Is Freemantle dead before it's out of the gates?
 

So I was listening to the latest episode of The Linux Actionshow podcast today (Season 10, episode 9) and even though I am far from a Linux expert I found the following interesting e I want to see if I was even close to understand what was said :)

Based on their discussion, about the NITs, they said that Freemantle is the operating layer of the next NIT - but after that it will not be continued as after that Nokia is continuing with QT.

Does that mean that Freemantle in essence is being developed to run on N900 - and that's it?

Re: Is Freemantle dead before it's out of the gates?
 

to my understanding the underlaying libraries won't be changed, so I wouldn't call it DOA.

Re: Is Freemantle dead before it's out of the gates?
 

I think you're right but my understanding is that...uh..correct me if I'm wrong..is that Freemantle is not really an OS in the traditional sense but rather a presentation-layer running ontop of GTK developed apps.

What I gathered from listening to the show is that N900 is the only device that will use this presentation-layer (in much lack of a better word) and that Nokia is moving to QT and a different presentation layer.

I do believe though, to make things even more complicated [for me] that items developed in GTK or QT will still run on the later devices and 'all' that is different is that Freemantle will not be there.

Am I kinda-sorta understanding it correctly?

Note: This is from someone who two months ago thought that Linux was the name of a Penguin.

Re: Is Freemantle dead before it's out of the gates?
 

Good thing so many things will be fixed in Fremantle.

Re: Is Freemantle dead before it's out of the gates?
 

Drink!

[message to short]

Re: Is Freemantle dead before it's out of the gates?
 

Thanks for the pointer!

To answer your question: no. Fremantle brings a lot of new stuff that needs to run in real devices and needs to improve with the feedback of real users and developers. The success of Fremantle is needed for the success of Harmattan.

We have started sharing our Harmattan roadmap soon to clarify a couple of questions that were coming frequently: what is the toolkit strategy and what is the relation with Symbian.

It's also a good practice in open source platforms. Now the Qt path can work with clearer goals in mind, knowing that Maemo is going to be a platform with good support and a significant investment in innovation. And the GTK+ path can start considering what future they want for the Hildon contribution and what is the support level they want to achieve in Fremantle. Since Fremantle hasn't even seen the light as final version there is time to discuss and get organized.

You might be interested reading and listening the original source:

Maemo Harmattan: Qt and more

There is also another thread where this topic has been discussed:

Quim keynote on Maemo's switch to Qt as the main toolkit

If you have further questions please ask. Perhaps better in the thread above.

Re: Is Freemantle dead before it's out of the gates?
 

Fremantle (Maemo 5) is built on later versions of the same libraries which have been used in Maemo since the 770 was released in November 2005. This includes gstreamer, DBUS, gconf and the Gtk+ widget toolkit. Several UI features in Fremantle will use Clutter (for the first time) to give whizzy dynamic effects. In addition, we'll start seeing more third party apps written using a different GUI toolkit: Qt. They'll still be running on the same OS, though.

Harmattan (presumably Maemo 6) is built on many of the same libraries as Fremantle. However, the roles of Gtk+ and Qt are reversed: the core of the user interface will be written using Qt; though third party apps will still be able to use Gtk+. Similarly, Qt has a technology similar to Clutter for doing fancy transitions - necessary to have a compelling user experience in this day and age.

More information, including links back into the appropriate thread here can be found in this LWN feature article on the subject:

http://lwn.net/Articles/341391/

Re: Is Freemantle dead before it's out of the gates?
 

Fremantle is version 5.0 of Maemo. Maemo, is a wide range of different services, APIs and it is a OS in traditional sense. On a standard Ubuntu, you have a lot of different services, APIs, etc as well, and you put them on top of those, X and GTK+, and you run GNOME environment on top of this.

In Harmattan one of the aims is to replace GTK+ plus Hildon with Qt. Many of the same services and APIs will be there, but the default widget toolkit will change. - which probably means that Nokia will rewrite most things facing the user (UI, etc) on top of Qt instead of Gtk/Hildon.

What does this mean for a typical extras developer? Not much, - he can still develop his Gtk+/Hildon application, put it in extras - because Gtk+/Hildon, in Harmattan, would be on maemo.org extras. Same way some people extend the platform with custom libraries - by putting them in extras.

