customized CPE - Android & Openmoko

[Wolfgang Spraul]

Dear khang,

> Is  OpenMoko CPE   a  good  choice  for  niche  or  special    
> applications  or  not ???
> Any  ideas ?
> If   not , many   of  the  developers  will   go  for  Android .   
> That  is  not  good .

Let me use this opportunity to talk a bit about Openmoko and Android.
First of all we really like Android! We don't see Android as  
competition, it is complementary to what we are doing and may help us  
in many ways.

If you look at the Android software stack, you will notice that they  
basically only use the Linux kernel and a few traditional 'helper'  
libraries, written in C (jpg, png, etc).
But the bulk of the system is written from scratch. They even have  
their own libc! Their own Java virtual machine, their own graphics  
system, etc. etc.
I do believe all of this is very healthy. Fresh blood. Take the Dalvik  
virtual machine for example. Basically they kick Sun somewhere, but  
that may turn out to be a nice wake-up call for something like HotSpot  
and similar established Java projects. IcedTea, GNU Classpath, etc.
At the same time they ignore pretty much everything the FOSS community  
built over the last 10-20 years.
No X, no d-bus, no standard packaging system, ...
It's a fundamentally different approach from Openmoko.

 From outside you first have to decide: Do you go with Android and  
step outside of 90% of what the FOSS community has built? Or do you go  
with Openmoko, which is a lot closer to something like Ubuntu?

There are many things to learn from Android. I like their 'intent'  
system for example. Lots of great software will become available as  
Open Source, and will find its way into many places, one of them being  
Openmoko.
Getting the complete Android to run side-by-side with GTK or  
Enlightenment will be a lot harder than it is with Qtopia for example.  
That's because even though Qtopia can be seen as one large monolithic  
code base, it is still relatively well connected to many other FOSS  
standards. Android is not connected at all, just sitting on top of the  
Linux kernel, and several times larger than Qtopia.

Bottom line: We hope the Android sources will be releaesed soon. We  
think Android is a great piece of software, will become really  
successful, maybe even become the long-awaited 'Linux desktop' one  
day? The future looks good for Android.
Openmoko will benefit in many ways, from Google pushing chip vendors  
to become more open to lots of high-quality Apache-licensed source  
codes being set free, allowing us to cherry-pick from Android into  
Openmoko.
Additionally, you can either play with Android itself, or run Openmoko  
software on hardware that was opened up thanks to Android.

Hope this helps, everything is moving so this is just a snapshot in  
time.
Feedback very welcome, a lot of this is driven by what the community  
wants to do. Often people show us how things should be done, not the  
other way round ;-)
Wolfgang

On Apr 4, 2008, at 11:32 AM, khang wrote:

> Taiwan  CPE  developers of  WiMax (etc..) are  looking  for  a new   
> OS/plateform ,
> so  their  applications  won't   be  limited .
>   We  are  planning
> 1.City  surveillance (live  project) ---  installing  digital   
> camera  wirelessly  sending  " clear "  image  back  to  control   
> center.
> 2.Remote  medical  care --- sending  medical  data / warning   
> signals   wirelessly  to  Hospital / doctors
> 3.Home  security   ---  remote  monitoring  home  status  including   
> old  people  and  child  safty
>   using   WiMax  network ( mobile  Internet ) .
> Is  OpenMoko CPE   a  good  choice  for  niche  or  special    
> applications  or  not ???
> Any  ideas ?
> If   not , many   of  the  developers  will   go  for  Android .   
> That  is  not  good .
>
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community