Problem: Compatibility breaks between OS and device releases
 

I'm on a rant here, I'm frustrated and quite upset right now, so take this rant with precautions, and accept advanced apologies if I offend anyone.

Qgil, it seems to me Nokia has a big problem right now: every couple of years Nokia breaks complete compatibility with old devices, and leave developers in the dust. This is why your comparison with laptop market doesn't hold water: almost every Maemo device Nokia puts out need new custom development from day one, and this is why Nokia needs to send out these devices at low low prices- to get enough developers behind the movement to create enough _interesting_ applications in order to have an applications market that would drive N900 sell. If you'd get previously compatibility one wouldn't need to get the latest and most expensive device in order to develop for it, and you wouldn't need to have developer devices sent out at discount prices for every device launch. And it's not just Maemo, it's Symbian too. It's like Nokia shoots itself in the foot everychance they have, and they love it.

So let's say I'm a diehard Nokia fan - my case; what should I develop on now, which device should I buy, on what should I focus ? I bought a 770 for a pile of money, just to have it WSOD'ed in a week, and drop another hundred to get it repaired out of warranty; just a couple of months later, N800 was launched, and I didn't manage to get in the developer program, and I didn't had any more money to spend on it. This year, Nokia launches 5800 with the message "platform is here to stay", so I get one, and start toying around with it, only to discover that again the platform is fractured - N97 isn't quite compatible, developing for both require you have both devices; and now the clear message that Maemo is the future, making v5 Symbian already dead on arrival.

Should I get a N900 then ? Not with my money on the full price for a still buggy device, considering only yesterday fakeshots of N920 appeared, with the message of launching in 9 months with maemo 6 and multitouch, and rumors possible N900 incompatibility - credible rumours if we account for the past.

Why can't we be more like Apple, with its incredible ecosystem and care to compatibility (three generations of the iPhone all capable to run the same latest software !), or like Google with its extreme predicability and openness ( compatibility across spectrum, clearly defined future intentions, huge market ) ?
I think Nokia wants to be in the same league, blurring the line between laptops and mobiles, wanting to have lots of developers committed to the platform, but cant' switch from its mobile-manufacturer mentality that every phone is unique and not necessarily both forward- and backward- compatible across an ecosystem.

So let me ask here: if, as a developer who doesn't have enough karma to get a nice discount for N900, if I put my money down for a full-priced unit, what do I get in return ? Big market - no ? Future prospects - no ? Ecosystem (compatible devices in the future, market place) - probably not? Is it all about the community and prestige ? This is not about me personally, but about the Nokia policies that make me think that if I throw away another several hundreds for a "development platform" in just a quick couple of months I'll have just another paper weight because all the latest focus and drive has moved to another device and platform.

One disgruntled Nokia user,
Rant over.

Re: N900 device program for maemo.org developers and contributors
 

1 - Completely off-topic.
2 - Glad to hear you're having luck running iPhone apps on your 2nd generation iPod.
3 - QQ

Re: N900 device program for maemo.org developers and contributors
 

1. Apple takes 30% of the gross revenue of any app you develop. They will not allow just any app on the device.

Contrast with Nokia: build whatever you want.

2. Google has poisoned their well by unreasonably rejecting applications on their platform.

Contrast with Nokia: upload whatever you want.

In addition;
- Maemo is built on debian. That means you can port many already existing applications to the device.
- Nokia is the #1 cell phone maker in the world with a market share in Smart Phones vastly larger than Apple and Google combined.

I think the wise developer will see that the opportunities here are huge.

Re: N900 device program for maemo.org developers and contributors
 

Not offtopic - it pisses me immensely that people that wrote significant software for N800 don't get in the dev program because their software is made obsolete for N900. If they'd put this kind of energy into iPhone app development they'd made some nice money off it. With N900 they have to but another device if they need to port the software, with still no prospect of getting cash in return. Why would they commit further to this platform. I'm NOT talking about me here !

2 - Nice though you can have a shitty 2-year old iPhone and getting the same benefits (on both sides of the market !) as one who just jumped on the train just yesterday.

What I'm trying to say: I'm not gonna risk a lots of money here and I'll take my game elsewere if I (and others who may feel left out of the program) can't get more transparency from Nokia about future plans and support.

Yesterday, Slashdot had a discussion about how Nokia learns to love and embrace openness. Congrats, Nokia, for not realizing that open souce != openness. Control freak Jobs manages to have a more viable platform than the guys running Linux on their devices!

