Crowdfunding an Ubuntu smartphone

Dr. H. Nikolaus Schaller hns at
Sat Aug 24 15:31:19 CEST 2013

Hi Radek,

Am 24.08.2013 um 14:22 schrieb Radek Polak:

> On Friday, August 23, 2013 10:21:33 PM arne anka wrote:
>> the missing money is just the indicator for this project's failure to
>> create sufficient public interest or even awareness.
>> while the GTA01/2 was a nice idea, it was already slightly outdated when
>> it appeared -- and since then nothing has changed, the gap between what's
>> considered standard and what the GTAxx is prepared to deliver rather has
>> widened.
>> to get even close to standard (and thus being a realistic alternative for
>> smartphone users), the project would need backing of a far more potent
>> entity than this tiny community is -- both money- and publicitywise.
>> since the GTA02 i spend about 2000€ on this project, maybe more (well, i
>> can afford it and it was worth supporting the idea behind it), and in my
>> very personal and subjective opinion, the GTA04 has been a huge
>> disappointment. i wouldn't spend the equivalent of a highend, state-of-art
>> smartphone or even tablet just to buy yet another even more outdated
>> device, free or not. compared to the shortcomings of the GTA04, even
>> complete freedom is not sufficient to justify that amount of money.
>> whenever i told somebody who had heard of OpenMoko that there's a
>> successor, they were surprised -- and when i told them the features and
>> the price, i got an incredulous grin and the question if someone really
>> believed that people would spend that amount for such a device.
>> i am still undecided if i should admire or pity the thread starter, if he
>> honestly believed that this community would be able to succeed where
>> ubuntu failed -- and on top of that to jump from todays GTA04 to the
>> device as imagined by ubuntu ...
>> all things considered, the realistic path is imo to cater to a tiny niche
>> of institutional customers -- like jörg(?) proposed.
>> while i am personally rather fond of the original GTA01 case (and think
>> that's almost the only tangible unique feature), i, too, would prefer it
>> to pick up where the n900 left. maybe then it could even tap into the pool
>> of still active n900 fans ...
> Very well written. I can add a few more points.
> For me GTA04 is not usable as daily phone. It's now collecting dust in shelf. 
> Why?
> 1/ poor power management
> 2/ bugged reenumerating modem
> GTA04 is good enough only in some situations. E.g. ok if i carry it to work 
> where i can anytime charge it. When i am on bike 150km from home i must have 
> reliable phone in case that the bike breaks up or in case that i need map.
> I took GTA04 with 2 batteries and N900 on my bike trip this summer. I used 
> N900 as a phone with SIM card. GTA04 was switched off - i just used it for 
> GPS. But after 4 hours one battery was empty and second battery was 50% empty 
> during one night in suspend. On the other hand N900 was working whole 4 days. 
> I used it for calls, for wifi in camps and in the end even for GPS.

That is something I still don't understand. From a power budget the modem is
specified to have 3-10 mA in suspend while registered to a base station.

I have even verified this with an ampere-meter.

And there is only one potential deviation - if there is some attenuation (e.g. wall)
there may be a situation where the modem receives the base station well and
tries to answer/register. But since the base station does not receive it at all, it
tries with high power. In this case the modem current increases to 50-100 mA.

BTW: this is the same with almost all mobile devices.

So it must be something in the OMAP system and that is IMHO the same as the
N900 has... So pure kernel code.

BTW: this would not change by redesigning the GTA04 into a spare N900 case...

> I though that self-made open source phone will have good power management. I 
> though it will be easy or at least possible to identify where the power goes. 
> I am even more dissapointed that noone except Neil Brown tried to improve the 
> situation.

Yes, that made me wonder as well. It appears that it needs some very special skills
we don't have in our community. Or we have but those members have no time.

> Event GTA02 is soooo much better in this area. For me this is so 
> important that GTA02 is now much more usable then GTA04. It's like having car 
> that can go just 30km.
> And the reenumerating modem - i really dont understand why we havent tried 
> newer firmware. For me missed calls is quite serious problem.

Because the upgrade firmware did not run on the kernels for unknown reasons
(failed with segfaults). And because we must make RMS believe that there is
no way to upgrade the modem firmware from user space.

> I dont understand why produce more phones until these problems are fixed.

I don't know how to fix these problems without producing more phones. Especially
there is one issue in the IrDA/RS232 driver which prevents the kernel to power down
both and the ITG3200 may also draw some mA. This all would be fixed in the GTA04A5
board design.

And there would be a different modem firmware if we buy new modules. And then
we don't even need to upgrade but can just compare two device releases.

BTW: buying new modules would be a good reason to get a permission to really
distribute the firmware upgrade package. But as long as we don't produce more
boards, we simply have no argument to knock at their doors.

> I even dont understand the efforts to make PVR, camera, radio and other 
> components working, unless we have working phone and not mobile heat 
> generator.
> Well you can answer why i am just writing mails instead of doing 
> something. But i always worked in userspace. I barely understand kernel and i 
> have no EE skills and equipment to contribute.

This may be the same answer for almost everybody else here. We simply don't
have enough kernel hackers to improve power management. And even my
personal skills in this area are very limited so that I wasn't able to contribute
anything. Adding PVR, Camera, Radio etc. is much simpler (just copy some
existing code into the kernel tree and fix some bugs),so I did focus on areas
where I did see a chance to succeed.

> I can contribute only as a 
> tester. I thought that i will deliver working userspace and IMO QtMoko is very 
> good at it. But without working kernel and HW there is not much point to 
> improve it.

Hm. That sounds quite different from the situation about 1 year ago when
you did the first releases of QtMoko and I always thought that the
3.7 kernel is working well enough, so that I started to add new features.

Has it become worse since then?


More information about the community mailing list