The difference is that Nokia is saying in Harmattan, - Qt is the toolkit we support/develop/etc. And I personally think that's a great choice to make. Targetting Qt specifically instead of a weird combo of Gtk+, with Hildon widgets on top, can only be a benefit - and allow us to put the Maemo platform (system daemons, APIs, etc) and our applications on a lot of different things - Maemo GTK+ is a bad mess for portability.

But, since Maemo is flexible, you can still develop your Gtk+/Hildon applications by simply depending on gtk+/Hildon libraries as usual and putting your application in Extras. The difference would be that gtk+/Hildon is now not part of typically installed SDK, but '3rd-party' and hosted in extras. This is how developers use Qt right now.

Does this mean that Fremantle is DOA? For many Nokia provided applications (user facing ones) - their time has maybe come. For all the new APIs introduced and existing ones, fixed bugs in system daemon A, B, C? No! They'll probably still be there in Harmattan (excepting maybe Clutter, but some may argue that this isn't a loss.).

So, what does it mean that GTK+/Hildon will be "community supported"? What does community want this to mean?

To me, it means that 'community' - our developers are still using these APIs in our applications. We can help out fitting Gtk+/Hildon to eventual Harmattan UI changes. This is already being done with Qt - fitting Qt towards looking/feeling like Hildon. And community is involved in this effort already. So, as long as there's an interest from community to run Hildon API applications on Harmattan, it would be possible to help out and make it happen. And it's probably easier than the Mer effort.

Re: Is Freemantle dead before it's out of the gates?
 

1:52:58! Any pointer (even if approximate) for the Maemo part?

Re: Is Freemantle dead before it's out of the gates?
 

A single use (once per device) OS with the majority of it's UI/presentational layer being replaced and those prior parts being deprecated almost as fast as it comes out spells out (to me) that Fremantle will be as stillborn as Diablo.

The kind ways of stating "no, it'll be continued to be supported by the community" yields no comfort from a future purchase because thrice over people have been limited to a Nokia device that has no clear method to upgrade to the next version using their re-appropriated machine due to either hardware or software limitations.

In this case, it's both. The feelgood speeches cannot deny that wholly. Fremantle will fix current issues in Diablo, require new hardware, allow the community to inherit some of it's architecture, while basically becoming a deadend itself shortly thereafter.

Re: Is Freemantle dead before it's out of the gates?
 

Which is why the effort - and results - of Nokia and the Mer team are to be lauded in providing a continued mechanism for supporting out-of-date hardware in a commercially realistic way.

Besides, we don't yet know if Harmattan is going to be targetted at RX-51 style hardware; which I guess will depend on the timescales.

Re: Is Freemantle dead before it's out of the gates?
 

The N810 and Fremantle/Maemo discussion starts at 24:15 on the video version (see link above).

The Ars Technica article they are referring to (and said is not 100% accurate) is at: http://arstechnica.com/open-source/n...o-platform.ars

(wish I had seen that before posting here as it does, along with the great answers here, address my question really well).

Edit: This episode of the show also discusses Google's Chrome OS and if by introducing that Google is 'admitting' that Android was a failure - or if it is as simple as Mobile=Android and Netbook=Chrome.

Thanks!

Re: Is Freemantle dead before it's out of the gates?
 

I agree with you on the risk of parts of it becoming a deadend shortly thereafter. Which is one of the reasons Mer is actually 'MaEmo Reconstructed", and not "Fremantle backport to N8x0". While a backport may have been easier, such a backport would be unsustainable and would become a deadend shortly after as more developers move on to newer devices. So, we decided to take the difficult road - which does take longer time but will be more interesting over time.

Re: Is Freemantle dead before it's out of the gates?
 

Cool! Just like Microsoft only 5 years later. :rolleyes:

Windows Mobile 2003 Second Edition

***

So... for a non-developer, plain ol' consumer, real world user of an existing Nokia iT... purchasing the presumably much more expensive Freemantle device is not a very good upgrade strategy.

BTW, what is a good upgrade strategy for an existing NIT user?

Re: Is Freemantle dead before it's out of the gates?
 

That does make a whole lot of sense. Thanks.

Re: Is Freemantle dead before it's out of the gates?
 

That way of viewing things is rather interesting to me as I am one of those people that are considering getting an N900 when it comes out and if it is well received. One of the things I do love about Linux and Maemo in particular as that is the only distro I've been exposed to - is the community support and involvement.

