Re: [maemo-developers] Porting Qt to Maemo
oh man, you're opening a can of worms with this one. I really wish Nokia had used Qt because we have a big library of apps for Qt/Embedded (or whatever it is called now), it just seems like Maemo is never quite done. At 02:51 PM 2/7/2007, Andrea Grandi wrote: Hello, today I did a little test with Qt and Maemo 2.2 SDK. I recompiled Qt 4.2.2 into the Scratchbox with maemo 2.2 installed and I was able to run a little Qt-HelloWorld inside the emulator. I wonder if it would be possible to realize a real port of Qt into Maemo and make it run on Nokia 770/800. In my opinion this could be possible. We should strip out parts that are not necessary, for example: examples, documentation, headers and possibly other stuff like opengl, variuos db drivers ecc we should try to reduce the size the most we can. Once this is done, we could also write application using Qt/C++ and not only GTK/C. What do you think about this? Anyone is interested in this project and want to help me? Best regards, -- Andrea Grandi email: [EMAIL PROTECTED] website: http://www.ptlug.org ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Porting Qt to Maemo
At 04:42 PM 2/7/2007, Andrew Barr wrote: That's just my preference. I don't see why anyone shouldn't try to port Qt to these devices...it would be an interesting technical exercise if nothing else. It would be nice if Nokia could open-source some more of the ITOS stack so that the community can have a little more flexibility in hacking at these tablets...regardless of your choice of GUI toolkit. you can't port, one is C and one is C++, it requires a total rewrite. Just my two cents, Andrew Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Maemo 2.0 API changes, signals and properties (was: ANN: Eagle)
I gotta say as a third party developer that the state of Maemo is rather frustrating, it seems like a far from done piece of work and a moving target. I've seen a lot of complaints on industry reviews of the device as well to the stability and selection of software included on the device. What's happening at Nokia to help with this and remove the perception that the device is a beta unit? thanks, shawn At 08:52 AM 4/19/2006, you wrote: On Tue, 2006-04-18 at 08:51 -0300, ext Gustavo Sverzut Barbieri wrote: Hello, Eagle (http://www.gustavobarbieri.com.br/eagle/) was ported to Maemo/Hildon (http://blog.gustavobarbieri.com.br/2006/04/eagle-in-maemo.html). In your blog you had mentioned this: However I opted to not port every component, like HildonColorButton or HildonNote because I think they're not well designed, they don't even provide signal changed, used by Eagle's DataWidget to persist data automatically. As API will change in Maemo 2.0, I won't bother with this until then. HildonColorButton and HildonNote APIs are not changing in any signifcant way, so don't let that stop you from including more widgets in the supported list. Hildon-related API changes are more or less limited to HildonApp and AppView and gtk_infoprint (and widgets no one was supposed to be using anyway) See http://maemo.org/maemowiki/HildonWidgets for a bit more details. HildonColorButton doesn't need a specific changed signal, it is already provided, though it's called notify::color and the callback signature is slightly strange (http://developer.gnome.org/doc/API/2.0/gobject/gobject-The-Base-Object-Type.html#GObject-notify) (The older version of generated API documentation misses some crucial bits like signals and properties, the new API has slightly better generated documentation in https://stage.maemo.org/svn/maemo/projects/haf/doc/api/hildon-libs/HildonColorButton.html) It's a less known feature in GObjects that all properties have implicit changed signals associated with them. It has the benefit that you don't need multiple foo-changed, bar-changed, ... signals for complex objects. So the lack of extra signal is intentional and we plan to use the same design in new widgets as well, see HildonProgram for example. -- Tommi Komulainen[EMAIL PROTECTED] ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Maemo 2.0 API changes, signals and properties
At 10:03 AM 4/19/2006, Marius Vollmer wrote: ext Shawn Gordon [EMAIL PROTECTED] writes: What's happening at Nokia to help with this and remove the perception that the device is a beta unit? We are hanging around on mailing lists and browse blogs, waiting for someone to tell us what to do. ;-) I'm engaged in private conversations with 2 people at Nokia involved with the 770 about our issues and they agree with everything, but we're still waiting for something to happen. My concern is that these conversations have been going for 5 months now with no visible progress. There could be progress behind the scenes, but I'm not privvy to it, so I'm expressing my frustration here. The other frustration is that we have a number of applications on other devices that we'd port to the 770 except that Nokia has given the impression that these are coming any time now, so we haven't done it, but the apps haven't shown up either. I think the device is really nice, but maemo itself just seems far from complete and it gives the whole experience a beta feel. Not to dredge up the whole decision of what to use on the device again, but Qtopia would have been a lot less of a hassle. ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Too busy to accept help? I'm not complaining
