Nokia's Maemo Switching To Qt

suka writes "During a keynote at the Gran Canaria Desktop Summit, Nokia's Quim Gil announced that a future release of Maemo is going to be built around Qt. Maemo Harmattan is going to switch away from GTK+ / Hildon, derStandard.at reports from the conference." Michael Pyne also writes with a post describing day one of the conference from a KDE perspective.

Comments

N900, please

[yk4ever]

That's all fun and games, but why are there no new products in the Internet Table line? C'mon, it's been almost 2 years since N810. The OS lives while the hardware was abandoned? Weird.

Re:N900, please

[migla]

I think the RX-51, aka "N900" is due "second half of 2009". The OS for it will not be backwards compatible with the n800

For OS developments regarding n8*0, check out the community project "MER" instead: http://wiki.maemo.org/Mer_Blueprint

Re:N900, please

[dbc001]

I'm an n800 owner and this is the first I've heard of Mer... What exactly is it? Is it a replacement OS? It looks like it's somehow related to Ubuntu MID. Is it stable enough that I should install it? Is it backwards-compatible with Maemo apps? Sorry about all the questions, but I've been anxiously awaiting some cool activity in Maemo-land!

Re:N900, please

[trampel]

AFAIK, it's a project to create an OS that is compatible with Maemo but entirely based on Free Software. One of the goals (IIRC) is to allow running programs using the Freemantle APIs on the semi-abandoned N800s and N810s.

Re:N900, please

[Anonymous Coward]

They do not develop new OS for it. They use Linux (or they port Symbian). The OS development is not easy. You can easily write all wanted applications for the phone, from SMS to emails and calenders and phonebook. But the OS is the most difficult to develop. Thats why Nokia use Linux (linux kernel) as their choise of the OS. It is robust, monolithic and all source codes easily available with great community support. The Qt developers already supports Linux well (not the Symbian) so porting is not necessary.

Starting over

[Vector7]

Hardly surprising, considering Hildon really wasn't very good. Sluggish, clumsy, and tending to waste a lot of very precious screen real estate - not that I see how switching to Qt changes any of those things. Still, it sounds like they're basically throwing the whole UI and all the software written for it out, and that sucks. I've long been tempted to write a little music toy app to run on my N800, but I should probably just buy an iPhone or a Pre (given that I don't actually carry the N800 around anywhere anyway).

Re:Starting over

[glebovitz]

take a look at the new animation framework, state machine, and the declarative UI if you want to see good reasons why they are making the switch.

Re:I know why..

[Kjella]

I know why.. Because QT was released under the LGPL, sorta recently.

Uh, maybe because Qt was bought by Nokia? They're the ones who decided to LGPL it, but they can do anything they want with it.

I wonder about this

[erroneus]

There is a lot of software for the Nokia N810 and below. Switching out to a new UI means a lot of stuff will either get uprooted or there will be a lot of libraries loaded into the machine's precious little memory.

Still, if the developers of software port over to the new environment quickly enough, it won't matter but I can't imagine things will be quick enough.

What can be done under Qt that can't be done under GTK? Is Qt more efficient in some way? What are advantages of Qt over GTK? I've never been clear on the differences... I just know they are different.

Re:I wonder about this

[abigor]

Qt is not just a gui framework. It provides a massive amount of extra stuff. Browse the documentation: http://doc.qtsoftware.com/4.5/index.html/

Note the WebKit integration, multimedia framework (Phonon, which was a part of KDE and later folded into Qt), OpenGL support, etc. etc.

Comparing it to GTK is like comparing a full-fledged desktop like KDE or Gnome to Blackbox.

Re:I wonder about this

[ricotest]

First and foremost Qt is not just a widget toolkit. It is a full development environment: it has a build system (qmake), a fully-developed IDE and widget layout editor (Qt Creator) and many, many extra libraries. To quote just a few examples, there are classes to handle tray icons (whether in KDE, GNOME, Mac OS X or Windows), classes for running TCP servers, integration with the Phonon media framework, the WebKit browser, SVG, databases, multi-threaded code and even scripting support using QtScript, an implementation of ECMAScript (JavaScript).

Qt is written in C++. GTK attempts to do object-oriented code in C and the result is a mess of explicit casting and macros. Seriously, most GTK C code looks horrible and is far less terse than the equivalent Qt program. This is mitigated when Python or Perl is used, but then you're sacrificing speed. With Qt writing C++ is basically as easy as using Java, C# or any other 'modern' language. All of the nasty stuff is taken care of. For example, Qt code is generally cross-platform.

Its signal and slot system is also very powerful. For example, you connect a button's click() signal to the QApplication's quit() slot, and the button will cause the app to close when clicked. These signal/slot pairs can even be set via the Qt Creator IDE, just like Visual Basic! Or you might start up a webpage download and assign a slot to handle the signal sent when the page has been downloaded. Qt's signal/slots are introspective and modifiable at runtime, and you define new signals and slots just like you define new methods for a C++ class. The drawback there is that Qt programs require a pre-processing pass by moc (the meta-object compiler), in order to generate meta-data for runtime signal/slot manipulation, and to offer some syntactic sugar around Qt's features. As a side-effect, Qt adds syntactic sugar for features some might find questionable, for example adding a foreach() loop for lists.

The build system, qmake, is quite simple: you list your source files, libraries and headers to link in a short configuration file (qmake can even generate this for you). qmake then generates a makefile from this data. This is useful as it also includes the 'moc' pass, but can be constrictive in some cases. You are, of course, not obligated to use qmake in your Qt project.

As far as widgets go, Qt's are comparable with GTK or any other toolkit out there. Qt does a better job of looking good on non-Linux platforms, such as Windows. It has a simple but flexible widget system that is much easier to use than GridBagLayout or any of Swing's more poweful layouts.

The main issue with Qt was that, up until recently, it was licensed under the LGPL and before that, it was under the restrictive 'Qt license'. This is no longer the case, so jump in!

Re:I wonder about this

[penrodyn]

Having used the GUIs in Visual Basic and Qt I would quite go as far as saying that signal/slot pairs is just like visual basic (VB). With VB you just click on the component (button etc) and it takes you directly to the code. In Qt you use this amazingly inept way of connecting stuff together visually by way of wires. I've never understood why other development environments have never taken the VB (or Delphi for that matter) route in GUI design. For simple GUIs it is highly productive, meas you can focus on user experience and coding rather than worrying about to to actually layout the controls.. But what is even more amazing most developers don't even know there are simpler way to build GUIs. Having said that I would agree that Qt is a very excellent toolkit, I just wish they would modernize the GUI construction end of it.

Re:I wonder about this

[dbIII]

There is a version of Qt for low memory devices (was called qtopia, now something else) which would give it an advantage on the platform instead of having to custom strip back the gimp toolkit. There's also the C vs C++ choice.

Re:I wonder about this

[umeboshi]

First, gtk+ is for C, while qt is for C++. Another major difference is that qt is more than just a widget toolkit, but an application runtime environment that provides widgets. This means that qt provides string handling, database connectivity, etc., although you don't have to use anything but the widgets and application objects, if you wish.

I thing maemo is mostly written in C, so some parts will probably have to be rewritten in C++.

This article may help a bit, although it only compares qt with gtkmm (the c++ bindings to gtk):
http://www.telegraph-road.org/writings/why.html

This article should be taken with a grain of salt, as it's pretty old, and may be inaccurate today.

I started using gtk+ with python, way back in the 1.x versions. The 2.x bindings for python were much better, allowing me to write more pythonic code using gtk+.

Later on, I decided to try out qt3, and I haven't looked back since. While it took a bit of getting used to, I found that it was easier to use qt, rather than gtk+, although I'm hard pressed to figure out exactly why.