If the N900 would come with a presentational layer that is in lack of better words dead-ended in that it will not continue onto other h/w devices I must admit that I, as a novice, would feel that I would be buying into a device that will soon lack support and s/w updates as focus moves on.

Even a community as active as this can't possibly juggle updates and ports between, what, five?, OS versions (Diablo/Chinook/Mer/Fremantle/Harmattan).

I guess some would phase out - but aren't people still using Chinook and maybe there will still be users clinging on to Diablo after Mer comes out. It still feels like efforts would/could be spread very thin.

I know that my views are simplistic but I also think that I may represent a part of the potential customer base for the N900 and I'm curious as to how Nokia is/will be addressing such concerns).

Re: Is Freemantle dead before it's out of the gates?
 

And I'll state this until I'm blue in the face. Your (the Mer team) efforts are nothing short of amazing, made possible by Nokia getting the right stuff to you all... and I thank you for that.

Trust me. Without the community and the Mer Team; I don't know what I'd do other than fully lose faith. I'm waiting... and that's fine. And where I can help - which again isn't much - I will.

Thank you.

Re: Is Freemantle dead before it's out of the gates?
 

Plain ol' consumers, real world users will start thinking their purchasing strategy when they see a new device announced with a price tag attached. Discussing Harmattan with Diablo in your hands is just too abstract for end users.

This was an announcement done to platform developers primarily, also to application developers. The basic conclusions a user could get is that Nokia has a mid term strategy for Maemo that is based on certain technologies that would make it easier to have applications available also in Symbian and other mobile platforms. If such users are concerned about open source, the announcement gives them an idea that Maemo will keep being an open platform with open devices welcoming all kinds of toolkits, bindings, languages and alternative platforms.

Re: Is Freemantle dead before it's out of the gates?
 

And my views are seen as largely unpopular. No worries about how you view things... I'm just quite sure that Nokia will not address anything further than they have in the past.

Fixed in <enter name of upcoming build>

At least Mer gives me a bit of hope.

Re: Is Freemantle dead before it's out of the gates?
 

Surely Fremantle will be as DOA as diablo is now.
Isn't the whole purpose of each release to essentially EOL the previous one, and build on it's successes?

Personally I'd like to see a tablet orientated kde4, which I am hoping this is essentially what Harmatten will be. (from a user point of view not a developer)

Re: Is Freemantle dead before it's out of the gates?
 

Nokia will switch from Gtk to Qt, but this doesn't mean that Gtk will be unusable. It's at a very good stage at the moment, surely better than what is Qt for now, since they're still working on the Maemo/Qt.

Gtk will be only community-supported like, for example, is Python support now. This doesn't mean you cannot use it to develop something interesting, look for example the grat job is doing that guy with gPodder or the Canola team.

I personally do prefear Qt toolkit, but it's just my own opinion, and I'm very happy about this switch.

I think that the next N*** will be a good product ;)

Well I cannot be sure about this, but.... I trust them :)

Re: Is Freemantle dead before it's out of the gates?
 

yes, but reviews count for a great deal in an over-populated market sector (phones) and reviewers will point out that the UI/OS is basically redundant. "Consumers" aren't stupid, and in hard times they tend to get even smarter.

Re: Is Freemantle dead before it's out of the gates?
 

Don't know if there is universal "good upgrade strategy". I can only say what I will do: everything depends on price *and* if N900 hardware will be supported by Harmattan.

If starting price > $400 or hardware future not clear I will skip and wait for Harmattan. My aging N800 is still good enough for most of my uses.

Re: Is Freemantle dead before it's out of the gates?
 

Why don't you say what you really think. :p

What interests me is how much of this full-on switch of loyalty (for the lack of better word) is down to Nokia's desperation to replace the much-maligned Symbian mobile phone platform as fast as they possibly can and their takeover of Qt software, and how much is down to Nokia's lack of control or lack of faith in the direction and abilities of GTK and Hildon?

I am not arguing the decision to switch per se; it must've been made at least a year or two before Maemo users and ordinary developers were told about it... Ergo, I hope Nokia succeeds while managing to keep the new system open enough.

Re: Is Freemantle dead before it's out of the gates?
 