I've already got Nils slamming me privately because I dared to mention Qtopia, but let me provide some perspective as a company who was very successful with Qtopia and the Sharp Zaurus and what Sharp and Trolltech did both right and wrong that Nokia could learn from (I don't care if they use Qtopia at this stage, I just honestly think it would have been faster and cheaper than going the route they did, but I am not privy to the information that went in to making that decision). Now by the time the Zaurus was commercially available, my company already had a dozen products running on it, maybe more, and there was a big and healthy open source movement that was also producing software, I don't remember how many apps, but it was a good amount and grew very rapidly. What was done right: Sharp actually located about 50 companies and individual developers (of which we were part) about 6 months before the release of the device, flew us all to San Jose, gave us a limo ride to the hotel, put us up with food and lodging, gave us a day seminar on the device and gave us devices. They hired some people that were specifically meant to interface with the developers and actually were almost always available in IRC for immediate chat and feedback. They worked with Handango to create a web site where commercial and free applications could be hosted. I don't care for the site much, but at least it was a central repository that Sharp would point to. Trolltech hired a liaison to work directly with the embedded community and keep the line of communication open. What was not done right: Sharp in Japan wouldn't trust Sharp Americas decisions and really pulled the rug out from under them. One example was that while they got the device in places like Best Buy they never sent anyone out to the stores to educate the sales people, so consequently they steered people away from the device. Sharp Americas support team kept getting gutted and the contact people in Japan kept changing till the point that you could no longer get information and it basically killed off any interaction between Sharp and the developers to the point that I was actually told by Sharp that they didn't want third party developers to do anything for the device. Sharp Japan making major changes to the OS and backend engine without telling any of the 3rd party developers and we only found out after the devices were released and then documentation was thin or non-existent. Confusing licensing - the Opie and OZ initiatives were really pushed by one of the people at Sharp US to try and commercialize his own embedded system, the problem was with the licensing because if you strictly followed the license, then a commercial application could not be legally sold for a device running Opie and OZ (I don't want to get mired in this again, I spent a lot of time working on this with lawyers at the time, and this was the end result). It was confusing to the point that Trolltech couldn't even explain it. Those are some bullet points of what we went through. We still sell a good amount of software for the Zaurus and Archos every single day, even with these issues. My point here is not advocacy for one windowing toolkit over another, it is to illustrate what works and doesn't work in this environment. I'd be more than happy to have a really detailed conversation with Nokia on this and share more details of my experience. At 10:51 AM 4/19/2006, Kasper Souren wrote: The product is on the market for less than half a year. There are already tens of usable free software applications ported or created. That's pretty impressive for the first 'open product' of such a big company. I'm not complaining. I'm a pretty satisfied customer _and_ developer myself. Just a little thank you to all the Nokia folks who made this possible... ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Maemo 2.0 API changes, signals and properties (was: ANN: Eagle)
I think you're not appreciating what I'm trying to say, email is notorious for not getting the proper tone across. I'm not attacking anyone, I'm expressing my concern at the current situation. I would also say that it is not at all obvious to me that the developers are working hard. I do not in any way mean that as a slam against them, I'm just saying that I haven't seen anything that would tell me that. That said, I don't read all the email on this list either. I think the device is gorgeous, I do think that some of the UI design ideas are not at all intuitive though, and the decision on the MMC card was really not a good one, they are expensive, hard to find and do not go very high on capacity. No one is going to forgive the price point if it is too under powered for what they expect it to do, it could be $50, it doesn't matter. I'm a programmer and I was a reviewer for magazines for year as well, I know how these things are approached. If there isn't an option for a beefier unit with more features, then it is judged on its features and stability regardless of price. Nokia is famous for their rock solid hardware, so something that comes out and seems more like a test unit is going to upset a reviewer even more. I've got to say that this review http://www.washingtonpost.com/wp-dyn/content/article/2006/04/15/AR2006041500125.html is pretty much spot on, much as I like the device, from a reviewer perspective there is nothing I disagree with in this review, I had all the same experiences other than the network connection one. With the lack of software, it's really not clear what you should do with this device. Nokia (or maybe it was just the press) has made it seem that things like VoIP will be coming any day now, but they haven't. My company has all sorts of cross platform VoIP solutions, I would have ported it to the 770 months ago but