Do we need DBusMenu in the Shell?

[Giovanni Campagna <scampa giovanni gmail com>]

Open Planet GNOME. Almost half of the latest posts relates on GNOME, KDE
and Canonical collaboration story, which is all around libappindicator
and StatusNotifier / dbusmenu.
One year ago, libappindicator was rejected because it needed integration
in GNOME Shell and Gtk+. Three days, Unity was unveiled, and everything
went well with considering Ubuntu community the "bad guys" that forked
GNOME, starting with Ayatana. I of course know this is not true, but
this is the impression I got from the outside (I only started
contributing to GNOME some months later).
Today, on the other hand, things reversed. Public opinion is now that
GNOME rejected, and still is opposing, Canonical per se. And of course,
all the flaming around GNOME Census didn't help here. We don't want a
large part of our user base to consider GNOME Shell a Red Hat project
(again, this is what is perceived from the outside).
First, because this is not true: not just Novell, Intel, and all the
various other companies, but also a great of individuals are making up
what GNOME 3 will be, in Fedora, OpenSuse, Debian, Gentoo, Arch and
maybe Ubuntu as well.
Second, because even if all developers were paid by one organization, we
would have failed if we didn't pass the message that GNOME is a body, a
project, and an organization, but most important GNOME is a community of
people, working for the advancement of free software.
We're about to release GNOME 3. We need the best publicity to have this
adopted by the majority of people. After all, our goal is a GNOME
desktop on every system, right? But if people switch to Xfce or KDE or
Unity (or, ugh, Mac OS and Windows) for political issues with the
project, rather than usability, design, technical bugs, then we're
wasting our precious developer time.
So I think we should give Canonical and the general public a big signal
of collaboration. It doesn't matter if this will have no direct effect
in terms of code and upstreaming of patches, we'll have done the right
thing and people will know.
The small step I'm talking about is support of DBusMenu icons in GNOME
Shell Message Tray. Yes, it does not fit the design (which calls for
notifications, not menus) and it is not used by anything inside GNOME
now. But code is there, in a bug where I proposed for the status area;
it would require some changes to adapt to last Gnome Shell version, but
is mostly fine, and has no external dependencies (implements the DBus
protocol directly).
Anyway, the reason I'm proposing this is not technical (libnotify with
persistence is by far better than StatusNotifier, I think we all agree
here), it is political.
We need to show we're open to technologies developed elsewhere, no
matter how dirty they are. Look at browsers and HTML5, which is the most
horrible application platform ever invented by man: they're all
competing on who is be the most compatible with the others and with the
spec.
I hope this will start a positive discussion, at least to make sure
we're not ignoring this, which is a serious issue from a PR standpoint.

Giovanni Campagna