I would like to add that Maemo is much more game-developer-friendly. While in the past, you didnt have OpenGL (ES) , at least you had SDL and direct access to video memory and all other hardware resources. With SDL, you have zero-effort to develop for Maemo;
I've never written a single hildon line of code in my life (I cant say anything about GTK, as I use MonoDevelop with GTK# all the time), but I've been a Maemo game developer for 4 years. While it is not really wonderful, looks nice enough to me - http://batterypoweredgames.blogspot.com/2009/07... =-)
On the other hand, Android's java-only approach was a huge no-go for me.I would have to write everything from scratch. Why cant it at least support MIDP out-of-the-box? I know, there are a few solutions out there, but thats not enough for me.
2 my cents
Daniel "NeoStrider" Monteiro
· 3 months ago
Post-comment recall: I agree with your post scriptum, but It would be even better if you add WebOS to the scene.
WinMo x iPhone (UNIX!) x Android (Linux) x WebOS (Linux) x Maemo (Linux) dead: MOTOMAGX / EZX (Linux) Zaurus (Linux)
the picture looks very nice to me.
Pavlo Zahozhenko
· 3 months ago
Thanks for your comments, Daniel!
Currently I'm exploring Palm's WebOS development platform by porting a simple game to WebOS. As I see it now, WebOS has a great potential for developers, especially those with web development background. However, its main weak point is its closed nature and binding to a single company - Palm, which is simply too weak to compete with such monsters as Apple or Nokia. Until WebOS remains closed (and, I guess, it will remain closed forever, as it is the only hope for Palm to survive), I don't think it will become major competitor to either Android, Maemo, or iPhone OS. Most likely, WebOS will occupy a small but lucrative niche and represent an interesting alternative to major platforms. Anyway, it will contribute to Linux success on the mobile.
By the way, you have forgotten Symbian OS in your equation - while this platform is completely uninteresting from developer's perspective, it still occupies lion's share of smartphone market ;)
Daniel "NeoStrider" Monteiro
· 3 months ago
Writing the game logic itself on js is not a problem at all. indeed, its quite a pleasure. I do it with Unity3D for iPhone (day job). The problem is writing 3D math in js. (even writing it in java can be a problem. 3D math HAS TO BE C/ASM - not even C++! But you can always use it from C++). Does it have OpenGL ES?
Pavlo Zahozhenko
· 3 months ago
No, it lacks OpenGL ES. As far as I know, even WebOS itself doesn't use OpenGL ES internally.
Daniel "NeoStrider" Monteiro
· 3 months ago
yeah, but Symbian is dying. And mostly, we're talking about a diferent smartphone kind. Symbian phones tend to be somewhat limited and its operating system is very difficult to develop for. As in Android, you dont get direct access to video memory unless you play dirty and ruin your chances to run it on other devices. Now they do have OpenGL ES, but is a bit too late for that.
Just for you to know, this project of mine, started on Symbian, SDK S60 1.2 (testing on my then "N-Gage" - great phone, btw, but limited in the light of my current Motorola A1200i - Linux phone), but when Maemo was announced I was more than relieved to make the transition. On the N-Gage, running on top of all the graphic gizmos the engine achieved 0.5~1 frame per second for a 9 polygon scene. Now, on the 770 , for a ~3500 polygon scene (with BSP processing), I achieve respectable 5~10 fps. So, I wonder how good is WebOS for some more intense gaming.
Pavlo Zahozhenko
· 3 months ago
Surely, Symbian is very developer-unfriendly, and it is one of the main reasons of its decay. This may change when Qt becomes official development framework for Symbian, but it'll probably be "too little too late".
As for intense game dev on WebOS, I don't know what it is like yet, as I'm developing a simple puzzle game. I'm actually worried about WebOS capability in serious gamedev department: imagine writing a game in javascript, it's something like that.
Daniel "NeoStrider" Monteiro
· 3 months ago
oops. posted the reply to the wrong comment =-P
Mike Bradshaw
· 3 months ago
Some thoughts/responses:
* I would prefer Maemo to stay with ARM over x86 as; + ARM is designed for low power usage, x86 drags all the history and desktop baggage with it + Developing for multiple architectures can reveal subtle edge case bugs (always a good thing!
* I think the advice is to develop _NEW_ apps using QT as the UI layer rather than Hildon/GTK+. By using QT you also have the potential to release on the multiple platforms that QT supports and this includes S60 (the largest volume of smartphones currently ;) )
Daniel "NeoStrider" Monteiro
· 3 months ago
Starting fresh is always good, but I wonder if in its rush to catch up with x86, ARM is not bloating its cores. Also, the idea of using a separate development inviroment is better than having a plain userspace enviroment.