Rant over, got a meeting to attend to and need to cool off.

Re: N900 device program for maemo.org developers and contributors
 

??? I would argue the contrary, those that had popular software for the N8x0 were those who could build karma. Budding N900 developers are nowhere to be found on the karma map.

Re: N900 device program for maemo.org developers and contributors
 

100% from 0 is 0 : there is no market place for Maemo.

This is why I can't get online radio to play on my devices !

This is why I can't get a basicly decent stock market application.

This is why no _real_ innovation in software happens on Maemo: only ports from desktop. Why I'm pissed off: I just noticed Layar and there is no way to run it on either symbian, maemo, N800 or N900. Talk about innovation.

They have incredible innovation, support and presence. They rejected few shitty applications, news at 11: nobody cares!

Of course, on Maemo I can put anything: what's the killer app for this platform, 'cause I kind'of missed it on garage ?

This is why iPhone owns 60% procent from the mobile applications market, and 50% of
mobile internet traffic. Nokia was in this market before anybody dreamed of iPhone. But I cant' install a symbian application intended for another phone is my life depended on it.

To do what exactly with a device that will be obsolete in 6 months ? This is my rant.

Re: N900 device program for maemo.org developers and contributors
 

Who are you talking about then? If the person wrote significant software for the N800, they would surely have the karma to be included in the program. So...

Anyways, again, your *****ing and moaning is completely off-topic here. Open a new thread if you want to continue telling us how wrong you are.

Re: N900 device program for maemo.org developers and contributors
 

You really wrote that in ultrarant mode I fear. My experience is completely different. In my opinion, everybody who made a medium contribution to the community got in the dev program (just see me with the N810 and the N900).
In addition, you can use the scratchbox environment to develop whatever app you like, then test it remotely on a real device using the service from Forum Nokia.
If you're truly interested in the success and possibilities of the platform, you will understand why the N900 breaks so much - at least in terms of UI. Maemo 5 is for fingers, and this needs fitting UI concepts.
And in the end: the device program offers a discount of how much? 200 EUR perhaps? This completely pales in relation to the hours of work needed for any developer to contribute just a small app. The discount is absolutely insignificant, it's just a nice gesture from Nokia towards the constructive part of the community.

Really, these rants are absolutely unnecessary.

Re: N900 device program for maemo.org developers and contributors
 

Hmm, are you sure it was really intended as a nice gesture? Because indeed for some people it really may be just a gesture and 300EUR is a no go no matter what is the retail price and discount.

Plenty of people here in the forum got N810 (or any previous device in fact) after the price dropped to $150 range.

Re: N900 device program for maemo.org developers and contributors
 

I can't speak for developers but as a user I completely agree. I am very hesitant to invest in an N900 when Nokia has not indicated how long they intend to support the platform.

I have an N810. Fixed in Fremantle may be a joke to some but it is a serious indictment of Nokia in my mind. Apple's support of all iPhones over several hardware and software releases has been mentioned. I would add Microsoft's support of Windows.

WinXP was released in 2001 and was actively supported until this year (and will continue in legacy support for several years). The vast bulk of programs written for WinXP will run without modification or recompilation on regular Win7. For those that won't, Win7 provides an XP compatibility mode.

I realize the environment is different but other companies seem to have a greater concern for their legacy customers and their devices and OSs than Nokia seems to. I suspect it is the fact that Nokia comes from the phone world where the OS and apps were largely device specific.

For whatever reason, the idea that Nokia is developing Maemo 6 without any expressed intent to make it compatible with the N900 is a big factor in my decision whether or not to ditch my current phone and, I assume, a big factor in non-Maemo developers' decisions as to whether or not to develop for Maemo 5 and the N900.

Re: N900 device program for maemo.org developers and contributors
 

A rant is a rant :). It is a personal opinion, and other people realities don't necessarily matches one's reality. If I wrote that, I deem'd it necessary.

About the device program, I'm not sure about N810, but on N800 it was supposed to be an idea contest. The totally obscure way Nokia decided to handle this upsets me: the karma point computation is not transparent, and the criteria based on Karma seems somewhat arbitrary.

About the developer in question, he's not very active on the forum, but his application is in top 30 most downloaded applications for OS2008 - still he doesn't qualify. I think that's just not right, but I'm not gonna name him/her, if they want to come out it will be their choosing to do so. So how's that karma computed again ?

The monetary value is not important here - the recognition is. This is why I'm pissed off. If someone at Nokia wants to check all this, please be my guest.