One of the things I liked about qt over gtk+ was the separation of the layout widgets and the interactive widgets. Coming from gtk, this was something that took me a while to understand, but once I got the hang of it, I liked it, and think that it's a better way to organize the widgets. With gtk, a vbox holds child widgets, such as buttons, labels, etc. So if you want to rearrange them in an hbox, you have to destroy those widgets and make new ones in the hbox. In qt, the layout widgets are of type "layout", and you can only have layout children in layout widgets. The interactive widgets are children of the main widget (or a child widget of the main widget). These widgets are "placed" into the layout, but can be removed without being destroyed, allowing you to rearrange the layout more easily.

I also prefer the signal/slot mechanism in qt over the callback mechanism in gtk. On the average, it makes it easier to glue your widgets together, but there are a few circumstances where a callback mechanism is preferred, in which case you have to invent a new signal(s) and chain them together. This is because there is no order of slots called when a signal is emitted.

Also, the qt documentation was better, more organized, and easier to read than the gtk docs (at least around the time I switched ~2004).

Probably the largest reason why we're even having this discussion is due to licensing. Gtk gained a lot of popularity, due qt being licensed under the trolltech license, which restricted developers from using the free version in commercial products. The switch to gpl didn't do much to change this, although you could then create commercial products, but you also had to release the source for those products. So if you wanted to keep the source closed and use qt, you still had to purchase a commercial qt license.

Re:I wonder about this

[Eil]

There is a lot of software for the Nokia N810 and below. Switching out to a new UI means a lot of stuff will either get uprooted or there will be a lot of libraries loaded into the machine's precious little memory.

As it is, minor Maemo releases can (and sometimes do) break compatibility with applications while major releases are generally not expected to be backwards compatible at all. It works the same on any Linux distro or desktop environment. Development of Maemo has moved at a glacial pace, so when Nokia switches to Qt, I assure you it will be a major release.

I'm looking forward to Maemo on Qt 4 if for no other reason than it will make WebKit support a cinch. (The current official Maemo web browser uses Gecko and using it is generally an unpleasant experience.) In fact, if I recall correctly, there are some KDE folks trying to get KDE 4 ported to Maemo, with all the interface enhancements necessary to make it usable on small-screen devices.

Re:I wonder about this

[shutdown -p now]

You forgot one thing: performance. Qt guys take it very seriously, and have numerous tests showing off just how fast their rendering and layout code is. I would imagine that, for resource-constrained devices, this can be a big deal.

Re:I wonder about this

[shutdown -p now]

With VB you just click on the component (button etc) and it takes you directly to the code.

It would only do that for the "default" event of the component (e.g. Click in case of Button). To wire up other events - such as KeyDown or MouseMove - you still had to edit events in the property grid.

In reality, Qt signals/slots are exactly the same concept as VB events/handlers. "Default events" are a minor convenience feature, nothing more

I've never understood why other development environments have never taken the VB (or Delphi for that matter) route in GUI design.

Er, which ones didn't? WinForms is event-driven, and very similar to VB (down to the "double-click the button to auto-generate event handler for Click" you've described). WPF is broadly similar. Swing uses different terminology (listeners), but same concept. Really, it's one of the basic OO patterns, and most UI toolkits these days use it.

The thing that VB (and Delphi) truly lacked is the way to do dynamic layout of controls - this is absolutely crucial for DPI-independent and theme-independent code (you've got to be able to reflow the UI when font size changes, for example), and it simplifies localization a lot, as well. It's why all new (or just better) UI tookits - including Qt - are centered around the concept of dynamic automatic layouts; but at the same time, it's not something that you can easily edit visually (as has been demonstrated previously in case of HTML).

That said, Qt still lets you do absolute positioning of controls in a visual designer, if you really want to have it that way (shame on you!).

Re:I wonder about this

[Draek]

My main issue is that Qt is pretty strongly tied to C++, and I *despise* that language. Whereas GTK in C may be horrible, but the bindings to Python, Ruby and C# are all excellent and a newbie dev could easily believe they were designed for them from day one.

However, all the other advantages you mention are still valid, plus Nokia controls Qt so overall I support this move, it was the most logical thing they could do. They've lost me as a developer, but I don't think anyone's gonna cry over that ;)

Re:I wonder about this

[StormReaver]

> The main issue with Qt was that, up until recently, it was licensed under the LGPL....

Slight correction: until recently, it was licensed under the GPL; but is now licensed under the LGPL.

Re:I wonder about this

[abigor]

Qt absolutely has bindings in other languages. For example, check out PyQt: http://www.riverbankcomputing.co.uk/news

Re:I wonder about this

[drizek]

Unfortunately, the n8x0 development community has been kind of dead since the introduction of The Great One. It isn't completely gone of course, but it is a lot less active than it was a year ago. The current stable of apps aren't going to be as useful as we would like them to be. The other thing is that there never really was that many to begin with, nothing even approaching what the iPhone has. Most of the development seemed to revolve around fixing deficiencies wit hteh OS(alternate environments, media players, web browsers and other things that got done right the first time in android/iphone/pre). Also, many of the apps were just straight ports of desktop gnome/gtk apps, and it would be rather trivial to do the same thing with qt/kde apps.

Another thing is that by switching to qt you can tap in to the KDE development community. KDE devs who were never interested in writing gtk for the n800 may now get excited about the 900 and pick it up.

Re:I wonder about this

[Kjella]

My main issue is that Qt is pretty strongly tied to C++, and I *despise* that language.

Did you try C++ with or without Qt? I must admit, I don't like C++ outside of Qt, it brings the whole platform to another level. QStrings and QByteArrays are a godsend compared to std::string and char *. Using the QObject system I easily write applications with no memory leaks because it will delete any child QObjects when it goes, making it easy even without amy garbage collector. Finally, using signals and slots makes your application more robust - screw something up and nothing will happen because the signal never reaches its destination but it won't crash hard on an invalid pointer. Granted, I've heard you can do the same with STL and boost and duct tape, but I never managed to do it.

Re:I wonder about this

[ricotest]

Whoops, thanks for pointing that out.

Re:I wonder about this

[mibus]

Most of the development seemed to revolve around fixing deficiencies wit hteh OS(alternate environments, media players, web browsers and other things that got done right the first time in android/iphone/pre).

Fully agree. Some of my most-used apps are competitors for built-in ones (Canola and MPlayer top the list, MaemoMapper, I also used Modest in competition with its built-in email some time ago).

KDE devs who were never interested in writing gtk for the n800 may now get excited about the 900 and pick it up.

...and the existing pool of Gtk+ developers get frustrated after Yet Another API Change, and leave.

There's never going to be the glut of third-party apps that the iPhone enjoys, when the API isn't stable - and you'd have much better luck again if you could keep the ABI stable for more than a year or two...

I really love my N810 in some ways - but in others, it's fallen flat. Can't play full-res video, GPS is slow to lock, the built-in-browser is slow and painful, finger-scrolling doesn't always trigger properly (and text highlights instead), there's inconsistent use of finger-sized vs. stylus-sized controls, the touchpanel needs entirely too much pressure to be comfortable for extended use... my replacement for it will come when I (eventually) buy an Android phone.

Re:I wonder about this

[drizek]

If the HTC Hero wasn't so ridiculously expensive to get unlocked, I would already have one. I doubt I will be upgrading to an n900 from my 800. It just isn't fun using the device. By far the most absurd thing to me, considering that this is an internet tablet, is that the browser is still running on a pre-release version of Gecko for Firefox 2, taken from svn some time around december 2007. This may have been updated in diablo, but I don't notice it. It is nowhere near the capability of webkit/3.5. Nokia really dropped the ball here. THere is no technical reason why web browsing has to be so slow with so many amazing open source rendering engines around.

Granted though, the n8xx is playing with one or two generation old technology. All it really needs is just multitouch and a faster CPU and just a tiny bit of attention to detail to the software and it could be great.

