Hi Ben,

> Your dichotomy is exactly right.  Chandler makes most sense either as a
> component of the system, somehow integrated with the Journal, or as an
> Activity (basically just an application like any other).  Ideally, I think
> I would prefer the former.  In Sugar, Activities run in a sort of isolated
> virtual machine that prevents them from communicating with other
> Activities or accessing user data unless explicitly authorized by the
> user.  This seems a bit unnatural for Chandler, which is designed to
> manage the rest of the system.  However, it's much easier to write an
> Activity, which can be installed by the user and is little more than an
> arbitrary Linux program, than to alter the Sugar system itself, which
> requires modifications to the Sugar code.
> 
> I found Davor's suggestion to implement a Chandler backend that presents
> an interface over D-Bus especially interesting.  Sugar's architecture is
> very much designed as a large number of independent programs that
> communicate over D-Bus.  Every Activity has a D-Bus service.  Even the
> collaboration system uses D-Bus, tunneled over the network by Telepathy.
> If the Chandler backend could run as a D-Bus service, I suspect we could
> make use of it most easily.
> 
> When would be a good time for the Chandler developers to meet?

I forgot to answer this question when you asked it; right now Grant and
I are focused on thinking about architecture at a different level than
this, but I think in a week or two it would be great to have an IRC
meeting.  Beyond that, we're pretty flexible about timing.  What
timezone are you in?

Sincerely,
Jeffrey
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev

Reply via email to