I didn't want to compete with a free solution included with the device, but nothing has come out yet. My whole intent in getting involved with the conversations today is to try to do something productive because I'd really like to see the 770 (or its offspring) succeed. At 10:48 AM 4/19/2006, you wrote: On Wed, 2006-04-19 at 09:14 -0700, Shawn Gordon wrote: I gotta say as a third party developer that the state of Maemo is rather frustrating, it seems like a far from done piece of work and a moving target. Are you seeing API breaks, or are you complaining that improvements are happening? Is this an attempt to complain about progress while also demanding that progress happens? I've seen a lot of complaints on industry reviews of the device as well to the stability and selection of software included on the device. What's happening at Nokia to help with this and remove the perception that the device is a beta unit? I think it's fairly obvious that Nokia's developers are working very hard. They appear to be working on general stability, bug-fixing, and new functionality. Attacking them will slow them down. I've seen only one or two reviews that think if suffers from a lack of functionality, but they have compared it with either larger and more expensive products (such as mini laptops) or with less capable devices such as phones and PDAs which don't fullfill the main 770 use case (using the web). The other reviews seem to be very positive, and forgiving of a few glitches with the recognition that it's a first, that it's meant to do one thing well, and that it has limited resources to keep the price down. -- Murray Cumming [EMAIL PROTECTED] www.murrayc.com www.openismus.com Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Maemo 2.0 API changes, signals and properties (was: ANN: Eagle)
At 12:39 PM 4/19/2006, Philippe De Swert wrote: Hi, On Wed, 2006-04-19 at 12:10 -0700, Shawn Gordon wrote: snip 770 praise I've got to say that this review http://www.washingtonpost.com/wp-dyn/content/article/2006/04/15/AR2006041500125.html is pretty much spot on, much as I like the device, from a reviewer perspective there is nothing I disagree with in this review, I had all the same experiences other than the network connection one. There are indeed very valid points mentioned in the review. However I think we need to change something in the attitude that is taken towards this product. This is meant to be an open device, so some compromises had to be made. AFAIK it would be impossible to distribute the device with wmv codecs without some complicated agreements/demands from MS. And honestly I really want to avoid any DRM stuff. I'm totally against DRM, see my iTunes competitor www.mindawn.com that I started a couple years ago, we use FLAC and OGG formats. My company was the first one to make a portable ogg player by producing software to play Ogg on the Sharp Zaurus years ago. Also his complaints about flash are annoying. Nokia cannot help it that Macromedia is plain evil and writes hugely inefficient software. Their flash plugin eats up to 90% of cpu on my old P3 800Mhz laptop, which is absolutely horrible for the amount of displaying it needs to do. Also they do not support recent flash for GNU/Linux. It even gets worse if you look at flash support for ARM. I don't even think the reviewers websites would be viewable on a normal PDA either. On top of that I recommend anyone to read the EULA that comes with the flash plugin. Personally it kept me from installing the bloody plugin, and I probabely never will again unless I really need too. I hope that GNASH will be in working order soon. yes, FLASH blows, but this is about managing expectations, you have to control as much as possible what peoples expectations are for this device, and without a clear description to the market then they are going to have some standard expectations for the device, right or wrong, and that is what you can control. With the lack of software, it's really not clear what you should do with this device. Nokia (or maybe it was just the press) has made it seem that things like VoIP will be coming any day now, but they Also I don't know if expecting a PIM suite was realistic on a device called Internet tablet. Nokia never marketed this as a PDA. Though the reviewer did look at it as it were one. And who needs a PIM now that we have google/yahoo calendar... EVERYTHING has a PIM on it, you'd be amazed at how many people use a PIM on every dang piece of equipment they have. Simple example: I never wanted to do a PIM on the Zaurus because it already had one, but Lineo asked us to put together a proof of concept for a new calendar app for possible future devices, which we did and they really liked then they went under, so we decided to release it and people gobbled it up and demanded we do an address book, memo and todo apps and email - pretty much all of which were already on the device by default. To this day our PIM apps are our biggest selling applications. For the rest it is perfectly adequate as an internet device (at least for me). I would just love official ogg endorsement. You'd need to port Tremor or write your own integer replacement library for the floating point. Tremor isn't particularly efficient, but it worked ok with the 206mhz StrongARM when we used it. haven't. My company has all sorts of cross platform VoIP solutions, I would have ported it to the 770 months ago but I didn't want to compete with a free solution included with the device, but nothing has come out yet. Well that was your choice. And you can hardly blame Nokia for you not porting