Re: Question for Bastian Nocera

[Ruben Vermeersch <ruben savanne be>]

On vr, 2010-06-04 at 15:19 +0100, Bastien Nocera wrote:
> On Wed, 2010-05-26 at 20:36 +0200, Javier Jardón wrote:
> > Hello,
> > 
> > in your application you say:
> > 
> > > - (Re-)defining GNOME:
> > > The Foundation charter defines GNOME as a loose collection of
> > > independent project, though we need to stop considering it as such if
> > > GNOME is to take an important role in the future of computing, be it on
> > > the desktop, or in devices, where it would provide the infrastructure.
> > 
> > Could you elaborate a bit more about this?
> 
> Look at the upper and lower bounds on this diagram for GNOME Mobile:
> http://www.gnome.org/mobile/gmae-arch-diag.png
> 
> Where does GNOME start and stop?
> 
> Do we go from the kernel up? From the user-space bits up? Is something
> still GNOME when it doesn't use GTK+? When it doesn't use Matchbox (as
> per the diagram), or metacity/mutter?
> 
> I would think it being fine to say, GNOME is:
> - Linux kernel
> - D-Bus
> - NetworkManager/BlueZ/PolicyKit/udisks/upower
> - X11
> all the way to GTK+/Clutter combination and apps
> 
> And this is what we need to focus on. There's a lot of swamp-draining to
> be done in the lower levels, and working on GNOME means working on one
> of those things in the stack.
> 
> In the same way, I think it doesn't shut out other OSes, be they other
> free Unices, or even Mac OS X and Windows, where the stack is just
> shifted (pretty much everything underneath what we currently consider
> the GNOME stack).
> 
> Defining the GNOME OS is required if we want to avoid getting cornered
> working on the bits at the top of the stack, and working around
> problems, rather than solving the solutions "The Right Way" all the way
> down our stack.
> 
> Obviously, this would require discussions...


Is this related to (I think it was Mark Shuttleworths) idea of defining
a reference platform with synchronized versions of the core components
(which we could then target)? I can imagine this to be massively useful
for ISVs.


-- 
Ruben Vermeersch (rubenv)
http://www.savanne.be/