[maemo-developers] Next steps in the future of our beloved Nokia 770

[From: Quim Gil]

Hi,

On Mon, 2007-04-30 at 21:55 +0200, ext Frantisek Dufka wrote:
> So what is the future, do you want just fix bugs or can this be used as 
> a test bed of future device vs OS independence?
>
> As there are probably more devices in the pipeline and some will 
> definitely run same major version of OS but have different hardware this 
> problem must be solved in future anyway.

Yes, in fact we are working already on device vs OS independence (i.e.
integrating HAL or Hildon scalability), being IT OS 2007 and the N800
our targets. There are other technical processes going on around the
improvement of the IT OS & maemo architecture. This is one of the
reasons why the story of the N800 will be different than 770's. 

It is really out of my scope and technical knowledge how much of this 
work done for the IT OS 2007 can be made available in the Hacker
edition. Carlos (and probably also prominent maemo hackers like
yourself) will know much better. My blind guess is that the community
effort will be important to make all this happen in the 770 & hacker
edition. 

>  How we can reflect this with 
> maemo SDK and repository structure? Looks like there are such things as 
> subarchitecture and flavors in Debian.
> http://www.debian.org/releases/stable/powerpc/ch02s01.html.en

Looks like we have some Debian developers in the house and we are
familiar with this paradigm. :) It is not trivial to incorporate this
(or a similar repository-based structure) to our current setup, though.
It has remarkable technical challenges per se, and it has also
additional challenges for our current development process and business
context. 

The question is how far we can and want to go (and how). This is an
uneasy discussion we are having now. 

But well, it is not a coincidence that we are using Debian as upstream
project, packaging system, APT, etc. The public roadmap already lists
features like system meta package and automatic codec installer that go
in this direction of flexibility in system components compared to the
release of monolithic FIASCO images. So we are progressing.

> I have reported the top of the iceberg as a bug here
> https://garage.maemo.org/tracker/index.php?func=detail&aid=804&group_id=164&atid=681
> But it needs some thinking. How can developers target specific devices 
> with same OS but different hardware? Some developers may already take 
> eqaution Bora = N800 as granted which means some packages will not run 
> on n770 due to compiler flags and other hw specific assumptions.

I guess this shows that the software architecture needs some fixes and
improvement, as well as some of our processes. Again, it is out of my
scope and current knowledge how much of this have we already solved and
are aware of. I tend to be optimistic but filing bugs is always a good
method to make sure your points are taken.

Again, we want to provide one day this software vs hardware detachment.
This means your points are fully on topic. It is probably too early to
know how much of this will end up benefiting the 770, though. I hope we
are able to build a story of collaboration together. 

-- 
Quim Gil - http://maemo.org