You have the dual slot SDHC, the stereo sound, high res, giant screen, excellent build quality, the kickstand and just so many things that make it stand out. Now that I have a netbook though, I don't think the n800 is worth the space it takes up in my pocket.

Re:I wonder about this

[Anonymous Coward]

GTK attempts to do object-oriented code in C and the result is a mess of explicit casting and macros.

Take a look at GTKMM, the C++ binding for GTK+, which uses:

• signal handlers with full C++ type safety (no macros)
• the standard C++ library (including STL containers and iterators, unlike Qt)
• object compositing
• automatic memory management, including for dynamically created objects
• internationalization with full UTF-8 support and C++ std::strings
• C++ inheritance to define your own widgets
• everything in nicely defined C++ namespaces

The GTK+ people chose C because back then the C++ compilers were not as mature as they are now (which is why Qt uses its own language as preprocessor, to fill in the old C++ compiler gaps), and practically any language could call C libraries, but not C++ libraries (not without extern "C", anyway). GTKMM provides a nice alternative, too bad so few people choose to use it...

Re:I wonder about this

[ardor]

gtkmm is sort of ok, but Qt is still superior. It has a much cleaner API, better documentation, a MUCH more powerful canvas widget, support for reflection of QObjects (very useful for stuff like RPC), etc.
gtkmm on its own is fine, but it suffers from having to inherit the gtk wall bangers, most prominently, the API.

Re:I wonder about this

[Anonymous Coward]

Take a look at GTKMM, the C++ binding for GTK+, which uses:

The problem is that gtkmm is always "something to take a look at" instead of being something people actually use. Why is that? The C mentality of gnome devels? General badness of gtkmm?

One advantage of Qt is that it was C++ from birth, so C++ is the unquestioned first class citizen in the Qt world.

Re:I wonder about this

[Aladrin]

I quit writing C/C++ code years ago because of the insane amount of micromanagement I was having to do. Your post has about convinced me to give Qt a chance, though. Thank you.

Re:I wonder about this

[Klivian]

My main issue is that Qt is pretty strongly tied to C++ This is repeated over and over again and even if people believe it, it' still not true. That is pure FUD. Whereas GTK in C may be horrible, but the bindings to Python, Ruby and C# are all excellent Fact is, the Qt bindings for those languages are more comprehensive and more up to date than the GTK counterparts. The Qt bindings all have powerful automatic tools for generating bindings for those languages, making it easy to keep up to date.

Re:I wonder about this

[GeneralAntilles]

Z

What can be done under Qt that can't be done under GTK? Is Qt more efficient in some way? What are advantages of Qt over GTK? I've never been clear on the differences... I just know they are different.

Simple, Nokia is trying to build an ecosystem that will allow developers to deploy their software across several different platforms using only one toolkit (Qt). Since Qt will be available on both Maemo and S60.

Re:I wonder about this

[EsbenMoseHansen]

I quit writing C/C++ code years ago because of the insane amount of micromanagement I was having to do. Your post has about convinced me to give Qt a chance, though. Thank you.

The thing about C++ is that it is easy to back yourself into a corner. E.g, if you write "delete" anywhere, you are most likely doing something very wrong. Get past this, and it is a quite powerful and effective language, though I would love to have some legacy stuff cleaned up --- like the default conversions.

Re:I wonder about this

[Roxton]

E.g, if you write "delete" anywhere, you are most likely doing something very wrong.

Can you point to any articles, framework sample code, or discussions of best practices that reflect the methodology you're describing?

Re:I wonder about this

[JudgeJackson]

The pattern is known as "Resource Acquisition is Initialization", or RAII. Search for that and you'll get a ton of hits. Simple rule of thumb: new in constructors, delete in destructors. There are smart pointer classes available to help you do this.Boost provides a set of them and it sounds like Qt has something similar.

Re:I wonder about this

[jrumney]

What are advantages of Qt over GTK?

I'd imagine the fact that Nokia now owns it might have something to do with this move.

Re:I wonder about this

[Draek]

Without, and from reading a few Qt tutorials it seems that as long as you stay within it, it does make C++ programming much nicer and easier than 'pure' C++, so I'm gonna try writing a 'real' app in it now, thanks!

Re:I wonder about this

[Draek]

Fact is, the Qt bindings for those languages are more comprehensive and more up to date than the GTK counterparts.

Do you have a source for that statement? because I've just tried PyQt again and, while its very nice and a far cry from the mess I tried ~3 years ago, so is PyGTK and I can't find any difference between either that doesn't come from the toolkits themselves rather than the bindings. And while I haven't tried Qt# yet, I have a hard time believing it could be better than Gtk# which is maintained by some of the people behind Gtk and Gnome themselves.

Re:I wonder about this

[Anonymous Coward]

Using the QObject system I easily write applications with no memory leaks

Good luck trying to verify that with Valgrind - it gets totally confused by Qt's little memory management tricks and reports all kinds of memory leaks.

Have you verified that there weren't any memory leaks or are you just guessing?

Re:I wonder about this

[Klivian]

Do you have a source for that statement?

What about the PyGTK and GTK websites. According to the download links the latest stable and updated PyGTK are for GTK 2.14, and GTK 2.16 has been out since the middle of March. For PyQt you don't find many minor releases of Qt that has not been followed by a updated PyQt release inside a week or two, for the last 5 years or so.

As for being comprehensive, the Qt bindings(at least PyQt) has a close to full coverage of the Qt classes. Last I checked PyGTK did not offer the same. And since Qt are not only a GUI library, the difference gets even larger as the bindings provide so much more.

Re:I wonder about this

[Hognoxious]

Can you recommend any resources (books or online) for learning it? I've use QT/Qtopia apps on my old Archos PMA and I like the simple clean GUI. Just right for a small device.

In my persoanl case, I've programmed a fair bit in 4GLs but only a little in C and Java and no C++.

Qt != KDE, GTK+ != GNOME

[Anonymous Coward]

It seems like they're still planning on using a lot of GNOME components, but putting a Qt skin on it. I just wonder if it is the best of both worlds, or the worst of both worlds...

Re:Qt != KDE, GTK+ != GNOME

[FooBarWidget]

Why would it be worse of both worlds? They're just libraries; mix and match however you like.

Nokia owns Qt

[iYk6]

What can be done under Qt that can't be done under GTK? Is Qt more efficient in some way? What are advantages of Qt over GTK? I've never been clear on the differences... I just know they are different.

Personally, I prefer Qt because of the superior documentation. I think Nokia prefers Qt because they own it now. It's sort of like when Microsoft bought Hotmail, and moved the servers from FreeBSD to Windows. It wasn't because it was better, but because it was theirs.

Re:Nokia owns Qt

[erroneus]

That's a really bad example... or perhaps a really dark example. In your example, that move was rather poorly done initially. They had all sorts of problems as Windows simply couldn't handle the task. Windows had to evolve to the task and as I understand it, the change was only superficial with BSD still on the back end or has that changed since I last heard?

At the very least, if you are accurate in your parallel, then that forebodes a dark and troubled time during the change over. Of course they would make up for it if while they are creating a new face, they would also create or migrate any heavily used or highly popular programs and utilities to go with it starting with a GPS map and route program that makes use of various types of map data including but not limited to openstreetmaps. I'd be rather pleased with it if it already had everything I want on it to begin with.

Re:I know why..

[interval1066]

Exactly. Has nothing to do with the quality of either toolkit, some one at Nokia said "Why are we building this product with the competing technology?" I always make a point of supporting both myself where it makes sense. Even though Qt is a nutty blob of nonsense.

Re:Ubuntu?

[Antidamage]

Ubuntu is a distro. QT is more of a graphics and application framework.

Re:I know why..

[mrsteveman1]

You don't think they bought Qt because they thought it was better?

