Submitted by OpenMoko, the project that creates not only a Linux operating system for phones but the hardware to run it as well, has announced some major changes to its software stack. Traditionally a Gtk+ endeavour, this is all going to change rather drastically.The main changes are rather drastic, as you can see below:
- Switches the window manager from matchbox to Enlightenment (E17)
- Switches the main applications from the GTK-based (developed by Openmoko and OpenedHand) to Qtopia (but using X11, of course)
The switch to Enlightenment should not have been news to anyone now: Rasterman, the man behind the Enlightenment project, works for OpenMoko, after all. The announcement makes clear that despite the switch to Qt, Gtk+ applications will remain fully supported on OpenMoko phones.
Well, that’s a real kick in the teeth for GTK.
I wonder: with the OLPC going to the “dark side”, Maemo almost certainly switching to Qt (since Nokia have bought Trolltech), and now this, have we seen the end of the “Gnome Mobile” endeavour?
I dunno if that’s the case. They’ve certainly said that they won’t be switching to Qt on Maemo. Just making Qt more well supported there. I don’t think anyone particularly feels like rewriting the existing GTK code.
Probably not. GTK has its advantages (the license for example). But when it comes to getting stuff done, it seems the openmoko people have decided that Qt is the better way to go. Of course if the community pours a lot of effort into bringing the GTK apps up to snuff then that could change.
I thought it was funny that in that thread, one of the responses was from someone that was so fundamentally opposed to C++ (claiming it is completely unusable) that he immediately rejected the whole openmoko project based on this switch. I can’t see how anyone with such a closed, unyielding mind can call themselves a developer.
> darkside
hahahahaa! qt is not dark, it’s opensource, run by a really nice group of _people_ that have been very supportive to the freesoftware movement. you make it sound like it is microsoft. shame on what ever being that calls a useful gpl code producing entity ‘darkside’.
i think qt is a good thing for opensource, and opensource mobile, it’s a pity it didn’t became the standard in the first place.
> GTK has its advantages (the license for example).
i dont see any underlaying advantages. if you want to write a freesoftware app with qt its cool and free. if you want to write a commercial app you pay a little (yes it is not much taking in to regard what the developer cost are for a commercial app). with the money trolltech^H^H^H^H^Hnokia actually does something: they develop the most kickass cross platfrom toolkit available. i only see everyone benefit here.
i dont work for troll^H^H^H^Hnokia, i do code some kde app, i did try gtk and gnome development and based on experience with the two of then decided to develop my app on qt/kde as it develops a lot faster for me.
[/q]
Read again, he was talking about the OLPC, and by dark side he meant windows.
Don’t have to tell me. I even own a Qt commercial license. But the GTK license is still an advantage over Qt. It just happens to be that it’s not enough of an advantage to convince me to use it.
If you want people writing software for your platform, that pretty much sucks, because you are alienating a large userbase that doesn’t wish to publish their code under the GPL and has no intention in purchasing a Qt License.
That would be the common myth, but so far there really isn’t any evidence for it. Maemo is built on GTK, yet there is not much (any?) 3rd party commercial development for it. Also, Qt allows you to use many free software licenses (BSD, Apache, etc) for your code even if you don’t buy the commercial license.
Well, the one thing about the licensing here is that alot of people prob. feel that it is unfair that Qt-users can buy themselves out of the GPL-licencing. (And would rather prefer LGPL) Although it definetely makes bussiness-sense to do the dual-licence thing. As for busssiness interest in the toolkits I wouldn’t be suprised if Qt gives lower TCO than GTK+. (Since Qt seems to be a bit more RAD’ish)
I didn’t know that you could use other licences than GPL on Qt apps, so thanks for that info.
I still do prefer GTK licence though. But that’s simply because I think someone(as in “poor” indidivuals that can’t use money on Qt-licence) should be allowed create closed source apps if they want to. And if they shouldn’t be allowed to do that then large money-strong corporations shouldn’t either.
No disrespect to Qt, just personal preference. (And the dual-licencning thing is probably partly what have financed Qt to become what it is.)
A lot of people could argue that it is unfair that people can opt straight out of GPL licensing, without contributing code or anything else back to GTK. There’s no incentive to make GTK better, either through more code going on or a contribution of costs.
It’s not a myth, it’s a reality. Maemo and other platforms don’t have a lot of propertary/nonGPL software simply because they are not massive enough. Compare it to palmpilot, winCE, jave phones, etc which had a lot of software written for them simply because they are.
Excuses excuses. The Palm Zaurus was Qtopia based and had probably similar selection of software. So you have absolutely no evidence that Qt based platforms scare away proprietary developers. Until you can point at a Qtopia platform that actually doesn’t have any 3rd party proprietary apps you’re just guessing.
