Re: Proposed module: conduit

[Dave Neary <bolsh gnome org>]

Hi,

Vincent Untz wrote:
> Short description:
> ==================
> Conduit is a synchronization architecture for the GNOME desktop. It
> provides an intuitive GUI for synchronizing things and a DBus interface
> for external applications to do the same.

Stepping back for a second (obnoxiously creating a whole new
sub-thread), I'd like to say it's a pity that "conduit as architecture"
has become conflated with "conduit as GUI application". It's kind of
like talking about gconfeditor's GUI when considering gconf for
inclusion in the platform.

Conduit's value, to me, is in applications using it as a
transport/heavy-lifting layer to do synchronisation stuff. A back-up
tool with Conduit as the back-end would rock. Having a way to keep
photos with a certain tag in f-spot automagically sync with Picasa or
Flickr would rock - of course, the best place for a GUI for this would
be f-spot itself, and not a Conduit client.

What is the best way for Conduit to get in? If John were to provide a
plug-in for synchronisation of Tomboy notes that would integrate into
the Tomboy UI, and Tomboy developers added Conduit as a hard (or soft)
dependency, would that be the way to go?

I see huge potential for a synchronisation framework like this, it could
hook into applications from Evolution and Thunderbird for PIM stuff,
online desktop-type apps for local/remote data sync, or a back-up tool
for simple user back-ups (I'd love that by the way:
http://blogs.gnome.org/bolsh/2007/07/10/easy-personal-back-ups/).

So if it's not getting in now, can we at least get constructive and plot
a path where the framework gets accepted next time around? And if it
*is* getting in now, can we point some maintainers towards docs that
will allow them to start using it straight away in 2.26?

Cheers,
Dave.

-- 
Dave Neary
GNOME Foundation member
bolsh gnome org