more than that. It was never optimized to run on slow hardware and it is fully featured. It has TONS of features and wide variety of libraries. There is GTK and\or Qt, full-weighted X server and composite manager and even some python, perl and other heavy and bloated stuff "by default". This one was never intended to be toy and therefore expects quite powerful hardware and apps are not intended to be mobile toys with two buttons either. Could be a bit too much for small devices... so either they have to abandon their goal to change as few things as possible and minimize bloat or it will be useless, slow and overbloated system barely tolerable on such devices. Let's say, 400-600 MHz CPU with limited RAM amount is not too fast even on native code. Now if we're about to run Java, Python, Perl and other crap - we're about to enjoy by slow system and lagging applications and interpreters\runtimes eating all RAM available.
..and looks like the new tablets are going to be more PC-like again, I presume? I mean.. Ubuntu has its Netbook Remix UI, but this is still basically a menu on top of the normal Linux Desktop.
I also wonder if this means "OS2009" is going to be Ubuntu-based or if there are two different "OSes" for the "N900" - OS2009 and Ubuntu. And if so, Nokia would need to support multi-boot, or would it be an either-or choice between Maemo OS and Ubuntu?
Also, is "April 2009" any indication that the new tablets will ship by then? I mean, it would be kind of pointless to have a complete distribution ready, but no devices to put it onto.
thp
I would like to see Ubuntu on the new Nokia tablet as default. Of course adding a UI on top is up to the discretion of the provider and would be handled independently of the underlying OS.
I doubt the tablet will ship April 2009, only the beta SDK is due around then but that makes a lot of sense. When the OS is finalized from the Ubuntu team, it gives Nokia a few months to get its custom software finished before shipping the final product.
I left wondering, why ?
Why is it so important for you to see Ubuntu running on your Nokia table per default.
Currently it is running Debian and guess what Ubuntu is made of... well... Debian....
So there is not a hole lot of difference... and for devices like the N8xx series it kinda makes
more sense in my mind to go with the stable and well supported Debian releases then with
the more volatile Ubuntu versions. I mean you do not seriously want to see a full blown normal
Linux desktop on your N8xx right ? Cause for the underpinnings i really don't think Debian vs Ubuntu
will matter much. Maemo/Nokia already has to change a lot of things to get everything supported
and get a good mobile look and feel, the short boot up times, extra kernel support, etc... This i kind
means that the biggest benefit of what Ubuntu did for Debian is gone in Maemo. Maemo does not
need a better self configuring Xorg nor does it need all the other desktop usability enhancements.
I love that Ubuntu is coming to the ARM, i think it's a good move and that many could benefit from it.
I just don't see why Maemo and the N8xx would benefit from it.
Just give me more onboard memory (memory is cheap right?), onboard 3d/cellular connectivity
(everybody wants that right? and isn't this in the making?), a *much* better camera (and turnable so
i show things other then myself on video calls), multi-touch touchscreen support.
And last but not least some more cooperation with the Canola people for an Alternative desktop environment,
i say alternative here because i would like to switch between the current style UI and a super-fancy UI.
Good to see that the mutterings from various companies at OSiM are bearing fruit.
I wonder if this press release is one of the blockers for the release of the Fremantle SDK?
Cheers,
Jaffa
I think that is exactly why.
The time frame for the Maemo 5 release fits in perfectly with the Ubuntu one. Nokia holding back the Alpha SDK until this announcement? Beta out when Ubuntu finalize their OS?