So I've read and heard several people talk about how easy it would be to port over the QNX kernel to Android. So I was wondering, if it's so easy why isn't someone doing it? I mean if you think Google should do this it would make a great proof of concept and could maybe sway Google a bit. Now that's not to say that I myself think Google should do it, I don't really know much about kernels, but I'd be interested in trying it, if nothing else to see if benefits do exist.
The fact that QNX is a proprietary OS might be one of the reasons.
it became closed source when RIM bought it.
You mean porting entire Android software stack to run on top of QNX kernel ? (in order to run it natively). That is not that straightforward and Google would never do it anyway. You can't just "switch" to QNX kernel, it has completely different architecture.
Android is based on old fashioned monolithic Linux kernel, with all the stability, security and performance issues that come with it. QNX has a fantastic real-time microkernel. If Android was QNX-based, that would be a dream come true.
Related
So I've been having a discussion in another thread regarding the use of older versions of MSOS's on PPC. That spawned a question on my part:
Is there a development group here somewhere that is working on a Linux OS, or another OS for PPC?
Linux will run on just about anything, its' lightweight OS needs little memory and cpu power. So how hard would it be to design a light Linux based OS for a PPC?
Obviously it would take a group of people, much like those groups developing Linux distros and programs.
I think there is memory to be saved, and speed to be had. And if someone were smart enough to wrap a dialer and vendor agnostic connectivity around it, it would take off.
Any interest in this?
http://wiki.xda-developers.com/index.php?pagename=Xanadux
or android
or
http://www.google.co.uk/search?hl=e...H_en-GBGB243GB243&q=linux+for+pocket+pc&meta=
Wow, I'm disappointed.
There are hundreds of WM5 & WM6 custom ROMs' being developed by hundreds of top notch developers...... and only ONE Linux port?
very underwhelming...
You may also want to check out OpenMoko (http://www.openmoko.org) or just try and put together your own.
Splitter said:
Wow, I'm disappointed.
There are hundreds of WM5 & WM6 custom ROMs' being developed by hundreds of top notch developers...... and only ONE Linux port?
very underwhelming...
Click to expand...
Click to collapse
It's a whole different thing. All those roms you are talking about are just modifications of an existing OS.
The linux port amounts to building an OS from scratch, and it's a lot harder.
edzilla said:
It's a whole different thing. All those roms you are talking about are just modifications of an existing OS.
The linux port amounts to building an OS from scratch, and it's a lot harder.
Click to expand...
Click to collapse
Agreed.
Actually porting Linux to an HTC device wouldn't be so bad. Some people have found out how to flash ROM's theoretically without needing a bootloader even.
The problem really boils down to drivers for Linux. We can't even get proper video drivers working with the Kaiser under Windows Mobile (the proper drivers were never included, so video output is slowwww) though the hardware supports 3d acceleration! HTC denies our requests for hardware specs to develop our own. And this is trouble we're having with drivers for Windows!
Really it boils down to this hardware. This type of hardware being proprietary as you can get. You've got processors and controllers that are highly proprietary and the vendors are tied in to 100 different non compete non-disclosure agreements and can't provide specs. Even the qualcomm chips borrow code from broadcom -- which means qualcomm can't publish how those portions of their chips work! Microsoft then licenses code from these vendors with promises not to share source. HTC licenses code from broadcom and qualcom swearing not to publish it. Etc etc...
Now, your a Linux developer. How do you integrate drivers in to your kernel when the chip instruction set isn't even documented? Control codes aren't published? Reverse engineering is the only way, which can take years. Developers here have learned simple controls such as to change LED's or discovered the standard interface for USB/SD cards. That's about it.
It's hard for an open source OS to survive in a closed-spec hardware world. PC's are open and well documented and very standard. However, every phone is different, and different production runs may even have significant changes in internal hardware design.
It's really a waste of time to seek Linux on mobile devices until hardware becomes standardized. Which is never because companies like qualcomm and broadcomm via and others are not fans of open source. This is the market and those who dominate it.
If this saddens you, it should -- but it's just the way it is.
I read both on Twitter and online about Palm Pre's "WebOS" being based on the same kernel or structure that is designed for Android/G1.
Is anybody in the works on seeing if its possible to run WebOS on a G1?
Note: I guess I'm asking to not to replace the Android but to say that if its possible and that it can be done.
webOS runs on the Linux kernel, but nothing else beyond that is open source. How are you going to port an operating system to a different microarchitecture and hardware specification without its source code? Answer: you're probably not.
Why would you want to though?
It may be cute, but the platform has yet to prove itself. Developers are going to be cautious about developing for that platform due to how palm has so severely stagnated over the last few years. Unless the platform shows signs of really catching on, I would just ignore it.
I remember a while back when Google said that they were going to keep Android OS fragmentation at a minimum or I think it was not at all...
Yet along comes honeycomb a version of Android built specifically for tablets.
-Sigh-
What are your thoughts?
It's going to be very tricky to have tablets and phone with the same processing power, let alone have people wanting the same things from both.
I don't think it fragmentation at all, just a version built for tablets. I wouldn't want that on a phone <5".