that application. And to be honest bringing out one of these solutions might have been beneficial for the whole development story. Please understand the point I'm making. It discourages the developers. Why will I spend time on that if I think Nokia is coming out with something shortly instead of doing something else? My whole intent in getting involved with the conversations today is to try to do something productive because I'd really like to see the 770 (or its offspring) succeed. Well I think everybody is waiting for some constructive action. I think there has been enough conversation. This is an open device, with a starting community. So if you come with a good idea, a tech-demo and a clear plan I am sure you will find people supporting it to make it a success. we're a commercial software company. We have plans for a few freebie apps, but our focus is going to be commercial software, to me whether it is open or not doesn't matter (no offense, it just doesn't), I'm interested in a viable product with a good user base and a way to access that user base to let them know about software
Re: [maemo-developers] Too busy to accept help? I'm not complaining
At 12:23 PM 4/19/2006, Philippe De Swert wrote: Hello Shawn, On Wed, 2006-04-19 at 11:59 -0700, Shawn Gordon wrote: I've already got Nils slamming me privately because I dared to mention Qtopia, but let me provide some perspective as a company who was very successful with Qtopia and the Sharp Zaurus and what Sharp and Trolltech did both right and wrong that Nokia could learn from (I don't care if they use Qtopia at this stage, I just honestly think it would have been faster and cheaper than going the route they did, but I am not privy to the information that went in to making that decision). Why would Qtopia be faster and cheaper? faster because it is done, has been used, refined, debugged and developed for for years, so other than device drivers in the kernel it wouldn't have taken hardly any time at all to get it up and running. cheaper - I'm assuming cheaper based on what I know of the licensing costs and the costs to hire a bunch of developers for years to develop and support the software. Nokia is not going to just rely on the open source community for something that they depend on, they will certainly have their own developers and these are going to be far more expensive than simply paying a small per unit license cost (I'm talking ones of dollars per unit). I am not trying to troll here but both have wide support in the Free Software community. (However in the embedded space GTK might have an edge. GPE is atm better supported and has more active developers than Opie for example. And I am not stating that because I happen to be involved with GPE, but because both Opie and GPe are involved with familiar we know about each other projects and we even co-operate on certain parts.) Keep in mind that Opie is simply a fork of the GPL version of Qtopia, so they get the advantage of all the things I spelled out above. Now by the time the Zaurus was commercially available, my company already had a dozen products running on it, maybe more, and there was a big and healthy open source movement that was also producing software, I don't remember how many apps, but it was a good amount and grew very rapidly. Well Nokia might not have had applictions readily available before the product was released. But I do remember porting an app in the maemo SDK before the device was actually available What was done right: 1.Sharp actually located about 50 companies and individual developers 2.They worked with Handango to create a web site where commercial and 3.Trolltech hired a liaison to work directly with the embedded community and keep the line of communication open. Ok. That indeed are valid points that could contribute to success with an open project. However *again* I see no reason why Qtopia would have been an advantage. In it's own right I believe the technical choice GTK vs QT(opia) has nothing to do with the success of these projects. As you point out political choices are much more important. Nokia has supported Gnome and has hired professional companies to support them as can be seen in Ari Jaaksi's presentation from Boston see (slide 7): http://www.kotiposti.net/jaaksi/ME9_LinuxWorld_2006_AriJaaksi_.pdf . So the one thing they really need to do is having somebody that can put time in co-ordinating the community and pass on all interesting developments to the people in Nokia who take the decisions. This last part has not yet been done. And if they manage to pick up the good things from the community it will become a killer product. So they definitely need to work on point 3. Apart from that the used toolkit point is completely moot. The 770 would probabely be as good/bad as it is now regardless of GTK or QT. I'm not espousing the benefits of one technology over another here, I'm simply making the point of how the business was and was not well handled in my opinion. Regards, Philippe -- | Philippe De Swert | | GPE developer: http://gpe.handhelds.org | Emdebian developer: http://www.emdebian.org | | Please do not send me documents in a closed | format.(*.doc,*.xls,*.ppt) | Use the open alternatives. (*.pdf,*.ps,*.html,*.txt) | http://www.gnu.org/philosophy/no-word-attachments.html Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
[maemo-developers] first commercial app from theKompany
We've just released our first commercial game for the 770 called tkcPanels. You can see the screenshots at http://www.thekompany.com/embedded/tkcpanels/screenshots.php3 and read about the product and purchase at http://www.thekompany.com/embedded/tkcpanels/index.php3 There is not a demo for the 770 at this time. Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