Then those people are not developing free software or open source. Any proprietary software on the openmoko phones would be going against openness of the project itself.
We want more free, open source software, not more closed proprietary software.
Well, but doesn’t C++ suck anyway?
This is quite common, sadly. When VLC switched to Qt for their main interface because of the cross-platform benefits (right-to-left languages for one) and because VLC is a real cross-platform open source application, the usual objections of licensing (VLC is still GPLed anyway) and the use of C++ were wheeled out.
People totally misunderstand C++, and have misunderstood a lot of comments people like Linus Torvalds have made about it. There’s nothing wrong with C++ itself, but as an object-oriented language you inevitably find yourself using a library like Boost or the STL. This is where pretty much all the problems come from.
I think the problem rather comes from the fact that you can program in C++ for ten years but still have a hard time taking over code from an C++ developer that has been using it for fifteen years.
There is such a thing as too much complexity and too much freedom. C++ has its uses, but imho application development isn’t one of them anymore these days.
I do agree that those ditching a project like OpenMoko for such a switch is an overreaction. You can still keep programs simple, the problem arises if you get that kind of C++ developer on the project who is complicating things for the sake of bragging.
You can say the same thing about Perl, and a lot of languages. Your post sounds generally like a rant against featureful/complex languages. It’s not bad to prefer simple, but let’s be honest about it.
I think most of the time that the complexity of C is ‘hidden’. Most C devs use some good libraries that make annoying things less annoying, which is roughly what you get at the language level from other languages.
That said, I personally prefer C to C++.
I think the STL and Boost are well designed. They are certainly easier to use than the equivalent C libraries! Especially std::string. And consider something like sort(). Way simpler with the STL than C.
*And* you get fewer security bugs and memory leaks…
Didn’t Trolltech already port Greenphone (or whatever qtopia phone edition is now called) to OpenMoko hardware? Sounds a bit like OpenMoko cutting their losses.
Qtopia Phone Edition was never called Greenphone. The Greenphone was a development mobile phone that ran Qtopia Phone.
and yes, Qtopia already runs well on the Neo.
i would buy a phone that ran E17.
While I am happy to see OpenMoko and Maemo edge toward Qt (my preferred GUI tk for Ruby development) I think all of these people should get going and put out a product. They have made some of the phones available to core devs, but no one else can get a damn phone.
They made a pretty CAD custom hardware designsetup…but no one can get the damn phone…
Are these folks for real? It seems they are a conglomerate of people that just want to collect some developers and pimp them out…maybe to Nokia?
Time to quit talking about what they will do and do somethings…money and action talks…BS walks…
the openmoko platform is as real as the Indrema gaming console…Until they have something to sell, they should quit their shoulda/coulda/woulda BS and get some product out…
Its seems just a bunch of vaporware…
Let us hope not!
Don’t let the door hit your behind on the way out, then.
I think you underestimate what it takes to get complicated, regulatory-compliant, mass-producible, *open* hardware out of the door, which is what they’re trying to do. Admittedly, they made some mistakes in the first generation of the phone, and the features list had to expand to satisfy all the people demanding WLAN support and other stuff, and even after all this there will be people who claim that it isn’t as good as one of the many different devices out there (doesn’t have a camera, play high-definition video, is “no good” for gaming or movie editing, and so on), but I think we should reserve judgement until it ships, which will most likely be in the next few weeks.
On the matter of the software, I think the last software environment looked reasonable, despite employing a far too desktop-like paradigm with respect to running tasks, and the icons and general interface hints weren’t particularly nice. From what I’ve seen of Qtopia, perhaps the icons will be nicer, but the paradigm isn’t quite what I’d expect from something new.
It is not that I don’t appreciate the hurdles that must be attacked and overcome, and I understand the needs to change hardware multiple times et cetera. Those hurdles consume resources, which is partially my point.
The “Are they for real” comment sounded sarcastic, but I didnt mean it that way. In an internet era with claims of non-apple PCs running OSX, SCO saying that UNIX is their property, and Microsoft claiming they want to coexist with FOSS, it is hard to verify the legitimacy of anything on the net until someting ships. Call me cynical, I guess…but only because of past experience…
On the other hand, I really, really want OpenMoko to succeed. The idea of being able to write stuff for my own phone rocks, as does the ability of people to do what they want with their technology (Listen carefully Apple!)! I just don’t relish the disappointment many (including myself) will experience if the barriers to entry consume all of the project’s resources or if the whole thing is a sham…
so don’t worry about “fully” anything. Just get something that boots and dials out the door. Release early, release often: it also applies to hardware.