Re: Problem: Compatibility breaks between OS and device releases
 

Maybe we should just forget this whole Fremantle thing let alone Harmattan, all they bring is tears and sorrow. Let's just use Diablo now and forever. Let's not change anything on it. Everyone can live a happy everafter in an eternally backwards compatible world!

Re: N900 device program for maemo.org developers and contributors
 

Re: Problem: Compatibility breaks between OS and device releases
 

If we would pay for support contracts I'm sure Nokia would put pay more effort in backwards compatibility.

Meanwhile, there is community effort project like Mer. If you appreciate that effort maybe contribute to that.

To ease pain of lack of backwards compatibility knowing in advance the path of compatibility is useful, and these paths are more clear with Maemo 5 and Maemo 6. For Maemo 5 has been pretty clear to me ever since Fanoush posted about Linux kernel changes by Nokia.

There is also retroactive work to get functionality in older devices. While Apple enables FM on their iPhoneOS devices... OK, enjoy. I rather have full Bluetooth profiles, and thank Nokia for providing hardware rendering on unsupported devices, while Jobs forced me to pay 10 USD for iPhoneOS 3.0 only to figure out his definition of Bluetooth support equals 'AD2P profile'.

These iPhoneOS devices have been existing for 2 - 2,5 year, 3 generations, and did not have radical changes, nor have the advantages of open source drivers and such. That is not to say Apple is doing bad in this way; I applaud that they maintain firmware fon different devices, but question is when their support is going to end. Also, I already heard announcements of Leopard being unsupported leaving my Mac unpatched while Microsoft supported XP longer than Tiger. Heck, I still receive security patches formy Windows XP.

Speaking of which IMO the absolute bottom-line is security and reliability fixes. These should be free, for a period longer than 1 year, but I've already posted a few times about that recently.

Re: Problem: Compatibility breaks between OS and device releases
 

I feel that I must also say something to this as I have been seeing many posts like this across the web..

I must say that I can't fully understand the comparison between Nokia smartphones and Apple iPhones backward compatibility. Why?

How many different phones does Apple have?
- 3, as I understand it.
How many categories of phones does Apple have?
- 1,

both of these questions are easy to answer, now, what if I ask you how many categories of phones and how many different phones does Nokia have? Now you may see my point.

With a company as large as Nokia, you just can't expect total or even widespread backwards compatibility as they move forward at a high speed, constantly developing new models in many different categories. As Nokia has said, they have many different phone models that are meant for many different usage situations.

From Wikipedia I found that until March 2009 Apple had sold 21.4 million iPhones. 21.4 million phones that are designed for the same purpose and are basically the same. Of course they have made improvements but still they have to be for the basic parts of the system the same for the sake of backwards compatibility.

As for Maemo, Nokia has openly said that Maemo 5 is a step 4 in a 5 step program. If you demand for backward compatibility then there is no way that step 5 would differ from step 1 too much and were is the progress in there?

As for Symbian devices that Nokia sell, if I remember correctly Nokia last year sold 13 phones per second, thats almost 410 million phones in a year. 410 million phones that are designed for different kind of users and for different kind of usage situations, how on earth are you going to keep compatibility between that amount of different phones?

Ps. Sorry for any possible grammar issues, english is not my native language

Re: Problem: Compatibility breaks between OS and device releases
 

CoreFusion just said most of what I was about to write.

I want to add, though, that starting with the release of Maemo 6, assuming Nokia comes out then and says that "the real Maemo" has truly arrived, I would be disappointed if the OS didn't start providing ongoing compatibility from one generation to the next. At least up until (and including?) the N900, they've pretty much said these were development devices, a sort of hardware/software beta development process. Fine so far; as long as it ends with Maemo 6 and a new era of compatible-over-time hardware and OS releases begins.

(Meantime, I'm sticking with my N800. For a while I was disgruntled that I wouldn't be able to upgrade to the new OS, and that Mer isn't turning out to be that great a replacement for Diablo*; but then I realized that I'm darn satisfied with the N800 running Diablo, it so effectively does so much of what I want.)

*Mer cannot contend as a Diablo replacement without having (or until it has) a good stylus keyboard.

Re: Problem: Compatibility breaks between OS and device releases
 

I have the impression that it usually doesn't take that much work to port an app from Maemo platform to another, mostly GUI rewriting.