[maemo-developers] tkcPanels for Nokia 770
Our first application for the Nokia 770 is now available, it is a port of our tkcPanels game. You can see the screenshots at http://www.thekompany.com/embedded/tkcpanels/screenshots.php3 and read about the product and purchase at http://www.thekompany.com/embedded/tkcpanels/index.php3 There is not a demo for the 770 at this time. This is really a fascinating device, you should check it out - http://europe.nokia.com/nokia/0,1522,,00.html?orig=/770 Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] first commercial app from theKompany
At 09:56 AM 3/1/2006, you wrote: On Wed, Mar 01, 2006 at 09:45:48AM -0800, Shawn Gordon wrote: We've just released our first commercial game for the 770 called tkcPanels. You can see the screenshots at http://www.thekompany.com/embedded/tkcpanels/screenshots.php3 and read about the product and purchase at http://www.thekompany.com/embedded/tkcpanels/index.php3 Clicking on the 'add to cart' link, I got a You must choose a valid product to add to the cart error. :) You have to select the item, there are 3 or 4 versions of it so click on the combo box just above the 'add to cart' that says 'tkcPanel Versions' -r Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] tkcPanels for Nokia 770
I accidently sent the announcement to our general announce mail list to this list as well, instead of making the situation worse by sending yet another email and saying woops, I figured I'd just let it lie. At 10:39 AM 3/1/2006, Kalle Vahlman wrote: On 3/1/06, Shawn Gordon [EMAIL PROTECTED] wrote: Our first application for the Nokia 770 is now available, it is a port of our tkcPanels game. You can see the screenshots at http://www.thekompany.com/embedded/tkcpanels/screenshots.php3 and read about the product and purchase at http://www.thekompany.com/embedded/tkcpanels/index.php3 There is not a demo for the 770 at this time. This kind of announces/advertisments would probably belong to the maemo-announce list, if even there (I'm not sure about the policy, the listinfo page states that any announcement related to Maemo but that might be a little broad). This is really a fascinating device, you should check it out - http://europe.nokia.com/nokia/0,1522,,00.html?orig=/770 I'm pretty sure anyone subscribing to maemo-devel would already know that ;) -- Kalle Vahlman, [EMAIL PROTECTED] Powered by http://movial.fi Interesting stuff at http://syslog.movial.fi Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
[maemo-developers] screenshot tool?
Hi, We've been hunting around for a tool to take screenshots on the device but haven't found one. Any suggestions? thanks. Best Regards, Shawn Gordon President ProgRock Records www.progrockrecords.com www.mindawn.com (949) 713-3276 --- Tune into the best progressive rock station on the planet - ProgRock.com at http://boa.mediacast1.com:9288/prog1.ogg ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] RFC: Qtopia Nokia N770 (maemo) porting layer
I've been watching this thread with interest, and I have some feedback for you. Qtopia is about a 3mb footprint, it is pretty small, has its own framebuffer and doesn't rely on X, this is for Qt2x. Qt3 was never made in to an embedded version because of its size. I'm told that Qt4x was modularized better so it could be used for embedded again, I don't think anyone is doing it yet. OPIE is a problem from a licensing and compatibility point of view. OPIE is a fork of the GPL version of Qtopia and as time has gone on it has become less compatible. The other issue is the license, you cannot bundle a non-GPL app with OPIE and it is technically a GPL violation to use a non-GPL app with it because the application is now using the GPL versions of the libraries. The other problem for a commercial software company is that you can't rely on the fact that it is going to be available unless of course you have found some way to static link everything together, it would be like running a Qt app under GNOME basically. I have a huge investment in Qtopia development, nothing I'd like better than to be able to just run them on the device without having to rewrite the UI, but short of Nokia changing their mind and licensing Qtopia, any effort is just going to remain in the realm of the hobbyist IMO. At 11:11 AM 11/8/2005, you wrote: 2005/11/8, Clemens Eisserer [EMAIL PROTECTED]: ...which makes these sound a bit funky. If you use *anything* GUI-related that is already implemented, you'll drag Gtk+ in. Which will mean that you need to have both Qt and Gtk+ libraries in memory at the same time, which I think is not too hot on the current hardware. From this point of view I think GTK2 was the completly wrong choice as Maemo's GUI toolkit. Its slow (mameo contains even a hacked version which tries to speed it up a bit) and heavyweight (megs of code splitted in many different shared-libs), but it was choosen for compatibility as the whole Xserver based approach. (fox-toolkit or fltk are much more efficient) If Maemo would get a hacker every time somebody said that... I assume Qt has not got megs of code then? The Qt site boasts that: The Qt Class Library is a growing library of over 400 C++ classes [..] That doesn't sound lightweight to me ;) Gtk+ is, in my opinion, fast enough for the 770, the bulk of the performance issues come from the limited memory from which the (AFAIK non-gtk) system services occupy a fair share and from the fact that there is virtually no acceleration for the graphics. I think with this decision in mind a port of QT would not be that bad either since it would allow running apps on your Maemo powered device which would not be able to run otherwise or would not look that good. I'm all for a Qt version, diversity is good. It's having both of them in memory at the same time that is bad. Which also means that running programs for both at the same time would be bad too. And as said, most of stuff already done is in Gtk+, so you'd effectively have to write the GUI bits from scratch. But hey, in case you haven't noticed, Maemo is free software ;) -- Kalle Vahlman, [EMAIL PROTECTED] Powered by http://movial.fi ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Gtk vs. Qt