This sounds wrong

[BRSloth]

Fromt TFA: "Nokias motivation for this move as being mostly driven through the desire for easier cross-platform-development, citing Maemo, Symbian and the desktop as examples."

One thing that sounds incredible wrong to me is the fact that they are saying that Qt was chosen to make "easier cross-platform-development". The applications that were ported directly from desktop to Maemo (Xchat is the first one the comes to my mind) have an incredible bad look in the device. Building an interface for a device that runs in a small screen (4.1 inches) with a small resolution (800x480) that also uses a large pointer (e.g., most of the screen is designed to thumb usage) is not the same as building an interface for normal computer screens and resolutions.

The move is simple political: Nokia controls Qt now, so they will use their own toolkit. It's not based on merits of the toolkit (or problems of the other.) But hey! Why tell people the truth, right?

Re:This sounds wrong

[Yokaze]

> One thing that sounds incredible wrong to me is the fact that they are saying that Qt was chosen to make "easier cross-platform-development". [...]
> The move is simple political: Nokia controls Qt now, so they will use their own toolkit. It's not based on merits of the toolkit (or problems of the other.) But hey! Why tell people the truth, right?

And the reverse couldn't be possibly true: That Trolltech Qt was bought based on the merits of the platform and because Nokia expected easier cross-platform-development. Why do you think Trolltech started porting Qt to the S60 platform?

> Building an interface for a device that runs in a small screen (4.1 inches) with a small resolution (800x480) that also uses a large pointer (e.g., most of the screen is designed to thumb usage) is not the same as building an interface for normal computer screens and resolutions.

Yes, it isn't. But I doubt, having larger entry barrier by having to learn a whole new API (Android, Symbian OS) or even language (iPhone OS) makes it easier to create a good application.

Re:This sounds wrong

[JohnFluxx]

I think it came across slightly wrong in the wording.

The point is not that you can take a full desktop application and run it on the maemo. The point is that you can develop it on the desktop. You don't need special hardware, you don't need an emulator etc.
You can use your normal environment, use Qt Creator for the IDE, and write your program. You can use your normal debugger and profiller, and so on, since it's all native code.

Then at the end of the day, you can just click and button and it recompile it for the arm for the machine.

Re:This sounds wrong

[Hognoxious]

The Arcos PMA uses QT. It's fine on a small screen, or at least can be if it's done right.

Re:This sounds wrong

[vakuona]

The reason for buying was that it could be bought. GTK couldn't be bought, hence Nokia couldn't buy it.

Thank you.

[itomato]

I was really liking the Debian chroot environment, but the Maemo overhead put the squeeze on.

Nokia owns Qt.

[Futurepower(R)]

Before you read too far, realize that Nokia owns Qt. It is not surprising that Nokia products use Qt.

Re:Nokia owns Qt.

[PiSkyHi]

Above post should be modded "-1 trolltech"

This is the Death of Maemo,if it really ever lived

[keithpreston]

I can tell you right now, this will kill Maemo. QT is a pretty good GUI toolkit, but this is going to draw in QT Embedded (QWS server and such). I personally have been working on an Embedded QT device for 2 years and can tell you, QT Embedded is horrible. Nothing more then a Demo written by Trolltech to try and expand the market share. The biggest pain with QT, is that since it tries to be cross platform is it re-implements everything (Networking, Audio, Mutexs etc... etc..). They make it fairly easy to use their bad, slow code, while the "beautiful" non-standard signal slot system makes it a pain to integrate with real C or C++ code. If they wanted C++ they should of gone with GTKmm.

Re:I know why..

[Anonymous Coward]

What else should they do, buy FSF?

It was either that or switch Symbian to Hildon

[eean]

Nokia wants a common platform across their internet tablets and smart phones. Given that the Symbian is going to support Qt, and the Symbian user base is much greater, its makes sense that Maemo would want to have access to the 3rd party apps written for the user base that numbers in the millions

And really it was clear in the talk he gave that the Maemo stack is still mostly unchange, and still using most of the Gnome libraries including crucial stuff like Tracker. Really even with the change in UI toolkit, its more Gnome then KDE, especially as none of the Maemo stack actually originated from the KDE community, where as much of it did from the Gnome camp.

example?

[eean]

I can think of a couple like GStreamer and Telepathy, but in both cases the support isn't 100% yet. And both are really crossdesktop from the beginning (Telepathy is just a DBus spec after all)

Re:example?

[t_hunger]

I can think of a couple like GStreamer and Telepathy, but in both cases the support isn't 100% yet. And both are really crossdesktop from the beginning (Telepathy is just a DBus spec after all)

Dunno about GStreamer, but Telepathy is a really broken spec that puts every pecularity of every protecol it ever thought about supporting into their apis. Considering that the only useable telepathy implementation is jabber at this time (in two different flavours) and hardly anything else works that project is a lot of hot air.

GNOME just need to die

[metamatic]

With the Mono infection and the reliance on GTK, the best thing would be for GNOME to go away. It started because Qt wasn't LGPL. That no longer applies, so let it die.

Re:GNOME just need to die

[Anonymous Coward]

With the Mono infection and the reliance on GTK, the best thing would be for GNOME to go away. It started because Qt wasn't LGPL. That no longer applies, so let it die.

If you want GNOME to go away, there first has to be a halfway decent Qt-based desktop environment. Even after a year and a half after it's initial release KDE4 isn't up to par.

KDE4 needs to die.

Re:GNOME just need to die

[ardor]

I am using KDE 4.2 right now, and it is amazing. Everything works well, no bugs, no problems. The underlying tech is easily superior to gnome as well.

Honestly, how many KDE4 bashers have actually *tried* it?

Re:GNOME just need to die

[Quantumstate]

I tried it just the other day. It seemed fairly nice but I was editing some options and it crashed on me, I had to kill it from the command line. I had only been using it for about 20 minutes. I have never had gnome crash on me at all. It did look fairly nice though once I had configured it a bit. I find gnome works really well though and the only minor gripe I have with it is the odd obsession with making everything look huge.

Re:GNOME just need to die

[BlackCreek]

Speaking as a former KDE user: I did try kde4.2 and IMHO it sucked.

Does not having positive words for the current KDE release makes me a basher?

Re:GNOME just need to die

[Timmmm]

I agree. It's just too damn slow and the whole 'widget' UI is pretty messed up. Some simple examples (there are many more):

* Why do widgets allow you to rotate them? Why on earth would anyone want that?
* When you resize widgets the centre is always fixed. I.e. to move a single edge/corner you have to resize the widget then move it.
* WTF is up with that 'cashew'? Nobody wants it.
* It's really damn slow. I've tried with both nvidia and ATI cards but couldn't get it at all smooth.
* Plasma and Qt use different themes. Makes everything weirdly inconsistent.
* The layout everywhere is generally much less clean and simple than gnome or KDE3 (ignoring the toolbars).

I used to use KDE3 and it was great, but they really have messed up with KDE4. Anyone want to start a lightweight Qt desktop? :-)

Re:GNOME just need to die

[puppetluva]

+1 for the parent posts.

Although the gtk/gnome UIs have always been awkward and clunky, they've never been dangerous before -- With the mono infection, they're downright dangerous.

Re:GNOME just need to die

[ardor]

I cannot reconstruct any of these points (minus the centre resize thing and the rotate, though I do not consider the latter to be a problem). What distro are you running? Note that Kubuntu is among the WORST KDE based distros, riddled with bugs, slow etc. I am using debian squeeze with KDE 4.2.

Re:GNOME just need to die

[metamatic]

And Linus began Linux because he didn't know FreeBSD existed.

FreeBSD 1.0 was 1993. Linux started in 1991. Fail.

Re:GNOME just need to die

[bill_mcgonigle]

I am using KDE 4.2 right now, and it is amazing. Everything works well, no bugs, no problems.

