Re: GEP 10 [was Re: GNOME Development Series Snapshot 2.3.0: "Mighty Atom"]

[Andrew Sobala <aes gnome org>]

On Tue, 2003-05-13 at 17:58, Luis Villa wrote:
> On Fri, 2003-04-11 at 15:42, Rodrigo Moya wrote:
> > > Actually Abiword and mergeant are distributed in Fifth Toe, gnumeric and
> > > sodipodi have not been because at the time the GTK2/GNOME2 ports were
> > > not stable.
> > > 
> > right, sorry Will, I forgot about 5th toe :-) But, then, what sense does
> > it have to have specialized apps in the desktop release (such as
> > gnomemeeting, for instance) and apps that everybody seems to need
> > (abiword, gnumeric) in a separate release?
> 
> It is my sense that gnomemeeting (via VOIP) and $IM_CLIENT could become
> powerful building blocks for other things. If gnumeric  or abiword sees
> phone://617-868-9192, it could recognize that as a phone number, and use
> gnomemeeting to call me. Ditto IM. And there are tons of other
> applications that I can't even begin to think of- if IM and VOIP are
> part of the desktop, Evo can use it, $GNOME_SCREENSAVER can check if
> your friends are online, etc., etc. Gnumeric, Abi, etc., are things you
> build on top of building blocks- IM and VOIP can be building blocks that
> those apps build on top of, if the groundwork is done. We already think
> of TCP/IP and http this way- we need to expand our thinking and toolkits
> in those directions.

The "Desktop Integration" thread indirectly raises the problems that
GNOME has here. Instead of adopting a single client for these protocols
that are called with command line options, we should have well defined
interfaces for getting a [GNOME/FreeDesktop]-compatible-[VOIP/IM]-client
and well defined interfaces for communicating with it. And that
infrastructure - which I don't know enough about to even draft an
implementation, but wasn't bonobo designed for this? - is more important
than adopting the clients into the GNOME core desktop.

This helps us to support the user's favourite client, and it lessens the
importance to projects of being part of the GNOME core. We want a
mechanism in which we can support a single "GNOME-recommended" client,
over QA, I18N, and A11Y support, but *without killing all other
competition in that niche.* Although it's advantageous to a single
project to be the only one supported by GNOME, it's a problem for GNOME
as a whole.

-- 
Andrew Sobala <aes gnome org>

"A freudian slip is when you say one thing but you mean your mother." -- unknown