Kate, your initial work looks very promising, but I'm curious why you're doing a port of Qt itself instead of Qt/e, which is significantly smaller (less features of course)? Lorn mentioned qt/e4, I don't know what the state of that is currently, but if that's what get's running in Maemo then we might as well port the UI to Hildon since porting to qt/e4 is probably going to be just about the same effort and we'd rather be native. Basically there is a huge library of free applications and our commercial applications that run on qt/e2 that ships with all the Zaurus's (Zaurii) so any effort to get some emulation working would be most effective if it targeted the platform that is currently in the largest deployment. I'm apparently jumping in to this whole topic late, so I don't know what has been discussed before, but I think we're the only hardcore commercial developers of qtopia applications around, so we have some specific ideas in this regard. best, shawn At 05:15 AM 7/26/2005, Kate Alhola wrote: ext Eero Tamminen wrote: Hi, On Maemo platform I regard Qt a bit like libSDL as I think both of them to have similar problems: - User interface looks different (different colors, pixmaps, font family size and they don't change when device theme is changed) - User interface works differently (no special widgets for touchscreen usage, uses menubar to open menus instead of a titlebar) - Do not integrate with the input method (in Maemo this is integrated with the individual widget usage so that input method comes up automatically only when needed and goes away when not needed) - Naively takes pointer/keyboard grabs unnecessarily and/or doesn't release them when required (in some rare cases could end up locking the UI) - User cannot switch to SDL application through Task Navigator (should be easiest to fix) I.e. there would be quite a lot work to integrate Qt library properly to the Maemo platform in addition to it taking a lot of additional memory as Qt libraries wouldn't be shared with the other application Also the Qtopia emulation layer would be usefull for porting Qt/Qtopia applications. There is work to do but it looks a like that there is lot of interest to have possibility to port QT / Qtopia applications and so there is good reason to do this work. I think that it is much more usefull to do this port as open source project and produce shared library that can be used multiple applications instead that every app developper will make their own port and spend lot of memory with statically linking. I just checked size of native arm libqt-mt.so.3.3.4 and it was 10917826 bytes ( 10Mbyte !! ) It is large and may be that many Qt applications need to be in RS-MMC card but still if some application is written with Qt, there is only two choices, run it even it uses lot of memory and has some diferencies in look and feel or then not run it all. Of cource the major things with Maemo intergration should be resolved like this input methods and pointer/keyboard issues that you mentioned. Also look and feel compatilble themes would be nice thing to have. Kate ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
RE: [maemo-developers] Gtk vs. Qt
At 08:26 AM 7/26/2005, [EMAIL PROTECTED] wrote: Hi, Kate, your initial work looks very promising, but I'm curious why you're doing a port of Qt itself instead of Qt/e, which is significantly smaller (less features of course)? Maemo (and therefore the Nokia 770 device) uses X11 and there is no place for Qt/e which does not use X. Qt/e emulation to get some software running on the maemo and the current maemo compliant device is a great idea if it will work, if you ask me. However, replacing the X11 with Qt/e / qtopia would be really bad idea (IMHO). you've actually done the work, so you know more than I am, I'm simply throwing out questions based on my understanding. My intent has never been to replace X11 or Maemo, but to have a qtopia app run in it, rather like when we sell Qt apps and someone runs them on a GNOME desktop, we include the libs required for it to run. You miss some things like cutpaste in this environment due to no shared clipboard (at least last time I checked). Now what immediately comes to mind is that your file dialogs will probably look and work differently, but even on the Zaurus we've always had our own look and feel in our apps. I have compiled both Qt3/X11 and Qt4/X11 for maemo. I haven't tried to put the Qt4 to the device yet since it is quite huge with the default options. We've done a little Qt4 porting on the desktop so far and it's a heck of a lot of work. Even moving Qtopia apps to Qt is not particularly trivial, it kind of depends on the function of the app though. Best Regards, Karoliina Salminen Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
[maemo-developers] Nokia developer contact?