Really? I switched from GNOME at KDE 4.1, and it's still rough at a well-patched 4.2. Plasma is much less crashy in the past few months, but it still does. There's too much Microsoft Windows influence in controls and some parts of the UI are really slow on old computers. Some of this is much better in 4.3, I'm installing that next week. Some things are just crazy ideas that must go, like the bizarre multi-selection controls in object lists. I'm been using computers for thirty years and the number of modes still confuses me - everybody else I've shown it to has been completely baffled. KWin compositing still crashes any multiuser setup I've tried. Their 'start' menu is still confusing, etc.

The KDE bug tracker shows just how many things are still rough. I file/cc bugs as I find them, there's tremendous work going on. Something around KDE 4.5 ought to be good for the typical desktop user.

The underlying tech is easily superior to gnome as well.

Definitely. They've taken on the task of outpacing the best commercial desktop platforms in a third to a quarter of the time, and they're going to do it. Once something like 4.5 hits parity, the sky's the limit. The only thing that can derail their efforts is if they decide 5.0 needs another complete rewrite and they don't maintain 4.x while they're doing that.

Re:GNOME just need to die

[Draek]

Aye, my memory failed me a little, it was regarding 386BSD, FreeBSD's progenitor, rather than FreeBSD itself (see here). But my point still stands: Linus' original reason for creating Linux is no longer valid, does that mean we need to kill Linux or that it serves no useful purpose anymore?

And you still haven't answered why are you posting your rant on a discussion about mobile platforms, which doesn't concern either Gnome or KDE in the slightest.

Re:GNOME just need to die

[socceroos]

* Why do widgets allow you to rotate them? Why on earth would anyone want that?

My wife is an average user. She's a primary school teacher. I gave her a fresh install of Gnome and a fresh install of KDE 4.2 and she refuses to use Gnome now - not because KDE is 'just better' but because she said it was more aesthetically pleasing and that she could rotate her picture widgets on the desktop and have a cool collage of her favourite photos. Needless to say, I couldn't care less about such things.

I'm not saying its the most useful feature - but she prefers KDE for the simple fact that she feels it looks better and she can rotate her widgets.

Re:GNOME just need to die

[metamatic]

Linux was started as a replacement kernel for Minix, so Linus could learn about 386-specific opcodes. BSD wouldn't have been a sensible alternative.

And 386BSD was released in 1992, which was still later than 1991 when Linux development started.

My reason for commenting on GNOME is that someone commented that GNOME should adopt Qt. i.e. I didn't raise the subject.

What's good on Hildon/Maemo?

[itomato]

Can you name three apps that are available for Maemo that you would honestly miss?

The mail app blows, the browser sucks, the media player is lousy, and the application manager is beta-quality, at best.

Don't get me wrong, I love my n810, mostly for the hardware - Maemo is a drag.

-

Qt vs GTK: http://slashdot.org/askslashdot/01/11/21/0227206.shtml

Re:What's good on Hildon/Maemo?

[erroneus]

Not three... but one -- the GPS mapping and routing app. But as someone else was describing how Qt covers a wide range of functions including SVG rendering, I think that perhaps such map software might be improved with a port from the existing Maemo to the future Maemo.

Just about any other apps that follow, I'm sure, will be acceptable. It plays music. It plays video. It plays some simple games. It's a portable device. It does what I wanted it to do and I am sure under a new environment, it might well be improved. It would be good to see some new life in my old toy.

Re:This is the Death of Maemo,if it really ever li

[ceallaigh]

The real story is the Nokia / Intel announcement of cooperation on Atom/mobile products. Intel seems rather focused on Mobilin for MID with a long term strategy for handsets. While Nokia will be pushing their Ovi stores/maps/content with a new UI for Symbian. I doubt that Nokia ever looked on Maemo for more than an R&D effort. Commercially it was never a success nor a viable consumer product - a geek toy yes, a popular consumer product never. Maemo is irrelevant. The real thing to watch is the Intel/Nokia relationship on handsets - see how that evolves from processor choice to OS. Sean

Re:This is the Death of Maemo,if it really ever li

[chill]

...except at least they were attemping to make a useful device from day one, while Nokia has totally let that ship sail into Apple's hands.

Is that Kool-Aid good?

Nokia sells 4x more smartphones than Apple does, with over 40% of the worldwide market. Nokia has won more design awards for phones than Apple, by a long shot. They even have smartphones (n97) that handily beat the iPhone. The problem is, Nokia caters to users NOT phone companies and thus the North American carriers don't sell their smartphones. All you can really get in the U.S. is their standard phones.

They're trying to get a bigger presence in the U.S. market, and are examining how to leverage QT, Symbian and Linux in doing that. At least they aren't sitting on their collective asses (like Motorola) and getting crushed.

Don't write them off.

http://money.cnn.com/2009/01/12/technology/hempel_nokia.fortune/
http://news.cnet.com/8301-13579_3-10245339-37.html
http://www.nokiausa.com/find-products/phones/nokia-n97/specifications

Re:GNOME

[Kjella]

Pretty much everything I've done with Qt tells me KDE should be a much better desktop than Gnome. But the truth is that most of the large desktop distributions use Gnome, Ubuntu is much bigger than Kubuntu and same goes for the others. None of the big three hitters Firefox, OpenOffice or GIMP are KDE applications - ok not all are Gnome apps either but there's not many "killer KDE apps" around. Don't get me wrong, they're all perfectly okay but nothing really rocks the boat.

Sometimes I just want to shoot the people that did usability studies for KDE, like for example making dolphin's file transfer dialogs into notifications that disappear whether they're done or not. When I'm moving a file it's very often a point to know when it's complete, and it irritates me to no end that every time I must click the notifications icon to know when it's really done. To be honest, I think I'd love to see Gnome/Qt. Maybe we'd see more head-to-head competition - or even more shared components? That's kind of hard today, both need to invent the wheel in C/C++ respectively.

+1 Insightful

[itomato]

Flamebait, maybe. These kids today do not respect a healthy, if inflammatory point. Makes sites like Digg a nightmare.

The reason GNOME was started is nearly the same the reason it should fade away. Because KDE relied upon (the then closed) Qt, GNOME was started, as a workaround using the GIMP Tool Kit. It's a similar situation with Mono, but hairier.

This is actually something to remember, thanks!

"I mailed Richard Stallman to let him know that this interesting project existed. KDE was licensed under the terms of the GNU GPL. I got a reply back from both Erik and Richard pointing out that KDE dependency on Qt resulted in a piece of non-free software. Qt did not end users the right to modify, redistribute nor distribute modifed copies of the code and violated the terms of the GNU GPL."

From The Story of the GNOME project.

Re:This is the Death of Maemo,if it really ever li

[Capt. Beyond]

I can tell you right now, Maemo will not use Qt Embedded.

'Bad code' is very subjective, and I would like you to prove Qt is slower than Gtk. Just because you say it's so, does not make it as such.

One reason they are going with Qt, is because they bought Trolltech. They could not have that much control over gtkMM, however ancient and unmaintained that code is.

Re:This is the Death of Maemo,if it really ever li

[Capt. Beyond]

oh ya, I just love working with auto generated c++ code.

Nokia's ownership of Qt was the only reason Qt got re-licensed under the LGPL, and no, just having an LGPL license does not mean someone has full
control over it's development. So the two are very different things.

Re:This is the Death of Maemo,if it really ever li

[thatkid_2002]

They even have smartphones (n97) that handily beat the iPhone.

I have a n96. Hands down the WORST phone I have ever seen. It is the buggiest and slowest piece of crap and I blame it on a Symbian implementation which has been hacked to death. I prefer the $100 AUD LG phone I picked up two years ago.

However, I really like the idea of maemo (it seems exactly like what I want, except it doesn't have 3G built in or a phone yet) and so far I would happily give it a design award, and I think that QT will only make it better.

