[maemo-users] Is OS2006 still supported?

From: Igor Stoppa igor.stoppa at nokia.com
Date: Fri Apr 18 14:19:55 EEST 2008
On Fri, 2008-04-18 at 14:07 +0300, Quim Gil wrote:
> Igor Stoppa wrote:
> > On Fri, 2008-04-18 at 13:52 +0300, ext Quim Gil wrote:
> >> ext Luca Olivetti wrote:
> >>> En/na Quim Gil ha escrit:
> >>>
> >>>> - The community is empowered to keep over time hardware and software as
> >>>> fit as they wish and are able to.
> >>> Quite difficult when some critical hardware parts are only usable with 
> >>> binary blobs.
> >> What is really the problem: the binaries or the license to distribute
> >> them? Or both?
> >>
> >> Do these binaries impede the community getting what they want? What do
> >> they want, by the way?
> > 
> > working wireless lan, battery charging, screen dimming and watchdog
> > pinging would be a good start.
> Do you mean that the community can't do a thing unless this
> functionality is covered by open source packages instead of binaries?
> Do you mean that Nokia 770 owners can't work on progress themselves
> until these components are opensourced?
> What would happen if the community could reuse these binaries in images
> build and redistributed by themselves?
> Also, is it feasible to think in open source alternatives developed by
> the community?
> One thing is not to do something and another thing is to impede others
> from doing things. Pointing the real stones in the way would help
> concentrating the attention on what matters.

Afaik the kernel module for wlan is binary only, this prevents new
kernels from being used.

Then binary only key applications introduce dependencies on version of
the libraries they are linked against.

For example if the community would like to get rid of initfs, then the
implementation would not be so strightforward, since initfs binaries are
built against a different libc.

About the community developing its own version, well, afaik in certain
countries it's illegal to reverse engineer sw and anyway we are not
really helping in certain sw areas. 

Look at the kernel code for retu and tahvo: it's quite close to be
obfuscated. And we haven't opened the specs for those asics.

Sure one can rewrite a piece of userspace code with no close HW
interaction, but these functionalities i'm talking about are too close
to the HW to be rewritten without actually having the HW specs.

Also i'm not sure about how open the API used by dsm and bme is,  that i
leave to you to check.

But to properly allow the community to come up with its own versions of
the closed components, we should make both API and related datasheets

Cheers, Igor


Igor Stoppa
Next Generation Software
Nokia Devices R&D - Helsinki

More information about the maemo-users mailing list