I would like to see a source for that too because the whole point of the OS is for the carrier/manufacture to customize it. They are the ones that need to update faster.
fragmentation is heavy on android platform, let it be because of freedom of customization, lack of support from phone builder for updates, etc. No need to quote source, just read on it, there's plenty of info on that.
Well he said Google says but didn't give a source for it.
I know there are multiple versions out that are lower than AOSP, it's a fact. There are also features left out of recent iOS versions that are not available on older versions but because they have the same v# people don't seem to complain about it.
But there is ZERO reason to blame Google for not updating blur/touchwiz/sense when it is the Manufactures obligation to update the software with their skin.
Honeycomb has nothing to do with phones though. The OS is made specifically for tablets. Its like windows ce. Yea, its called windows, but because it only runs on smaller devices doesn't make windows fragmented. Same thing applies here. Honeycomb is specifically made for tablets.
The windows ce example is a bad one. Have u used a wall windows ce device (not windows mobile - ce/embedded edition windows)?
In any case as long as APIs are kept consistent there should not be any big problems.
The fragmentation is only problematic because of the way devices are supported, tbh...
Sent from my SGH-T959 using XDA App
How come they got Android 2.3 on the iphone but noone can get ios on android hardware?
I would think at least partly it's due to lack of trying (doubt many Android developers would want to use iOS)... but the thing is, Android is designed to run on a lot of different hardware, whereas iOS runs on specific hardware, so they can do lots of specific checks if they really wanted to to make sure it's running on a real iPhone. This is just my educated guess of course.
The iPhone currently has a handful of carriers and only 4 hardware configurations. I'd imagine that a fair amount of information about the internals gets leaked as well.
I am unaware of the specifics of android development on that platform, but the biggest advantage their developer community has is less overlap and redundancy in their efforts (lack of fragmentation). Also note that Android is open source-IOS is not, so we only tend to see iPhone OS installed on Apple hardware.
I also have my suspicions that their grass isn't really all that green. Personally, I tend to favor fragmentation, as it fosters innovation(but YMMV!).
You didn't ask my opinion though
Cheers.
Android is open source so it can be adapted to almost all hardware. iOS is closed source so to adapt it you'd have to reverse engineer. And that is illegal. It's that simple. And been asked before. You should try searching, reading and thinking, it's wonderfull, doesn't hurt and you might learn something.
Sent from my HTC Desire using XDA App
The reason's pretty simple. In order to effectively port an OS to any particular device, you really need access to the source code.
On the one hand, with Android, Google literally gives the source code away. You know that AOSP term everyone throws around? It stands for Android Open Source Project, and is the website (source.android.com) where anyone can download the full Android source code and do basically whatever they want with it.
Then, there's Apple. They guard iOS' source code vigilantly and litigiously -- I mean, their over-protectiveness even extends to what apps they'll let run on (non-jailbroken) phones. So needless to say, they don't make it easy to take their OS apart and port it to other devices. Really, they make it as hard as possible.
Great answers guys thanx!
Yeah! Thanks for the answer Ik Desire! You Rock!
Maybe I'm the only one, but I am a long-time fan of the now-dead WebOS, (formerly PalmOS). Back in 2011-2013, I probably had 6 or 7 different WebOS phones, and found the hardware and the software to be superior in user experience than any others. The problem was that that was the big explosion era for Android and IOS, so WebOS got pretty overshadowed, and HP, (the new owners at the time), decided to axe any further development, and ended up selling the OS to LG, who has developed it further for TVs and stuff, (I don't have all of the technical details straight, but that's a superficial summary). Anyway, recently, out of nostalgia, I pulled out an old Veer, and bought a new Pre 2, and went back into the WebOS forums realm to get a fix of my old favorite, and I was pleasantly surprised to see the work of the community to keep WebOS alive. There is an archive of hundreds of app files available on ftp, (all abandonware available for free - even the paid apps from HP catalog), ongoing development of apps and patches to keep WebOS working with all of the changes in technology, (including updates to Preware - the unofficial dev app catalog, started way back then), and LuneOS, an open developers version of WebOS, which can be built for modern phones, and which was just updated - https://pivotce.com
I think the XC is a perfect candidate for LuneOS, being small, currently affordable, (one upside to the short lifespan of official support), and having pretty good specs. Plus, there's already plenty of foundational development needed for supporting porting, (Lineage, AOSP, and even a partly-working Halium - https://github.com/Halium/projectmanagement/issues/103). For anyone familiar with building/porting, the work needed to be done to build LuneOS for XC is minimal.
My problem(s): I am not familiar with building/porting. I am happy to test and learn, but I also currently don't have any regular access to a Linux computer, and even if I did, it would take me a looonng time to figure out from scratch what someone else could probably do in no time, since all of the elements are in place, (I checked with WebOS devs, and they said with what's available, building for XC is '... mostly a matter of configuration').
So, anybody out there with an XC want to do a build for us?
@harryharryharry @oshmoun ...?
Sorry, I'm not even remotely close to having the capabilities to pull a stunt like this, I think I'm currently in the same boat as you (minus the linux part)
Having said that, here's hoping someone more knowledgeable jumps in and picks up this interesting project. I'm willing and able to test and provide feedback as well of course.