However, they really need to be careful or they will end up like OpenMoko - which is damn near dead. They switched tookit 3 times IIRC.

Re:This is the Death of Maemo,if it really ever li

[johannesg]

You seem to know what Maemo actually is. Since the summary doesn't care to enlighten us, could you maybe do the honors?

Re:This is the Death of Maemo,if it really ever li

[ardor]

I have personally worked on QT embedded projects as well, for well over a year. Some platforms weren't supported out of the box, yet I didn't find it to be particularly painful. Neither was the signal/slot-system. I did use GTKmm before Qt though, and had to endure all the braindead API designs inherited from gtk.

Gtk+ is not Nokia's problem

[jipn4]

Nokia has its own lightweight GUI library that they use with Symbian--and their UIs suck. They have built applications with Gtk+--and their UIs suck. They have build Windows and OS X desktop apps--and their UIs still suck. I think the problem Nokia has with GUIs and software has to do with how they develop software, not whether they use Gtk+ or Qt.

Another problem with their choice is that it ties them to C++; the trend in mobile development, however, is towards other languages, like Javascript (Pre), Java (Android), Objective-C (iPhone), and C# (Windows Mobile). Only Symbian steadfastly clings to C and C++. That would be fine if Symbian actually ended up being the fastest and having the best UI of the bunch, but it's actually the slowest and least responsive of the bunch.

Re:Gtk+ is not Nokia's problem

[ultrabot]

Another problem with their choice is that it ties them to C++; the trend in mobile development, however, is towards other languages, like Javascript (Pre), Java (Android), Objective-C (iPhone), and C# (Windows Mobile).

Actually, I think this will end up being a competitive advantage in the long run. If Nokia smartphones end up being the *only* smartphones that run (mostly) raw native code compiled straight for the metal, they will end up being the fastest in the long run, given equivalent hardware.

That would be fine if Symbian actually ended up being the fastest and having the best UI of the bunch, but it's actually the slowest and least responsive of the bunch.

The problem with Symbian isn't C++ - it's C++ done horribly wrong, and series of unfortunate technical choices (e.g. pervasive client-server architecture).

Re:Gtk+ is not Nokia's problem

[jipn4]

Actually, I think this will end up being a competitive advantage in the long run. If Nokia smartphones end up being the *only* smartphones that run (mostly) raw native code compiled straight for the metal, they will end up being the fastest in the long run, given equivalent hardware.

No, they won't. C++ is fast for small inner loops because programmers there can take full advantage of its features. Big applications end up being slow and bloated in C++ because programmers simply cannot manage the complexity anymore: all their time goes into chasing pointer bugs and dealing with include files, and little remains for performance tuning and algorithms.

And what is this "long run" you're speaking of anyway? If it takes 5 years for Nokia to optimize their current C++ applications, do you think anybody will care? Few of the PDA applications from five years ago are even remotely up to date today. There is no "in the long run" for software; what counts is what you can deliver in 3-6 months, not in a few years.

Re:Gtk+ is not Nokia's problem

[ultrabot]

No, they won't. C++ is fast for small inner loops because programmers there can take full advantage of its features. Big applications end up being slow and bloated in C++ because programmers simply cannot manage the complexity anymore: all their time goes into chasing pointer bugs and dealing with include files, and little remains for performance tuning and algorithms.

That's bollocks. C++ is not really that much less productive than Java/C# if you have a good platform toolkit to go with it (Qt). With Qt, you don't really manage your memory manually most of the time, the classes do it themselves through implicit sharing.

Admittedly, C++ is much less productive than Python & other dynamic languages, but that's not the issue at table here; we are comparing against Java, C#, ObjC.

And what is this "long run" you're speaking of anyway? If it takes 5 years for Nokia to optimize their current C++ applications, do you think anybody will care?

The phone applications easily have a life span of several years. They get improved, but rarely rewritten.

This applies even more so to "platform" level stuff. If you write more of that in C++ than Java, you'll have a faster platform, given equivalent algorithms.

There is no "in the long run" for software; what counts is what you can deliver in 3-6 months, not in a few years.

It seems Nokia was able to turn a profit with Symbian, even if Symbian is widely dreaded as the least productive programming environment in existence. I believe they will do great with Qt, and attract a great deal of third party interest as well.

Re:Gtk+ is not Nokia's problem

[jipn4]

Admittedly, C++ is much less productive than Python & other dynamic languages, but that's not the issue at table here; we are comparing against Java, C#, ObjC.

That is exactly the issue. Objective-C is a dynamic language, and the CLR and JVM support dynamic languages. Java is a lousy language, but even it has garbage collection, full RTTI, and reflection.

This applies even more so to "platform" level stuff.

Qt is an application and GUI toolkit.

If you write more of that in C++ than Java, you'll have a faster platform, given equivalent algorithms.

But the "algorithms" aren't equivalent because the C++ programmer aren't even catching up with the level of functionality that users of better languages can implement in the same amount of time.

It seems Nokia was able to turn a profit with Symbian, even if Symbian is widely dreaded as the least productive programming environment in existence

Nokia phones succeed because Nokia makes great phone hardware. Their software is consistently rated poorly, and it is slow. And there is no real mainstream third party application market. I know: I use a couple of Nokia phones daily.

With Qt, Nokia is betting on the wrong horse. They should either switch to Android or buy Palm or do something entirely new.

Re:Gtk+ is not Nokia's problem

[ultrabot]

Qt is an application and GUI toolkit.

No, it (QtCore part) can be used throughout the stack.

But the "algorithms" aren't equivalent because the C++ programmer aren't even catching up with the level of functionality that users of better languages can implement in the same amount of time.

Or so you would believe. Have you ever actually tried Qt? Or is all of this based on Java/C# propaganda you've read & bad C++ experience with crappy frameworks?

Don't forget that things will only get better as the new C++ standard rolls in (we'll get lambdas, auto, ...).

With Qt, Nokia is betting on the wrong horse. They should either switch to Android or buy Palm or do something entirely new.

We'll see about that. I have to say my enthusiasm for Android waned the second I heard of Nokia buying Trolltech.

Re:Gtk+ is not Nokia's problem

[vakuona]

Actually, I think this will end up being a competitive advantage in the long run. If Nokia smartphones end up being the *only* smartphones that run (mostly) raw native code compiled straight for the metal, they will end up being the fastest in the long run, given equivalent hardware.

For that to be true, they would have to make sure their software was optimised for all different hardware they produce. If they are going to stick to one processor then maybe. But that is a lot of work for the developers, with the way current app stores work. Native code for applications is overrated. Users do not care about it. and oh, the long run absolutely does not matter.

Re:Gtk+ is not Nokia's problem

[jipn4]

No, it (QtCore part) can be used throughout the stack.

You could use Perl "throughout the stack" if you worked hard enough at it; that doesn't make it a good idea.

Or so you would believe. Have you ever actually tried Qt? Or is all of this based on Java/C# propaganda you've read & bad C++ experience with crappy frameworks?

I've been programming in C++ for more than 20 years and in C for more than 25 years, as my primary languages. My dislike of those languages is based on intimate familiarity with it.

Don't forget that things will only get better as the new C++ standard rolls in (we'll get lambdas, auto, ...).

The problem with C++ isn't what it lacks, it's that it's so complicated that people don't manage to produce good results in reasonable time. Although the new C++ standard fixes some things, it makes the basic problem with C++ worse--its complexity.

We'll see about that. I have to say my enthusiasm for Android waned the second I heard of Nokia buying Trolltech.

How is buying Trolltech going to fix the problem that Nokia's engineers apparently don't know how to organize a menu, organize a dialog, perform user testing, or perform bug tracking?

Re:Gtk+ is not Nokia's problem

[ultrabot]

I've been programming in C++ for more than 20 years and in C for more than 25 years, as my primary languages. My dislike of those languages is based on intimate familiarity with it.

