There is one wrong assumption in this discussion, that there won't be a Java VM under Meego.
Who knows?

Actually, the wrong assumption in this thread is that Nokia are interested in hosting apps for other handset manufacturers in Ovi. The MeeGo community repositories will host packages for each supported architecture. Nokia Ovi will host a repo specific for their devices. From an ISVs standpoint the benifits of using MeeGo are not being tied to Google app store, each vendor can have their own app store. This will be very liberating for vendors and will give them their power back.

But isn't that "old mobile"? Since the power of the phone's OS is partially its applications (like iphone's and android's), isn't that assumption a step back?

Anyway. What we really need here is a statement from Nokia people that have insight of what's happening there. The thread is not meant to be a good or bad Nokia/MeeGo/Intel/Whatever. It's just a question: Did Nokia/MeeGo/Intel thought about this? And if so, what's they're standing?

Yes, they've thought about it. Qt is their platform, write once deploy everywhere.

You may think having a single app store like google is the future. But given the opportunity any ISV would want their own app store, like apple, like rim and like Nokia. It's a no brainer income generator.

Android supports multiple non-official stores (read: APK repos). There is no requirement to use the Google Android Market to run Android.

Currently it seems that the Google Market is the largest by a wide margin, but the times they are a-changin'. Amazon is launching a Android specific app store with an interesting sales model. , Samsung, Verizon, etc, etc, also have very small stores. There are also stores unaffiliated with Android or devices like AndAppStore, App Brain, and Appslib.

I make no assertions in regard to said store's quality, but they do exist. Bottom line: Android OS is not irrevocably tied to the Android Market, though it may appear that way based on popular distribution.

The APK container format, Dalvik, and well thought out Android SDK make many apps extremely portable. I expect more stores to pop up with competing propositions, as well as a general quality to improvement given time.

I didn't know that, I presumed the store was just another Google service handsets were locked to, well that's good to know.

Honestly, this is not a problem, unless you're talking about software applications that no longer have support. If I'm a vendor, and a new hardware platform comes along whose only difference from my current platforms is a new CPU (upon which the old OS is running), I'll have no trouble rebuilding the app for that platform. If you're talking about an app whose designer has moved on to other things, well yeah, that orphaned product won't run on the new device, but that's kinda how it works for computers everywhere...

Ovi supports multiple platforms today! Search for Angry Birds; you'll be able to find it for both Symbian and the n900, and I'm sure those are not using the very same binary...

I Can see V13's point very clearly, As a former N900 owner, and a Former Maemo Qt developer, Now an android owner and developer,
I can already feel how much easier maintaining an android app is - compared to a C++/Qt application, and also how much more rewarding it is.

The Goal (as I understand V13 is thinking about it) - Give developers easy tools, to develop, distribute, maintain and sell applications, while rewarding them properly with income to keep them coming once more.
Once this goal is achieved, in my opinion, an OS which aims to run on multiple brands and architectures, has a good fighting chance in todays market.

Standing between that goal and the current situation are multiple aspects, let us analyze them.

1. Different form factor / onboard peripherals
2. Different CPU architecture
3. Different distribution methods.

The problem of having to take into consideration multiple form factors and added features, in the essence of peripherals, is a common one to all modern smartphone OSs. Android / Meego / And surprisingly, even apple, have this problem.
The developer must write his app to use the correct hardware, if available, and the UI needs to be able to adapt to screen size / resolution.

Different distribution methods is where so far, the N900 has been failing at, and we just don't know yet if MeeGo will address this problem with an adequate solution.
There is no doubt that the ease in which the Android market and Apple app-store allow developers to upload an application which will intern be able to be executed and used correctly on a very big quantity of end-user devices, is appealing towards developers.


As for CPU architecture :
Personally as a developer, I don't want to have to test and own 5 different devices just to be able to check if my compiled binary file really runs on all different 5 architectures. I don't have time to do it, and don't have the money to do it.
If the development IDE / SDK, allows me to select from a list of checkboxes, to which architectures I'd like to compile, and automatically upload all binaries to Ovi store, and this will really work, then we don't have a problem.
Currently in C++/Qt developers have to use #ifndef's to include/exclude hardware / distribution specific code to make things work both on Maemo and S60 and MeeGo.
If this will stop when meego is around, by supplying all the required tools via an API, this is again one step closer to allowing a batch compile to multiple architectures.

Finally, the whole thing should be easy to perform and require little time. Fast penetration with a large market of applications is necessary to succeed.

the reason for non JavaVM is QT(especially QtQuick) please read more on QT before all bashing about non javaVM. There is no reason to have it! writing more advanced games etc == you have to learn c++ thats the case in android too!

And about QtQuick: QtComponents(written in QML/javascript) will replace todays Libmeegotouch UI in meego that was written in C++ and now most done in QML language. also it seems symbian will use it (=arm cpu:s).

Latest release of QTCreator has already support for different targets: arm/x86 and even different os:es like Linux, Windows, Macosx, Maemo/Meego, Winmobile...

And about the progress in this areas: Public repository avaible at gitorius.

Except from one case: Future devices. Android apps are guaranteed to run on future devices as long as they run a compatible OS version and most probably they will run faster and take advantage of the new hardware.

But when using MeeGo, who will dare (or care) to introduce a new architecture or improved hardware? It's not only a matter of architecture. Even an improved instruction set or extra registers will be useless if you have the same binaries, just like it happens with Debian (all programs are compiled for ridiculously old hardware).