Reporter
· 3 months ago
I am interested in devices that run Maemo or Android largely for the fact that I can use the devices as wireless handheld touchscreen X terminals. Can you add some comments about the suitability of these operating systems with that in mind? I believe that Maemo ships with X, whereas Android doesn't, but that Android has one or more easy hacks to get X running on it.
Daniel "NeoStrider" Monteiro
· 3 months ago
Maemo is a great OS on itself. My 770 still works very nicely today (its a 2005 device!) and its system is very complete. Mine runs midnight commander and nano very well. I had to chance to even try running Mono on it. Indeed, I dont build debian packages all the time. I try my applications from the terminal first and then build a package.
Franklin Beansley
· 3 months ago
This article needs to be revised for spelling and grammatical errors. Also, paragraphs are your friend.
Pavlo Zahozhenko
· 3 months ago
Sorry for that, English is not my native language (I live in Ukraine and write mainly in either Russian or Ukrainian), so some mistakes are inevitable.
Name
· 3 months ago
excellent said!
b w
· 3 months ago
Actually you will be able to use QT on the N900 and Maemo 6 is coming in 2010. There The head of Maemo markiting talks about that at the end of this video http://www.youtube.com/watch?v=GzNbPAgcHMw
nip
· 3 months ago
some notes: The Qt libraries are available for Maemo, although Hildon uses GTK. This is very relevant if you discuss strategic positions. Nokia invests heavily into Qt and envisions it as a cross platform framework. You can develop WinMobile applications on an apple computer and port them to Maemo or symbian with ease.
Surprisingly Nokia recently announced a netbook running windows. Can;t wait to see at least dual boot with Maemo.
The larger install base can serve as a disadvantage, slowing down structural changes.
Pavlo Zahozhenko
· 3 months ago
Thanks for comment! I've got the point about Qt and reflected it in the blog post. Surely, it makes a big difference and brings new opportunities for Maemo application developers.
As for Nokia's netbook, I think it's just a test-drive on netbook market: putting Maemo on their first netbook would have surely resulted in failure. With this netbook Nokia will gain some experience on notebook market and [probably] proceed with Maemo netbooks later on.
yayati
· 3 months ago
Well some of the arguments above are off track, i guess the center of the discussion is mobile devices not gaming devices.
However i see maemo or similar OS ( may be moblin ) having the edge as well as being preferred by developers, who knows the dalvik VM could be ported easily onto maemo , and we have all the application stack from android to maemo!
Also end user perspective, which will define the success of technologies, a lot will depend on 2 other factors 1. Brand and 2. Hardware.
The more processing power, and more hardware capabilities or accessories means more attraction to buyers, how much would the Dalvik VM support the ever changing needs of hardware? Linux has show great appetite ( servers to embedded devices to any things supported ) for hardware drivers and can natively support many hardware changes and innovations to be seen on mobile hardware, so the advantage is OS's like maemo.
The next big question is brand, and which will be a great deciding factor in this war, I see for now its a war amongst equals Android ( with Motorola, LG, Samsung and mighty Google), Maeomo ( with current leader Nokia ), iphone OS ( with the promising and redefining market capable Apple ) and winmo ( blackberry and the likes , and also dominant Microsoft )
However my summary is: Linux + Interpretive lang ( ruby / python ) + C, C++ + Java > Dalivik VM based Java
Pavlo Zahozhenko
· 2 months ago
An interesting point: Dalvik VM is open-source and runs on both x86 and ARM Linux, so I think it is just a matter of time before it is ported to Maemo.
As for your worry about Android compatibility with newer hardware, it's not a problem at all, as what adapts to new hardware is Linux, Dalvik just runs on top of it, more processing power and memory simply means faster Dalvik's instances/more Dalvik's VM instances. Android runs even on x86 2+Ghz multicore processors, so porting it to something like Snapdragon or future ARM processors is a matter of simple recompilation and a few optimization tweaks. Actually, Android was designed for portability and vertical market from the start.
All Comments
I've never written a single hildon line of code in my life (I cant say anything about GTK, as I use MonoDevelop with GTK# all the time), but I've been a Maemo game developer for 4 years. While it is not really wonderful, looks nice enough to me - http://batterypoweredgames.blogspot.com/2009/07... =-)
On the other hand, Android's java-only approach was a huge no-go for me.I would have to write everything from scratch. Why cant it at least support MIDP out-of-the-box? I know, there are a few solutions out there, but thats not enough for me.