Commercially realistic?

a) Most if not all of the commercial developers are porting their apps (Skype, Gizmo, Flash etc.) to Fremantle and those apps will be compatible with Mer as well?

b) Some of the commercial apps will be ported to Fremantle and they will be compatible with Mer as well, the rest will still work despite linking to old Chinook/Diablo-era libraries?

c) All or some of the currently available (but generally aging) commercial apps will continue working under Mer thanks to compatibility layers with the old Chinook/Diablo platform, but the commercial developers will have nothing to do with Mer compatibility?

d) ??


I have a feeling that the majority of N8x0 users will stick with whatever was installed in their tablet or what they can flash on it using the Official Update Applet; i.e Chinook or Diablo. Most MS-Windows users certainly, and probably many Linux users too.

Re: Is Freemantle dead before it's out of the gates?
 

Doubt that; Gnome and KDE not only differ in toolkit, and while switching toolkits is something you can force yourself to, switching philosophies is not.

I personally expect we're going to see what Mark Shuttleworth's once dreamt: QT/Gnome (Mobile).

Re: Is Freemantle dead before it's out of the gates?
 

Reviewers are actually not that different from users: they pay attention specially when there is some product with known specs and price tag - something to look at and play with.

Let's see what they say when the announcements are made.

Of course consumers aren't stupid! Even less the potential Maemo users, quite demanding and well informed customers - with plenty of choices at hand. We are working to have something good for them (for you).

The UI/OS is not redundant: is evolving. One day Maemo 5 final will be launched with a device, and this will be the best OS/UI and the best device Maemo will have done. Of course we keep working on further releases and products that should bring better OS/UI and better devices, better developer offering, better compability with other platforms etc.

Also note that you are mixing 2 things: change of toolkit with hardware incompatibilities. Fremantle is not officially supported in OMAP2 devices (N810m N800) because we concluded that Maemo 5 was too much for that hardware if commercially supported. In comparison, the GTK+ and Qt toolkits are very similar in terms of hardware demands. They sit too high in the platform and, as we are insisting, the middleware underneath stays more or less the same, of course with significant version upgrades and enhancements.

So the discussion of Harmattan running or not on the next device has actually little to do with the toolkit officially supported and more on other aspects like what features will Harmattan bring, demanding what hardware capabilities and at what point of time.

Discussing this paragraph above when there is not even a Maemo 5 device announced is an exercise of pure Summer entertainment. (which is fine for the Summer, don't get me wrong) :)

Re: Is Freemantle dead before it's out of the gates?
 

Historically Qt has proven to be faster than GTK+ on slow and limited hardware, as example see the former Qtopia distro in OpenMoko. But of course a lot of people disagree with that point :D

Re: Is Freemantle dead before it's out of the gates?
 

i.e. it's commercially unrealistic to expect Nokia to constrain future devices' software by spending lots on supporting previous generation hardware.

However, they can garner goodwill (which translates into potential future sales) by assisting with the community taking on the work of maintaining - and enhancing - previous devices' OSes.

Re: Is Freemantle dead before it's out of the gates?
 

thanks for taking the time to reply qgil :)

that's the day we're all waiting for!

Re: Is Freemantle dead before it's out of the gates?
 

After reading this thread, I can only summarize what has already been said (sort of). This is really a developer discussion and could never be a user discussion. Sure, there will be a few users who are interested in all of this, but mostly, users are typically happy with their current device for quite some time -- and never come to maemo.org.

Being brought up on major mobile company contracts, most regular users expect to be with a specific device for at least two years -- and most users keep their devices longer. This changes when a new trend arises (RAZR, Blackberry Pearl, iPhone, etc.), but usually changes only occur when contracts are up.

But, that's hardware... When concerning software, it's a little different, but not much. I still know many PC users who are still running Win98 and many Mac users who still use OSX Tiger. They don't care either. (And, I dare you to show me a person who's really -- successfully and at an acceptable speed -- running the latest version of OSX on their old beige G3!)

So, thinking that all of this is a big deal for end-users (the larger market of end-users, not just us) is wrong.

As for developers... Well, from what I've seen, the Maemo Community are all a pretty smart bunch of people. And, being that you will still be able to use your beloved programming languages, I don't forsee any major issues in upgrading releases for new toolkits.

Tim

Re: Is Freemantle dead before it's out of the gates?
 

Right you are timsamoff.

Although I've only been a small part of this community and can't contribute much more than advice on getting certain programs to work properly via setting changes and workarounds, I still love to see all the various things happening with Mer and Fremantle/Harmattan. Although Diablo is no longer supported, I know the Maemo Team and Nokia is helping us as much as they can by lending a hand to the folks working on Mer, which will be a great OS.

