Active Topics
-
Is there a section to talk about Java ME? Apps, etc. (0)
to General by Kalatti - 2 days, 6 hrs ago -
Extra softwares in Sailfish using CLI, repositories, etc (117)
to SailfishOS by nieldk - 4 days, 3 hrs ago -
Firefox with Leste (6)
to Maemo 7 / Leste by teroyk - 5 days ago - more...
m4r0v3r |
2012-01-29
, 16:28
|
|
Posts: 1,746 |
Thanked: 1,832 times |
Joined on Dec 2010
|
#11
|
why not ask the Mer team for the SGX drivers? am not sure if they can give it out, but it wouldnt hurt.
|
|
2012-01-29
, 19:40
|
|
Posts: 249 |
Thanked: 277 times |
Joined on May 2010
@ Brighton, UK
|
#12
|
Originally Posted by m4r0v3r
A reasonable idea. IIRC, only Stskeeps has access under NDA to the source though, and there may be restrictions to only allow Meego DE (and now Mer) to use them. I'd like to be pleasantly surprised.
why not ask the Mer team for the SGX drivers? am not sure if they can give it out, but it wouldnt hurt.
Not to put words into their mouths...but there's also the not inconsiderable task of convincing the Mer guys to do some work on a sister project when they're probably prefer people just switched wholesale to theirs.
...at the very best we'd still be trading a known binary blob tied to old kernel interfaces with an unknown binary blob tied to newer kernel interfaces. They cannot release the source due to the NDA.
| The Following 2 Users Say Thank You to mr_jrt For This Useful Post: | ||
 |
|
2012-01-29
, 19:54
|
|
Posts: 1,671 |
Thanked: 11,478 times |
Joined on Jun 2008
@ Warsaw, Poland
|
#13
|
Originally Posted by mr_jrt
There's no restrictions on the SGX drivers (not even 'don't use outside nokia devices' or 'no commercial usage'), but you would have to use Debian armhf to use them as they're in hardfp ABI. If I was you I would use the new drivers, the kernel drivers are open and we support vsync. Userland is naturally closed, but redistributable.
A reasonable idea. IIRC, only Stskeeps has access under NDA to the source though, and there may be restrictions to only allow Meego DE (and now Mer) to use them. I'd like to be pleasantly surprised.
Not to put words into their mouths...but there's also the not inconsiderable task of convincing the Mer guys to do some work on a sister project when they're probably prefer people just switched wholesale to theirs.
...at the very best we'd still be trading a known binary blob tied to old kernel interfaces with an unknown binary blob tied to newer kernel interfaces. They cannot release the source due to the NDA.
BTW, Nemo is the project that contains the SGX drivers and handset UI, not Mer, Mer is a 100% open source core. Nemo builds on top of Mer.
Just out of morbid curiousity, why not base on Mer - what's the problem there that you see? (I'm not going to try to convince you either way). And I don't want to hear 'It's not Debian', give me some reasonable arguments - I've done the exact same path you guys are going down at, back in old Mer and Deblet, so I know what I'm talking about
I'm doing a talk next week at FOSDEM where one of my main points is that even the most open mobile project has the same kind of basic system work burden as a commercial project does. And you're going to have exactly the same kind of burden as we do in Mer, except we already took it upon us and have finished most key elements.
__________________
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
Last edited by Stskeeps; 2012-01-29 at 20:11.
|
|
2012-01-29
, 20:35
|
|
Posts: 249 |
Thanked: 277 times |
Joined on May 2010
@ Brighton, UK
|
#14
|
Originally Posted by Stskeeps
That's awesome news! Thanks.
There's no restrictions on the SGX drivers (not even 'don't use outside nokia devices' or 'no commercial usage'), but you would have to use Debian armhf to use them as they're in hardfp ABI. If I was you I would use the new drivers, the kernel drivers are open and we support vsync. Userland is naturally closed, but redistributable.
Originally Posted by Stskeeps
I know and respect your efforts greatly. Despite being a relatively late comer to Maemo, as I've dug deeper your fingerprints are everywhere, and usually in quiet important areas. You're quite the asset to any project in my opinion.
Just out of morbid curiousity, why not base on Mer - what's the problem there that you see? (I'm not going to try to convince you either way). And I don't want to hear 'It's not Debian', give me some reasonable arguments - I've done the exact same path you guys are going down at, back in old Mer and Deblet, so I know what I'm talking about
...however, I can't give you what you want in a form you'll find acceptable I suspect. I see great value in having Debian as the primary upstream vendor. You do not. There we are. Dismissing an opponent's argument out of hand as an unreasonable argument just prevents a proper discussion. We'll just have to agree to differ.
The slight irony is that your final paragraph states:
Originally Posted by Stskeeps
...and this is precisely one of the primary values I see in Debian as an upstream vendor and parent project. But there we go. I'd be more than happy to pull individual technical things from Mer, but given I find the architectural differences Maemo has from Debian that I've encountered irritating, Mer has an uphill struggle there.