I wasn't asking about C++, but about Qt - I developed a visceral dislike of C++ when doing Symbian development, even if it was not C++'s fault at all.

Although the new C++ standard fixes some things, it makes the basic problem with C++ worse--its complexity.

I have never been bitten by the complexity of the language. That problem is a bit overhyped.

OTOH, I have been bitten by crappy libraries (yes, that includes the standard library).

How is buying Trolltech going to fix the problem that Nokia's engineers apparently don't know how to organize a menu, organize a dialog, perform user testing, or perform bug tracking?

Subjective.

Re:Gtk+ is not Nokia's problem

[jipn4]

I have never been bitten by the complexity of the language. That problem is a bit overhyped.

It's not a problem if you develop by yourself. It's a problem if you try to run software projects with dozens of programmers.

Subjective.

Usability is objectively measurable, and Nokia's is objectively bad.

Re:This is the Death of Maemo,if it really ever li

[Anonymous Coward]

I have a n96. Hands down the WORST phone I have ever seen. It is the buggiest and slowest piece of crap When I first got mine I was not exactly enamored by the battery life or UI responsivity. Switching off wifi scanning gives me 2.5 days on a charge now and switching off UI effects has made the UI quite usable. I really thought I'd made a bad buy in the beginning but I'm quite happy now. Plus, not even in Linux have I found a podcast client as good as the one in the N96.

Why do they call themselves "trolls"?

[Futurepower(R)]

Nope. It's true, Nokia bought Trolltech. Quote: 'Eirik Chambe-Eng, Chief Troll and co-founder of Trolltech continues "We are thrilled to join forces with Nokia." '

I wonder if the people who work with Qt (cutie) will continue the tradition of calling themselves "trolls"? A troll is "an imaginary creature of human-like form, very ugly and evil-tempered".

I doubt very much that the people who work with Qt are ugly and evil-tempered. What I think they meant is that, originally, they had a feeling of not belonging.

I wonder if they will continue calling themselves trolls now that they are part of Nokia.

Re:Why do they call themselves "trolls"?

[PiSkyHi]

Well, it might be a little tricky now that they also changed the default pronunciation to "Cute".

Re:Why do they call themselves "trolls"?

[fatphil]

Moomintrolls?

Re:This is the Death of Maemo,if it really ever li

[ultrabot]

The biggest pain with QT, is that since it tries to be cross platform is it re-implements everything (Networking, Audio, Mutexs etc... etc..). They make it fairly easy to use their bad, slow code, while the "beautiful" non-standard signal slot system makes it a pain to integrate with real C or C++ code.

This "biggest pain" of yours is what makes it cross platform.

You know what? You don't need to use these "slow" wrappers, you can use the file descriptors directly if you wish, and call to posix to your hearts content, if you don't care about running the code outside Linux. Good luck explaining that to your manager though.

I invite to you to investigate how "slow" these wrappers are by just reading the code:

http://tinyurl.com/loerlj

They make it fairly easy to use their bad, slow code, while the "beautiful" non-standard signal slot system makes it a pain to integrate with real C or C++ code.

PIBKCAP, probably.

Re:This is the Death of Maemo,if it really ever li

[omz13]

Nokia has won more design awards for phones than Apple, by a long shot.

Yes, but Nokia has been in the phone business for how many years compared to the short time Apple has been there... so its hardly surprising they have more awards.

Re:GNOME

[Super_Z]

I recently moved from KDE 3.5.* to KDE 4.2 on my desktop at work. I find it pretty stable and so far I'm happy with the components I have used. The big exception is Amarok which is .. not that good.

Re:This is the Death of Maemo,if it really ever li

[gnud]

All the auto generated code is boring, boilerplate stuff that you could easily write by hand if you for some sadistic reason wanted to.

Have a look at a moc_*.cpp file some time, it's not rocket science. But I'm glad I don't have to remeber the order of the (integer) values in the meta data array - moc generates it automatically.

Re:GNOME

[Daniel Phillips]

Pretty much everything I've done with Qt tells me KDE should be a much better desktop than Gnome. But the truth is that most of the large desktop distributions use Gnome, Ubuntu is much bigger than Kubuntu and same goes for the others. None of the big three hitters Firefox, OpenOffice or GIMP are KDE applications - ok not all are Gnome apps either but there's not many "killer KDE apps" around. Don't get me wrong, they're all perfectly okay but nothing really rocks the boat.

The think I like least about each of Firefox, OpenOffice and GIMP is the user interface, for which I blame GTK. For example, Firefox's application chooser dialog makes me want to slit my wrists.

Re:This is the Death of Maemo,if it really ever li

[ecki]

...this is going to draw in QT Embedded

What makes you think this?

... their bad, slow code ...

Care to shed some more light on this too?

Re:I know why..

[Carewolf]

Uh, maybe because Qt was bought by Nokia? They're the ones who decided to LGPL it, but they can do anything they want with it.

Not anything they want. They have to keep it open source due to the Free Qt agreement with KDE.

Re:GNOME

[drinkypoo]

These interfaces are really and truly not GTK+'s fault. GTK's main failing is making everything look like a modernized CDE. Since they have already long since skinned away the Motif-look I see this as not being a bad thing. A certain stage of GNOME's development seems to have been to supersede CDE, also.

On the other hand, the GNOME project seems to be developing a serious case of "our way or the highway". If you take gnome-panel out of GNOME, it doesn't work right any more, and they've gone to great lengths to try to prevent you from doing so. The logout window has options for logout and shutdown options which you cannot use together... what? I seriously cannot have one dialog with all my shutdown/logout options? Basically you can easily replace every part of the GNOME desktop with a surrogate just to get decent performance (I am SO tired of tracker bugs cratering my #$#$^@# system... it just keeps happening!) and so what exactly are you getting, anyway?

Re:This is the Death of Maemo,if it really ever li

[drinkypoo]

Motorola didn't sit on their ass, they went nine conflicting directions at once. The result is similar, but more expensive.

Maybe I'm just a snob but I've found every Nokia interface I've ever used to be terribly painful. Mind you, Motorola is as bad or worse.

Re:GNOME

[makomk]

When I'm moving a file it's very often a point to know when it's complete, and it irritates me to no end that every time I must click the notifications icon to know when it's really done.

I'm pretty sure the tray icon indicates the number of running transfers now. It also pops up a message briefly when the transfer finishes. (Of course, I'm probably using a newer KDE version than you.)

Re:This is the Death of Maemo,if it really ever li

[BlackCreek]

I absolutely agree with you that too many people write Nokia off the phone market due to nothing but "kool-aid". Either Apple or Google (android) kool-aid. Or the fact that they only know the US market, but have no idea of what happens elsewhere.

One point people seem to miss is that not everyone wants/can/will buy such ridiculously expensive phones to start with.

However.... I own a G1, have messed around with an iPhone, and played a lot with my GF's Nokia 5300 X-Press music. Nokia's current problem is that their phones are not designed around the expectation that the user has a "all you can eat data-plan". While this allows them to sell phones to people without a data plan. It has lead to a mess of connection settings, connection permissions, and syncing.

As a phone, and multimedia player the 5300 beats the shit out of the G1 (lousy battery, heavy). But Nokia really needs to clean their act regarding internet browsing, application download (ovi sucks!), and syncing.

Slides & Audio

[ultrabot]

Ok, it seems they got the audio recording of the talk up too.

Check it out at:

http://flors.wordpress.com/2009/07/05/maemo-harmattan-keynote-at-gcds/

Qt uses only 'nice' parts of C++

[chris-chittleborough]

The Qt designers don't just create widgets etc, they design components that are easy to program with. As part of this, they avoid stuff that requires the tricky/ugly parts of C++. For instance, you rarely need to explicitly delete objects, because their libraries use reference counting to automagically delete objects at the earliest appropriate time.