Many of us end-users/inbetweeners, including myself, don't usually ever know what's going on behind the scenes, but the community is there to relay what is going on if the information is available.

We're waiting Nokia and Maemo Team and I know you will deliver. :D

Jesse~

Re: Is Freemantle dead before it's out of the gates?
 

I agree with timsamoff about it being primarily a developer problem.

As a developer, I had the similar questions about Fremantle/Harmattan. I'm just now at the point of starting development, and was trying to decide just which platform to target. It seems that at this point I have 3 options: non-Clutter, GTK for Mer; Clutter/GTK for Fremantle, or QT in anticipation of Harmattan. (Right now I am leaning towards sticking with non-Clutter GTK to target primarily (and maybe exclusively) Mer, because nothing that I've seen discussed so far seems compelling enough to purchase the upcoming Fremantle hardware). I'm waiting to make a final decision until more about the Fremantle devices is known.

While I am appreciative of Nokia's new attempts to keep the developer community informed, I somewhat think that they should have waited to even mention Harmattan until the actual Freemantle devices were out. Because after all, aren't we realistically looking at a 2-3 year time frame before a Harmattan device is even produced?

Cayle

Re: Is Freemantle dead before it's out of the gates?
 

I think the last day of "Summer" in my part of the world is September 7th...so is this a hint or is my imagination merely starved for an announcement and thus delusional?

More about Summer...

Re: Is Freemantle dead before it's out of the gates?
 

It still fits nicely to the Nokia World in September 2-3 that has been already rumoured for some time to be rather big event this year for Nokia and announcements for many new devices including Rover.

Re: Is Freemantle dead before it's out of the gates?
 

@cayle

I think Harmattan is planned for next year, not 2 - 3 years from now. I would imagine it will be an update for Freemantle devices a year after they're out. I couldn't imagine Harmattan not running on the Freemantle devices given they will be able to run Qt 4.5 when they come out via community support.

Personally the application I'm planning on writing will be Qt based given that it has the longest know path.

Re: Is Freemantle dead before it's out of the gates?
 

You can't draw any conclusions from that. Diablo devices are able to run Qt 4.5 right now and we know they won't even run Fremantle, let alone Harmattan.

Re: Is Freemantle dead before it's out of the gates?
 

Sure we can, there's a reason Nokia decided against backporting Diablo, and it has nothing to do with Qt: hardware. Based on that we can extrapolate most of the necessary information about that to determine Harmattan's likelihood of running on Fremantle hardware.

So, based on its likely timeframe (1-1.5 years away), and the likely hardware landscape at that time (OMAP4 not likely to be available), and what we know about current and future hardware (OMAP4 does not offer additional hardware features over OMAP3 that would greatly increase the difficulty of any backport), we come to the conclusion that running Harmattan on Fremantle devices should offer little to no difficulty.

Re: Is Freemantle dead before it's out of the gates?
 

I agree you can't draw conclusions, but I think the N8XX is a different issue. I imagine if Nokia really wanted to, Fremantle could have been supported on the N8XX (assuming accelerated graphics drivers were released). But from a business sense it didn't make sense from a cost perspective (development, testing, support, maintenance). Yes it would have made people here happy, but would do little harm otherwise. Instead they could invest most of their time and effort on the "next" device that has more breathing room hardware wise.

From an end user standpoint Nokia could upgrade from Freemantle to Harmattan and the "casual" user might not even know.

The fact Nokia has been able to keep the N9XX hardware secret this long probably means they wont release a device that has a 1 year software lifespan. Also that's not saying their won't be a N9XX+ that doesn't come with Harmattan installed with new features.

Re: Is Freemantle dead before it's out of the gates?
 

I don't believe we can extrapolate anything at the moment. Very little information about Harmattan is public, and none at all about the hardware it will run on.

These are pure speculation and/or wishful thinking currently. It could just as easily be 2+ years away like Fremantle, with one or more hardware iterations in between like the N810 and N810WE, and it could be based on a non-TI or even non-ARM CPU.

Even if there are no hardware or software surprises and that turns out to be true, it could just as easily be another case of "we can't devote resources to maintaining RX-[57]1, use Mer".

I'm not saying that Harmattan won't run on Fremantle-generation devices, just that right now we can't know - and I suspect even Nokia themselves don't.