2 my cents
WinMo x iPhone (UNIX!) x Android (Linux) x WebOS (Linux) x Maemo (Linux)
dead:
MOTOMAGX / EZX (Linux)
Zaurus (Linux)
the picture looks very nice to me.
Currently I'm exploring Palm's WebOS development platform by porting a simple game to WebOS. As I see it now, WebOS has a great potential for developers, especially those with web development background. However, its main weak point is its closed nature and binding to a single company - Palm, which is simply too weak to compete with such monsters as Apple or Nokia. Until WebOS remains closed (and, I guess, it will remain closed forever, as it is the only hope for Palm to survive), I don't think it will become major competitor to either Android, Maemo, or iPhone OS. Most likely, WebOS will occupy a small but lucrative niche and represent an interesting alternative to major platforms. Anyway, it will contribute to Linux success on the mobile.
By the way, you have forgotten Symbian OS in your equation - while this platform is completely uninteresting from developer's perspective, it still occupies lion's share of smartphone market ;)
Just for you to know, this project of mine, started on Symbian, SDK S60 1.2 (testing on my then "N-Gage" - great phone, btw, but limited in the light of my current Motorola A1200i - Linux phone), but when Maemo was announced I was more than relieved to make the transition. On the N-Gage, running on top of all the graphic gizmos the engine achieved 0.5~1 frame per second for a 9 polygon scene. Now, on the 770 , for a ~3500 polygon scene (with BSP processing), I achieve respectable 5~10 fps. So, I wonder how good is WebOS for some more intense gaming.
As for intense game dev on WebOS, I don't know what it is like yet, as I'm developing a simple puzzle game. I'm actually worried about WebOS capability in serious gamedev department: imagine writing a game in javascript, it's something like that.
* I would prefer Maemo to stay with ARM over x86 as;
+ ARM is designed for low power usage, x86 drags all the history and desktop baggage with it
+ Developing for multiple architectures can reveal subtle edge case bugs (always a good thing!
* I think the advice is to develop _NEW_ apps using QT as the UI layer rather than Hildon/GTK+. By using QT you also have the potential to release on the multiple platforms that QT supports and this includes S60 (the largest volume of smartphones currently ;) )
The Qt libraries are available for Maemo, although Hildon uses GTK. This is very relevant if you discuss strategic positions. Nokia invests heavily into Qt and envisions it as a cross platform framework. You can develop WinMobile applications on an apple computer and port them to Maemo or symbian with ease.
Surprisingly Nokia recently announced a netbook running windows. Can;t wait to see at least dual boot with Maemo.
The larger install base can serve as a disadvantage, slowing down structural changes.
As for Nokia's netbook, I think it's just a test-drive on netbook market: putting Maemo on their first netbook would have surely resulted in failure. With this netbook Nokia will gain some experience on notebook market and [probably] proceed with Maemo netbooks later on.
However i see maemo or similar OS ( may be moblin ) having the edge as well as being preferred by developers, who knows the dalvik VM could be ported easily onto maemo , and we have all the application stack from android to maemo!
Also end user perspective, which will define the success of technologies, a lot will depend on 2 other factors 1. Brand and 2. Hardware.
The more processing power, and more hardware capabilities or accessories means more attraction to buyers, how much would the Dalvik VM support the ever changing needs of hardware? Linux has show great appetite ( servers to embedded devices to any things supported ) for hardware drivers and can natively support many hardware changes and innovations to be seen on mobile hardware, so the advantage is OS's like maemo.
The next big question is brand, and which will be a great deciding factor in this war, I see for now its a war amongst equals Android ( with Motorola, LG, Samsung and mighty Google), Maeomo ( with current leader Nokia ), iphone OS ( with the promising and redefining market capable Apple ) and winmo ( blackberry and the likes , and also dominant Microsoft )
However my summary is:
Linux + Interpretive lang ( ruby / python ) + C, C++ + Java > Dalivik VM based Java
As for your worry about Android compatibility with newer hardware, it's not a problem at all, as what adapts to new hardware is Linux, Dalvik just runs on top of it, more processing power and memory simply means faster Dalvik's instances/more Dalvik's VM instances. Android runs even on x86 2+Ghz multicore processors, so porting it to something like Snapdragon or future ARM processors is a matter of simple recompilation and a few optimization tweaks. Actually, Android was designed for portability and vertical market from the start.