Just out of morbid curiousity, why not base on Mer - what's the problem there that you see? (I'm not going to try to convince you either way). And I don't want to hear 'It's not Debian', give me some reasonable arguments - I've done the exact same path you guys are going down at, back in old Mer and Deblet, so I know what I'm talking about
I'm doing a talk next week at FOSDEM where one of my main points is that even the most open mobile project has the same kind of basic system work burden as a commercial project does. And you're going to have exactly the same kind of burden as we do in Mer, except we already took it upon us and have finished most key elements.
Cheers for the good news though! Much obliged.
| The Following 5 Users Say Thank You to mr_jrt For This Useful Post: | ||
|
|
2012-01-29
, 20:51
|
|
Posts: 1,671 |
Thanked: 11,478 times |
Joined on Jun 2008
@ Warsaw, Poland
|
#15
|
Originally Posted by mr_jrt
As I stated, I won't argue. But I will leave you with one lesson that I did learn from old Mer and Deblet work - and I believe you'll end up with the same lesson as many others have.
The slight irony is that your final paragraph states:...and this is precisely one of the primary values I see in Debian as an upstream vendor and parent project. But there we go. I'd be more than happy to pull individual technical things from Mer, but given I find the architectural differences Maemo has from Debian that I've encountered irritating, Mer has an uphill struggle there.
There is very good reasons why Maemo is like it is and not just a UI slapped on top of upstream Debian.
It bends down to how the software performs in a mobile setting: the memory footprint, the wakeups software does, the storage footprint. Maemo and MeeGo did a lot of slicing and dicing that was needed to provide the quality mobile products you see now.
And you'll end up doing the same work because you'll realize it's necessary. I was foolish and didn't believe in the choices they made as it was not looking like standard Linux systems anymore. And now I understand them.
Knowing those reasons and how to deal with them is what is key to success, good performance in mobile OS'es, battery life, responsiveness, etc.
For the same reason, we're not basing on Fedora in Mer. The distributions that we love and admire are not mobile OS'es.
If you do succeed in making a Maemo re-make, in Debian that works properly and sanely for mobile, that's great. There is a need for that as no-one else has been succesfull in doing so in the open until now*
(* Tizen is Debian based, BTW, but I wouldn't base even an alarm clock on that published stack)
Anyway.. starting sometime after March, I'm returning to being just a community member around here, with no special accesses or privileges - I have to move on some day too
. Mer will continue to go on as there's a lot of open and commercial interest in it as well. Nemo will continue as well. And I'll be in those.Good luck with your project, if you need any technical advice, feel free to ask.
And set as your first priority to seperate hardware adaptations and your mobile core. Oh, and turning off the screen automatically when idle.
__________________
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
Last edited by Stskeeps; 2012-01-29 at 20:55.
|
|
2012-01-31
, 00:09
|
|
Posts: 1,513 |
Thanked: 2,248 times |
Joined on Mar 2006
@ US
|
#16
|
Originally Posted by Stskeeps
What "new" drivers are you referring to?
There's no restrictions on the SGX drivers (not even 'don't use outside nokia devices' or 'no commercial usage'), but you would have to use Debian armhf to use them as they're in hardfp ABI. If I was you I would use the new drivers, the kernel drivers are open and we support vsync. Userland is naturally closed, but redistributable.
There is very good reasons why Maemo is like it is and not just a UI slapped on top of upstream Debian.
It bends down to how the software performs in a mobile setting: the memory footprint, the wakeups software does, the storage footprint. Maemo and MeeGo did a lot of slicing and dicing that was needed to provide the quality mobile products you see now.
I sometimes think we might be better off letting go of the cellphone hw adaptations and target a N8x0 era solution consisting of a (let it be closed because we can't get it open anyways) small cellphone w/ bt pairing to open source tablet. That opens up the options a bit. I know that is limited and puts a lot of people off, but better a small success than a large failure.
Anyway.. starting sometime after March, I'm returning to being just a community member around here, with no special accesses or privileges
|
|
2012-01-31
, 05:27
|
|
Posts: 1,671 |
Thanked: 11,478 times |
Joined on Jun 2008
@ Warsaw, Poland
|
#17
|
Originally Posted by SD69
The SGX drivers that are practically similar to the ones in Harmattan that we also have working on N900 and published as part of the Nemo/MeeGo efforts, which have improvements that the Fremantle SGX drivers does not have.
What "new" drivers are you referring to?
True that mobile OS is not easy to get to from desktop OS. I assume when you say Maemo, you mean fremantle. And when you say mobile, you mean cellphone.
I sometimes think we might be better off letting go of the cellphone hw adaptations and target a N8x0 era solution consisting of a (let it be closed because we can't get it open anyways) small cellphone w/ bt pairing to open source tablet. That opens up the options a bit. I know that is limited and puts a lot of people off, but better a small success than a large failure.
Good to have you back.
N8x0 and especially 770 shared the same issues. That's why Maemo is the way it is.
