Hi,

Funny I missed this thread, even though I'd seen TexRat's blog post...

...yet...

The story for me is (and has been for a while) that there isn't a clear story for developers on the platform. This can be traced back to the Trolltech purchase at the start of 2008 - which was around the point where people started wondering about Qt's role in Maemo.

Let's track the timeline from January 2008 (pre-Diablo) through to today:

1. Jan 2008: Nokia buys Trolltech - no news on plans for Qt in Maemo, but people say "Nothing is changing"
2. Apr 2008: Nokia announces that Qt will be supported for Maemo in addition to GTK+/Hildon
3. September 2008: At the Maemo Summit, Nokia gives some more information on Qt plans in Maemo: Qt will be available but unsupported for Diablo and Fremantle, supported on Harmattan, and then after that "we don't know yet" (but it is obvious that Qt will become primary platform on Harmattan+1)
4. July 2009: During Gran Canaria Desktop Summit, it's announced that Harmattan will be Qt based, and GTK+/Hildon will be "community supported"
5. Sept 09: Fremantle final released with the N900, GTK+/Clutter interface and new UX guidelines, means most Maemo 4 applications need porting to Fremantle
6. Jan 10: MeeGo announced - MeeGo 1.0 will be a merger of Maemo 6 and moblin 2.1, with Qt as the primary UI development framework, connman/oFono/bluez as core telephony & connectivity stack - a common OS base for multiple user experiences
7. Apr/May 10: MeeGo Netbook 1.0 released, core applications include Banshee, Evolution, core UI is GTK+/Clutter based. Harmattan will be Qt based, but will use Debian package management, but will still be called MeeGo Handset 1.0.

Now, as an application developer, in July 2009, just as Fremantle is ramping up to get out the door and wow people on a Maemo phone, you've been told that your existing application will require porting, and that the platform you're porting to will be obsoleted by the following version. Qt 4.5 had some hacks to get it working on Maemo with qt4-hildon, but Qt 4.6 is the first fully supported version of standard Qt for Maemo, and it only came out last month on Maemo 5 with PR 1.2.

So what do you do between July 2009 and May 2010? Do you migrate your application to Fremantle? Try to re-develop it in Qt (or write a new app in Qt)? If so, Qt4-Hildon or standard Qt? None of these are trivial decisions - and importantly none of them are making your application better (which, as an application developer, is what you want to be doing).

And what is this inconsistency between the public messaging of January when MeeGo was going to be the all-singing, all-dancing stack you could build anything on, and now we see that there is some sort of common bits, covered by this hodge-podge of graphical parts which are using all sorts of toolkits & technologies, probably present because of existing partnership agreements (eg. Evolution Extress & Banshee), shoe-horned together and called MeeGo?

What of the compromises and technology choices that are going on right now unseen in the MeeGo project? The move away from NetworkManager to ConnMan for Novell's MeeGo, imposed by Intel. The requirement to move away from the (very cool) Clutter windowing system to some new & untested Qt replacement for it. The move to oFono for Nokia, leaving their tried & tested communications framework in the bin. It seems like key players are fighting to have their little bit of the world present in the required platform, resulting in a cluster**** which will not resemble any other Linux distribution, and which will once again not provide a compelling developer experience.

Yes, I am getting fatigued by the evolution of the project, and I have yet to see any real hooks to allow people to get involved in MeeGo - on the contrary, I have seen a good amount of stop energy in the form of "it's too early for that" and "that's all legacy stuff irrelevant to the platform" which is killing any good will left.

I may be wrong - maybe when MeeGo handset comes out it will be gorgeous and compelling and we will all be dreaming of millions of MeeGo phones, and application developers will come flooding in. I'd love to see Hildon ported to MeeGo and let all those Maemo application developers have their apps work well unchanged. I don't really think either is going to happen. It feels like we're back in 2006 again, starting afresh, and there has been a deliberate strategy to have Maemo be "the one we threw away", to use the phrase from the Mythical Man-Month.

Dave.