Hi, Seems that everyone I knew at Nokia from the Media Terminal days is gone, and we'd like to talk to them about porting our library of applications to the 770 (www.thekompany.com/embedded) - given that we use Qt/e-Qtopia, I've been watching the discussion on that topic with interest - has anyone tried taking a Qtopia app and getting it to run by just including the Qtopia libs required for it? Aside from that, a contact at Nokia would be really appreciate, I've seen some names around but no email addresses. Feel free to reply privately if needed. Thanks. Best Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Nokia developer contact?
At 01:26 PM 7/25/2005, Mattias Schlenker wrote: [EMAIL PROTECTED] wrote: Not unless you plan on porting the applications to Hildon Gtk. Perhaps I'm being misunderstood. KDE is a totally different beast than Qtopia, I don't think it makes sense to try and port that myself, and I'm not trying to replace a toolkit or compare environments, what I'm trying to determine is if there is a realistic way to have our Qtopia based apps run on Maemo as an intermediate step to a native port so we can get our apps out quickly. Oh no. Not another flamewar, please. Both Maemo and Qtopia are good (if not great) examples of free software. We had discussions on porting Qt and flamewars on the best environment for the N770 before and did not find a solution that satiesfies all people debating here. We do not have the RAM for supporting both Qtopia/KDE and Maemo/Gnome at the same time, but taking a subset of the other library, integrating the app and porting it this way is a good thing. As much as I do *not* like the K Desktop Environment, I like Trolltech's efforts of providing a portable library for any OS. Regards, Matt PS: I might just be some kind of car mechanic (and even worse journalist) tinkering around with his own distribution (called lesslinux), so do not take me too seriously... :-) -- Mattias Schlenker / Tel 0851 9441369 oder 0160 7352988 Freyunger Str. 42 / 94034 Passau / http://mattlog.schlenker-webdesign.de/ http://rura-penthe.schlenker-webdesign.de/steinchenspiel/c.php ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
Re: [maemo-developers] Nokia developer contact?
At 02:17 PM 7/25/2005, you wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Shawn Gordon wrote: At 01:36 PM 7/25/2005, you wrote: Shawn Gordon wrote: At 11:12 AM 7/25/2005, you wrote: Shawn Gordon wrote: Hi, Seems that everyone I knew at Nokia from the Media Terminal days is gone, and we'd like to talk to them about porting our library of applications to the 770 (www.thekompany.com/embedded) - given that we use Qt/e-Qtopia, I've been watching the discussion on that topic with interest - has anyone tried taking a Qtopia app and getting it to run by just including the Qtopia libs required for it? Aside from that, a contact at Nokia would be really appreciate, I've seen some names around but no email addresses. Feel free to reply privately if needed. Thanks. Do you mean rebuilding qtopia to use QT/x11, or bypassing X and using QT/e or (*gasp*) qvfb on X? Qt/e has its own framebuffer and doesn't use X, which is part of the appeal for it on embedded devices, I think the footprint is about 3mb (Lorn, correct me if I'm wrong) - for my purposes I don't care about X, I'm just trying to figure out if we can kind of carry our own environment along and have it work and look reasonable within the existing framework, I'm sure someone will port Opie to the device at some point and try using our applications on it, and some of them might even work like that. Our applications are typically well partitioned between the presentation and application layer, so a port isn't terribly hard, but anything to make the process quicker is a good thing from our perspective. disclaimer: this is my opinion, and I am not employed by nokia, movial, kernel concepts or any other company. Koen, seriously, have a tall cool drink of water - you are way off base on what you are accusing me of doing. I've got no idea what Qt/e propoganda you are referring to, my company has been happily making and selling applications for the Sharp Zaurus since before it was officially released, coming on 4 years now, with over 30 products. I'm not here espousing it to be better or worse than anything else, we are very aware of what it can and cannot do, but that is not at all the point of the discussion that *I* am trying to have. You are being rude as hell with your This does that DEAL WITH IT type comments, they are uncalled for and out of line. Try reading what I actually wrote. I sincerely apoligize for being rude. I just got pissed of by the fact that QT/e got dragged into the discussion again. The opie project proves that you can support X11 by changing a config option. I've used several apps from you company, which were very nice apps. I just found the qtopia env to limiting. Couple that with Lorn spreading FUD on opie-devel (see the QT4 thread) and I get a bit paranoid. I stopped paying attention to all the zaurus, opie, openzaurus discussions years ago, so I have no idea what might be getting