Re: Reasoning behind panel re-implementation

[William Jon McCann <william jon mccann gmail com>]

Hey,

On Sun, Nov 22, 2009 at 7:00 PM, Barry van Oudtshoorn
<bvanoudtshoorn gmail com> wrote:
> Hi,
>
> I've been playing a bit with Gnome Shell, and, largely, I like what I'm
> seeing. I'm just wondering, though, why the panel is being completely
> re-implemented in a non-backwards-compatible way. I understand that the
> panel is meant to be more closely integrated with the window management, and
> I applaud this, but I don't understand why support for the myriad panel
> applets that have been developed using the current system will be dropped.

Glad to hear that you like what you are seeing.  Good news.

Some background here may be helpful.  Before the Shell was even born
there was general agreement among core GNOME developers and designers
that the GNOME 2 panel had outlived its usefulness.  There are a lot
of reasons for this.  I won't go into them here.  Owen, Vincent Untz
(current panel maintainer) had discussions during GUADEC 2008 about
what we should replace the GNOME 2 panel with.  It was decided to
organize a User Experience hackfest to help answer that question.  The
GNOME Shell project was born shortly after that hackfest and used many
of the ideas that were presented or generated during it.  So in a way,
the Shell was designed to replace the GNOME 2 panel in a non-backwards
compatible way.

The compatibility issue is really only a factor for applets.  But
since we wanted to get rid of bonobo (GNOME 2 applets use bonobo)
anyway - a compatibility break is implied by GNOME 3.

So then question is to understand and define the role that applet-like
functionality will play in the GNOME 3 user experience.  That is
ongoing.

> I see in the mailing list that there's discussion about the clock. Wouldn't
> it make more sense to use the existing clock applet, which is rich in
> functionality, rather than reinventing the wheel?

There will be code (or at least concept) reuse where it makes sense.
I can sympathize with the idea of not just throwing away what is good.
 In fact, I worked on the clock applet too. :)  However, it is also
very important to capitalize on the opportunity to fix past mistakes.

> I know that Gnome Shell is powered by JavaScript, and that therefore there
> are some issues with using applets written in C, but I honestly do feel that
> writing some sort of wrapper to enable users to use whatever existing
> applets they want would solve an awful lot of issues. For example, users
> could enable the traditional GNOME menu as a way to 'get used to the
> system'. Less time would be spent redeveloping functionality that already
> exists in GNOME, and more time could be spent delivering a really great UI
> and UX. I also think that take-up (by users and distros alike) would be
> faster, as you wouldn't have people waiting for the favourite applets to be
> 'ported' to the new system.
>
> It's just my two cents, but I'm really interested in hearing the responses
> from the devs (and non-devs, of course. ;))

It is a reasonable request.  We have actually done something similar
already with the user-session menu.  It is largely copied from the GDM
user switch applet and is still mostly written in C.  However, it is
just begging to be rewritten or enhanced to make use of the
capabilities we have in the shell that we didn't have in GNOME 2 (or
at least couldn't expect to have).

Hope that helps.  Thanks for your feedback.

Jon