Back is a strong word and it's more likely I won't be spending terribly much time on N8x0, N900 or even N950/N9 except for my own personal interests except for anything that comes indirectly through Mer and Nemo efforts on top of hardware adaptations for these devices.
__________________
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
As you go on to other communities, remember to build them around politeness, respect, trust and humility. Be wary of poisonous people and deal with them before they end up killing your community.. Seen it happen to too many IRC channels, forums, open source projects.
Last edited by Stskeeps; 2012-01-31 at 05:34.
|
|
2012-01-31
, 08:40
|
|
Posts: 1,397 |
Thanked: 2,126 times |
Joined on Nov 2009
@ Dublin, Ireland
|
#18
|
Originally Posted by Stskeeps
That is precisely what I think. All those devices already have a great Linux OS and, at least by now, even the N900 is receiving some love from CSSU.
Back is a strong word and it's more likely I won't be spending terribly much time on N8x0, N900 or even N950/N9 except for my own personal interests except for anything that comes indirectly through Mer and Nemo efforts on top of hardware adaptations for these devices.
The problem I see in the horizon is that there are no more commercial alternatives that are real Linux and there is where I see Nemo, as an alternative that I would like to see running on Android devices from other vendors (and not only Nokia devices).
|
|
2012-01-31
, 09:47
|
|
Posts: 915 |
Thanked: 3,209 times |
Joined on Jan 2011
@ Germany
|
#19
|
The problem with all these "great Linux OS"es on mobile devices is that they are still not Free Software because they all depend on proprietary blobs.
We see the problems that brings every day when talking about kernel-power, CSSU and many more projects which have to waste a lot of work for not breaking outdated interfaces while trying to implement new features.
I realize that mobile devices have totally different requirements due to their focus on power consumption and system load, but that doesn't mean that the software running on them has to be proprietary.
Just imagine Maemo including all its drivers were Free! That would mean that kernel modules could be recompiled for newer kernel versions and I'm pretty sure we'd have a kernel 3.x by now. Instead newer features have to be backported, i.e. work that has already been done by others has to be done again.
Lots of mobile devices share similar hardware, so Free drivers could be exchanged. Instead every project has to do the same work again and again while trying not to break the interfaces of their special proprietary base system.
I also understand (even though I don't like it) that companies like Nokia have an interest not to release the source code (or only under NDAs) while a platform is new. But I see no reason not to release the sources under Free licenses after a certain time.
Of course that only works if you have that in mind when designing a product and make sure to use only components of which you can release the sources. But if you really mean it what should stop you from doing that?
I consider these proprietary blobs a mayor problem and quite frankly it annoys me that I have no choice to buy a qwerty smartphone that is absolutely Free. Due to these problems I don't consider Maemo to be a "great Linux OS", it just happens to be the least restricting one which is why I use it. But that doesn't mean it's good.
We see the problems that brings every day when talking about kernel-power, CSSU and many more projects which have to waste a lot of work for not breaking outdated interfaces while trying to implement new features.
I realize that mobile devices have totally different requirements due to their focus on power consumption and system load, but that doesn't mean that the software running on them has to be proprietary.
Just imagine Maemo including all its drivers were Free! That would mean that kernel modules could be recompiled for newer kernel versions and I'm pretty sure we'd have a kernel 3.x by now. Instead newer features have to be backported, i.e. work that has already been done by others has to be done again.
Lots of mobile devices share similar hardware, so Free drivers could be exchanged. Instead every project has to do the same work again and again while trying not to break the interfaces of their special proprietary base system.
I also understand (even though I don't like it) that companies like Nokia have an interest not to release the source code (or only under NDAs) while a platform is new. But I see no reason not to release the sources under Free licenses after a certain time.
Of course that only works if you have that in mind when designing a product and make sure to use only components of which you can release the sources. But if you really mean it what should stop you from doing that?
I consider these proprietary blobs a mayor problem and quite frankly it annoys me that I have no choice to buy a qwerty smartphone that is absolutely Free. Due to these problems I don't consider Maemo to be a "great Linux OS", it just happens to be the least restricting one which is why I use it. But that doesn't mean it's good.
| The Following 2 Users Say Thank You to sulu For This Useful Post: | ||
|
|
2012-01-31
, 12:10
|
|
Posts: 1,513 |
Thanked: 2,248 times |
Joined on Mar 2006
@ US
|
#20
|
Originally Posted by sulu
There's rumors of some reverse-engineered, open drivers for Mali being released at FOSDEM and open drivers for the new Intel SoC. Let's see if either of those pan out at all.
Just imagine Maemo including all its drivers were Free! That would mean that kernel modules could be recompiled for newer kernel versions and I'm pretty sure we'd have a kernel 3.x by now. Instead newer features have to be backported, i.e. work that has already been done by others has to be done again.
Lots of mobile devices share similar hardware, so Free drivers could be exchanged. Instead every project has to do the same work again and again while trying not to break the interfaces of their special proprietary base system.