Re: Problem: Compatibility breaks between OS and device releases
 

ddalex, if you want to be in the safe side then develop with Qt 4.6 and you will have your compatibility path granted for Harmattan, and probably beyond.

http://labs.trolltech.com/blogs/2009...t-on-the-n900/

Re: Problem: Compatibility breaks between OS and device releases
 

Still not there yet, though it's looking promising:

1. No packages available yet for Maemo 5
2. No instructions on using Qt Creator with a Maemo 5 profile (as demonstrated at the summit
3. No clear path for communicating or raising these issues

Hopefully these'll be addressed in the next couple of weeks, but I don't want people to get false expectations right now.

Still, my first impressions of Qt Creator were promising (C++ is still eugh, though ;-) and I wanna change the key bindings for debug)

Re: Problem: Compatibility breaks between OS and device releases
 

@ddalex: Got to agree with you that API breaks are one of the biggest things that sour developers to a platform (or prevent them from getting started in the first place). On the other hand, supporting an API until the end of time is a sure way to make sure your platform is full of buggy cruft.
That being said, it really is disheartening to see messages on here from developers that did good work in the past, but gave up up developing for Maemo because rewriting their already working programs because of API breaks isn't something they want to do in their spare time.
For my part, I'm willing to give Nokia the benefit of the doubt here, but I'm really hoping that there aren't too many more API breaks ahead of us before they're willing to commit to support an API for the long term across various devices and and Maemo releases.

-John
*goes off to drink more coffee. Please forgive any incoherence.*

Re: Problem: Compatibility breaks between OS and device releases
 

I think they have been pretty upfront about Product x being Step y in a 5 step plan. This entails change -- not everything has been major changes. Some of it has been pretty simple recompile. They have been working on getting the product to where they believe it needs to go. Think of it as Step 1 = Alpha 1, Step 2 = Beta 1, Step 3 = Beta 2, Step 4 = RC, and Step 5 = Final. ;-)

They needed to test the waters; and get community input. (Kinda like the open source model, release often...) Its a bit harder to release new hardware frequently without it costing them a huge amount of money. They have been extremely upfront what is coming for Step 5 --- I believe they have gone way above and beyond what was needed in this. They have seriously have risked taking potentially a lot of sales from the n900 by pre-announcing BEFORE the current device has shipped where they are going. This is pretty much unheard of in most companies, and considered horrible marketting. But, for us Developers this is HUGE, HUGE, HUGE! Large, kudo's for Nokia taking the time to make sure we are ALL aware of where they are going!


I don't disagree with you. I've preached this same line elsewhere. Developers are going to win this platform. But you still have to factor in Nokia and its shareholders. Most companies don't even go for "long" term wins; they go short term. Nokia has already been financing this "direction" for a long time. I believe I read that the 810 was the first device that actually made them any real money.
Part of the issue I believe is they are trying to re-build their core for the next generation -- they have strategically taken some awesome steps but everything is not in place yet. It takes time to get all the ducks in a row. So yes their will be some frustration as they make mistakes and transition. That is the whole purpose of Qt; cross-platform between Symbian and Maemo phones. This is the "missing" link between the lines of phones. Once they get it all working properly; they will be in a awesome postition. But it takes a lot of developers (that they are paying) to get all the libraries fully working to be able to see this vision come true.


Well, if you have been in this business any amount of time; you will know that the day you buy something it is outdated. ;-) Their is a good possibility it that Harmatten won't work. Also equally a good chance it will. I don't think Nokia themselves will even know if it will work until they get to the point of releasing it. They have already started an "official" (Nokia funded) Qt4.6 port for Fremantle in addition to the Qt4.5 team; so they appear to be strongly trying to make sure everything works in Fremantle.

If you want to wait; I'm in the position I have no maemo devices at all (so I develop using the scratchbox/sdk) -- I but I can see where this is going; and I think this is literally the ground floor of a product which will take Nokia a long ways into the next generation and way past Apple, Rim, and Android.


Because they are not quite at their goal for mass consumer devices (step 5) -- I would venture that once Harmatten comes out the API will be very stable (In fact you can target Qt 4.6 right now) and won't change hugely bring the stability you are seeking.


That would be me. I'm standing at 167 Karma. No discount. No device. So I will answer your question. ;-)

Getting on the ground floor of a revolutionary device. Think 8086 -> 80286 -> 80386. The development market went through a lot of upheaval between those generations of chips. 80386 brought the stability to the market.