So it is easy for any good programmer to learn enough C++ to use Qt effectively.

(Actually, Qt uses an extended version of C++, implemented via a preprocessor. The extensions provide "signals" (like no-op methods) and "slots" (methods which can be connected to signals), plus a limited-and-very-useful facility for run-time widget class information. As usual with Qt, these facilities are just extensive enough make it easy to do the things most people want to do, rather than trying to provide everything that anyone might want.)

!News

[Zantetsuken]

Anybody with a hint of interest in the Maemo platform already knows this from Wikipedia and the project home page (at least 4 to 6 months)...

Re:GNOME

[BlackCreek]

KDE is a highly integrated, cohesive environment, whereas GNOME mostly is a desktop, a file manager, a few applets and a bunch of assorted applications that happen to use gtk for drawing their ui stuff.

Gnome is also a project with decent "Human Interface Guidelines". Which KDE lacks. Badly.

Re:This is the Death of Maemo,if it really ever li

[Anonymous Coward]

Is your Google broken? You lazy git!

Re:This is the Death of Maemo,if it really ever li

[keithpreston]

Qt isn't bad code and is probably faster the GTK. The previous product I did used GTK. The problems is QT Embedded. It wants you to use QNetwork and QSocket and QWhatever. Most of these are bad buggy wrappers on top of something else. The build system for QT embedded is bad as an entire system builder. As a whole QWS is not multithreaded safe. I can tell you that I personally have fixed 100-200(our team fixed more like 1000s)bugs, race conditions, threading problems and work arounds in Qtopia for our embedded product.

I'm not arguing over QT and GTK, either will do the job. However there is a lot of difference between QT the GUI toolkit and QT the system platform. In it's embedded form (Qtopia, QT Embedded) it is a pain to deal with. It is nothing more then a Demo that has never seen the light of day. If they only use the QT GUI toolkit with the current setup they have (X11, glib, dbus etc..) it might not be bad. However if they adopt QT Embedded the system platform, they will die a quick death.

Re:GNOME

[EsbenMoseHansen]

KDE is a highly integrated, cohesive environment, whereas GNOME mostly is a desktop, a file manager, a few applets and a bunch of assorted applications that happen to use gtk for drawing their ui stuff.

Gnome is also a project with decent "Human Interface Guidelines". Which KDE lacks. Badly.

You mean like these??

Re:GNOME

[EsbenMoseHansen]

What?

Firefox application dialog is exactly what you said it is, FIREFOX APPLICATION DIALOG, it is *not* the Gnome's application dialog... So blame Firefox.

On linux, they are GTK+ dialog boxes, which is not-exactly-but-almost the topic at hand.

Re:This is the Death of Maemo,if it really ever li

[dfghjk]

"The sad thing is that people have been screaming for this now for years, and Apple essentially delivered it first."

For your definition of "essentially" perhaps but Apple was far from first.

Re:GNOME

[BlackCreek]

Do you realize you just showed a link to a page last updated in 2004?

KDE may (formally) have HIG, but it is not as if anyone is doing anything with it.

Re:GNOME

[Anonymous Coward]

Ignorance is bliss, isn't it? You do know that GNOME has a HIG because it NEEDS it, in order to specify every toothgrinding little detail that is mostly done automatically in Qt/KDE... Congratulations to the GNOME brainwashing team for making a virtue of necessity so successfully that people think it's a weakness of the competition to not have one as extensive. :>

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:This is the Death of Maemo,if it really ever li

[Capt. Beyond]

all fine and dandy, except I do not have to work/edit with the moc'd code at all.

Re:I know why..

[vakuona]

Because then they own it, and can do whatever they want with it, include release non-free software based on Qt as copyright holders.

Re:Ubuntu?

[socceroos]

It gives them Canonical. How significant that is remains to be seen - but Canonical have been pushing hard for business partnerships of this ilk.

Re:skype?

[socceroos]

I'm going to bite - and then some.

I'm not sure if you're referring to the fact that the Linux version of Skype is not as fully featured as the Windows version, or if you think that the Linux version of Skype is riddled with bugs or if you just think that the look and feel of Skype on Linux is not up to scratch - but here are a few pointers:

> Skype doesn't have as much incentive to work on their Linux version of their client
> Skype was heavily married to the Windows API's in their version 3 and 4 clients meaning that cross-platform support is hard

If you've read those two bullet points then it should become clear that this has nothing to do with QT, but rather with Skype's internal practices and motivations.

Re:GNOME

[socceroos]

Amarok 2.2 is the release to hang out for. It will make you coffee, pay for your kids intuition and clean your car - or so they say.

Re:This is the Death of Maemo,if it really ever li

[socceroos]

If Nokia wanted full control then they wouldn't have just announced their GIT repositories. Everyone can effect the development of QT. Its interesting to see QT 4.6 progress - because there is a lot of downstream work being done and itches being scratched. I like it. =)

Re:I know why..

[noundi]

Haha what? You're talking about the move of a huge corporation from a consumers angle. No they didn't buy it because they thought it was "better". They bought it because they considered it profitable.

Lets blame GTK for everything

[WebCowboy]

Clearly it is the most evil code on the planet. It kills baby seals I heard. GTK is responsible for all that subprime lending that triggered this massive economic meltdown too, right? Geeeez.

The think I like least about each of Firefox, OpenOffice and GIMP is the user interface, for which I blame GTK. For example, Firefox's application chooser dialog makes me want to slit my wrists.

Yes, GTK's ancestry is tied to GIMP in some long dark distant past...but to blame shortcomings in ANY of those applications you mention on GTK makes about as logical and sensible as the statement I made above. First and foremost it doesn't matter if you use the greatest toolkit/framework/tools/etc in the universe, you can make an application that is absolutely horrible to use. KDE and KOffice weren't horrible, but to many they've lacked in features, design and usability at various points in time. The massive upheaval between KDE 3.x and KDE 4.x wasn't just done on a whim--their development communities went headlong into what they had to know was going to be a long, painful process to make KDE relevant again, and all through that time qt adherents would probably argue that qt was a more elegant platform throughout. The thing is, GNOME had the upper hand from a usability perspective for some time because much was done to make the environment more usable, perhaps at the expense of the DEVELOPER experience. Now KDE is back to being competitive in that space and some might say has the upper hand again because in addition to being competititve as a user environment it has a more elegant native development toolkit in the form of qt.

With Mozilla and Oo.org in particular your assertion makes no sense at all--both of those applications don't even natively depend upon GTK OR qt at all! Integration with GTK (or, at a desktop level, GNOME) or qt (or KDE) in both of these apps is a "retrofit" and it sometimes shows. Since FF and Oo.org both place a very high degree of emphasis on cross-platform support they may make sacrifices in native UI environment integration from time to time--or at least questionable choices. FF does have its won XUL-based alternative if I recall, so you can still get around any GTK limitations. However, there is much that could be done by FF developers to better use the GTK wigit, and GTK being GPLed there is nothing at all preventing mozilla developers from contributing to the GTK project to address shortcomings pointed out by their users. Perhaps the argument is there that GTK ought to rethink the design of its file chooser, but that isn't really an example of an overall GTK design/architectural shortcoming but rather a higher-level discussion.

Re:I know why..

[interval1066]

You have a point. Actually, I thought of it from an engineering angle. I simply don't know enough about business to make that connection. Hence, I'm not in that line of work.

Re:I know why..

[LWATCDR]

"Even though Qt is a nutty blob of nonsense."
We need a new moderation.
-1 mindless zealot.

I have worked with GTK and it is actually a good toolkit if you only want to work in c. The C++ bindings are iffy and I have not used Mono yet and may not ever due to a lack of time.
QT is also a very good toolkit and a much better choice if you want to work in C++.