discussed. All I really know is that supposedly the next major release of Qt/e is going to be based on Qt4 but until there is a device using it, it doesn't matter to me. The point you seem to be missing in everything I wrote is that I stated the fact that our apps are Qtopia, I didn't say anything else about it or rewriting the rom or making Qtopia run on the device. I just want my apps to run in the current environment and if there is a way for me to do that and leave them essentially unchanged as a first cut, then so much the better for us to get software out by the release of the device. I could care less about any of the holy wars about what everyone likes, we did our technical review of toolkits years ago and made our decision and we've been happy with it, I don't need to convince anyone else and I'm not trying to. I have to say that this is the nastiest developer mail list I have ever joined, We are still celebrating a QT/e free device and platform (and we are probaly gnome zealots too), on a *gtk* centered mailinglist. This is no excuse, but I wouldn't expect a parade and a visit from the president when saying you want to use QT/e. I really don't get your animosity. Every toolkit has its plusses and minuses, but as far as I know, Qtopia is the only embedded desktop on embedded linux that is being used in the real world on real devices at the moment, it must be on about 10 different Zaurus and the Archos and some other things. I'm not aware of any shipping hardware that uses Maemo at the moment. This is not propaganda or anything else, just a statement of fact. Qt/e works and it works very well, but again, it's not germane to this list or to my question, so I don't care about anyones zealotry, I have problems that I need to solve and I'm trying to find the best solution in terms of time, money and effectiveness. Again my apologies for being rude, and I sincerely hope you can find a way to run TK apps properly inside the factory flashed image
Re: [maemo-developers] Nokia developer contact?
At 02:24 PM 7/25/2005, you wrote: I really think that this place isn't the right place for asking your question. It's the perfect place as evidenced by your well thought out response below: The difficulty involved in porting Qt based apps to Gtk can vary a lot. Personally, If I were you and I _really_ wanted to do it. I'd try to get libqt to work under the scratchbox environment. Then package it as statically linked to your application executables. Hang on, isn't that against Trolltechs UA? Anyhow. Remember there is a difference between the open source version and the commercial version, we have a commercial license, which might still be a problem, I'm in the research phase right now. You're part of a business that has chosen to specialise in writing apps for embedded devices (I'm guessing here). You chose Qt as your platform, which was shipped with the zaurus. Now, tell me if I'm wrong. I'm guessing you plan on trying to sell applications to users of the N770 when it gets shipped. Cool, nothing wrong with that. It isn't however a good idea, in my opinion, to write apps that: a) Will look crap on stock devices. (Unless Qt can be hildonised) b) Do not look like other apps on the device. c) Don't interface with the core platform. well, we don't know that any of those items will be true yet do we? That is the point of researching it. First we find out if it is a viable short term option technically, then we do a proof of concept to see how it works in reality. If it doesn't look and work naturally on the device (like most Java apps on the Zaurus) then it won't make sense to do it that way. So, personally. Even though the wording by koen may have sounded hard. Unless you plan on helping/starting/doing some Qt modifications to make it interface, at least a little, with hildon/libosso. who knows, we might - like I said, this is all research right now You should really get your development team to read some docs at 'http://www.gtk.org/'. It took me less than a week to get good enough to write apps in Gtk + Hildon + libosso. oh, we know all about Gtk, but take a look at www.thekompany.com/embedded and you tell me what is going to be faster, rewriting all those apps or have a way to make them run 'as is' and react naturally on the device. From the user experience perspective they don't care what OS is on it, or the toolkit used to get there, they want apps that do what they want, are easy to use and typically attractive. That is what my focus is. Maybe in a week we'll find that there is no reasonable way to have our apps run 'as is' and we will have to start porting. Like I said, our presentation layer is nicely partitioned from the transaction layer for the most part, so we would typically just be rewriting the UI, which might be relatively trivial, but we won't know till we get in to it deeper. ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers Regards, Shawn Gordon President theKompany.com www.thekompany.com www.mindawn.com 949-713-3276 ___ maemo-developers mailing list maemo-developers@maemo.org https://maemo.org/mailman/listinfo/maemo-developers