That depends. It is the #1 presale unlocked phone on Amazon. The dev-discounts don't apply to Amazon -- so this is normal users in the USA only! The device isn't even shipping -- we haven't a clue how popular it will be. (Peter, Quim, any presale numbers available?)

I've been developing for over 20 years. Not a lot of "new" technology really excites me. The Maemo/n900 really excites me. The amount of "future" prospects it offers is unbelievable.

If I had the ability to convince the wife that I "need" a $600 phone -- I would be purchase it in a heart beat. I can only hope that by the next dev discount program that comes around that I can qualify for so that I can much convince my wife that a $300-$400 device is worth while. ;-)

Most apps have really only needed to be recompiled. Some have needed to be re-worked. But I spent a couple nights of work porting several libraries from scratch and then fixing a couple bugs and the app runs really well on my SDK/Scratchbox.

Could be if you get your satisfaction from being known and "prestigious" -- their are several developers on this forum that I am very impressed by because of all the "unpaid"work they put into it. Some have done some amazing things at no cost to the community and cost them time and money.


Their are a lot of people who are still happy their 770/800 and 810. If mine wasn't broken; I would be still happy with my n810 (and would still be lusting over a n900 tho). But you can use the SDK for free and develop pretty much anything in it if you don't want to drop any money. The community will test your app on real devices for you. = No Cost!

I understand your frustration. You feel you have been misled and you want to make sure you aren't anymore. What do you have to insure you aren't being misled this time?

Well, I think in this case Nokia has been _VERY_ upfront about where they are going and have been putting their money where their mouth is by funding not only the Maemo Qt 4.5 branch, but the Qt4.6 branch of code for the Maemo and Qt4.6 for Symbian devices. That isn't even talking about the money they are investing in Qt in general. They are attempting to synchronize their platforms and by this radical of a change they will have accidentally burnt some developers. I think they realize that Qt4 is a/the major underpinning of their next generation and they are not pulling any stops to get it out.

Nathan.

Re: N900 device program for maemo.org developers and contributors
 

Err..

1) How many years has Nokia been in the cell phone market compared to Apple and Google.

2) How many years have they been in the smartphone market compared to Apple and Google?

3) How fast are Apple and Google growing?

Just because Rome was on top before doesn't mean they'll be on top forever. [Though I don't see Apple being on top either, they're the type that would rather own their own market and be content with that since they can never capture the entire market with their tactics].

My concern more with compatibility breaks with the OS and device releases will be how long the n900 is supported. It would be terrible if the lifespan of the device was only one or two years. Especially if Nokia pulls stunts like "fixed in freemantle" if the fix could also be done on the previous OS. But hopefully being built on an open-source community will ensure a long life to re-coup the 600 dollar investment into the device.

Re: N900 device program for maemo.org developers and contributors
 

based on past history it'll be supported by nokia for about a year, you know this, i know this, everyone knows this, its not some hidden fact its how they do business, look at how they support their phones for example


there is nothing wrong with this by me, as i've preordered with this knowledge in mind

Re: Problem: Compatibility breaks between OS and device releases
 

Hi, this spin-off thread belongs more to Development, if someone wants to move it.

The Qt team is working on the packages. They were busy with the Qt Developer Days. Local builds work.

The communication path is described at
http://qt.nokia.com/developer/qt-for-maemo-developers

Re: Problem: Compatibility breaks between OS and device releases
 

Note that this seems to be a write-only mailing list, with perhaps only one subscriber. It has one post, by one author. Me:

http://lists.trolltech.com/pipermail...er/thread.html

Re: N900 device program for maemo.org developers and contributors
 

I would question that "everyone knows this" and I haven't seen any official Nokia pronouncement that Maemo 5 will only be supported for X number of years or that the N900 will not be upgradable to Maemo 6. In fact, the phrase "Maemo 6" exists on only two pages within the site www.nokia.com; both being a press release about Qt. This information may exist but it certainly isn't easily accessible to the consumer.

This isn't a problem for the cell phone market. Cell phones are purchased with the assumption that if you want a newer OS in a year you buy a newer phone. For the smartphone market, this is assumption may change depending upon how backward compatible the new set of OS upgrades are. For the portable computer market, this has never been the assumption.

Whether it will hurt Nokia at the high end remains to be seen. It certainly doesn't help.

Re: Problem: Compatibility breaks between OS and device releases
 

There have been a few commits to Qt Creator about that lately (I didn't know it was demoed at summit :eek:), so I guess the feature is still under development and you may need to pull QT Creator git to use it.