Re: Planning for GNOME 3.0

[Sandy Armstrong <sanfordarmstrong gmail com>]

On Sun, Apr 19, 2009 at 6:31 AM, Emmanuele Bassi <ebassi gmail com> wrote:
> On Sun, 2009-04-19 at 14:34 +0200, Sebastian Pölsterl wrote:
>
>> I think it would be a big mistake to omit applets in the new gnome desktop
>> evolution.
>
> why?

Because users want some functionality to be conveniently available
from the time they log in to the time they log out.

> we've been changing the platform gradually over the years, mostly by
> deprecating stuff and including new functionality. nevertheless, I
> haven't heard a single justification for the continued existence of
> "applets".

See above.

> what do applets provide, nowadays, and are they even remotely useful?
> what can deskbar-applet provide that cannot be implemented with
> something that does not sit inside a 24x24 icon on the most valued piece
> of screen real estate? isn't a gnome-do approach equivalent to the
> deskbar-applet?

I personally use gnome-do to access Tomboy more than I use our
applet/tray icon.  Do you think you and I are typical?  Would be nice
if somebody did some work to research this instead of just saying "I
feel this way, so why don't you?".

> why tomboy-applet is so special? it's basically a
> launcher with a custom context menu. also, starting up deskbar-applet
> *and* tomboy as applets on my panel causes my desktop more to start up
> on login; not great turn ons, especially when there are developers out
> there trying to get the boot-to-UI process down in the seconds range.

Users put all sorts of crap in startup scripts, session management,
panel, etc, because they want it available at login.  This is not
going to stop if we get rid of applets.  The correct solution is to
fix the performance problems.  Of course, I 100% agree that any
applet's functionality should be available *without* it being a
required startup item.

But going back to your previous statement, doesn't the use of
something like gnome-do also increase my login time?  But don't a
great many users take that hit willingly?

> clock; and the volume is now becoming a notification area icon since
> basically everyone has media keys on their keyboard and don't need an
> on-screen slider anymore.

Wow, that statement strikes me as very out-of-touch with reality.

> yes, it was all good with GNOME 1.x, but even for 2.x the amount of
> applets has been steadily decreasing - also because writing an applet is
> not trivial (as it involves dealing with some of the most obscure and
> less documented parts of our development platform). people have been
> abusing the system notification area with all sorts of crap (beagle,
> tomboy, etc.) because writing an applet is *boring* (server files
> anyone?) and *hard* (weird build changes, hard to debug uses, completely
> different APIs for handling the menus), and it really doesn't provide
> you with much functionality (wow, an icon and a context menu!).

Yes, we have all abused the hell out of the notification area, despite
HIG to the contrary.  Tomboy supports it because users demand
applet-like functionality even without a GNOME panel.  One of the most
common ways Tomboy is used is to be added to GNOME session startup and
run in the notification area.  Do you think this abuse of the
notification area to emulate applet behavior is the right thing to
promote in the gnome-shell era?

> so, please: saying "it would be a mistake" without providing reasons why
> it would be good to have applets support in the first place it's not
> going to convince me that we should keep them.

I also agree with Sebastian's response, btw.  Applets (whatever they
are) should be *easy* to create.  The idea of ever-present desktop
objects is not going to go away, and users are not going to stop
wanting it.

I'm thinking that the best thing we can do is to run some (long)
usability studies on this, or find similar research that can help us
make an informed decision.  We are developers, and often have very
different ideas about how to use a computer, compared to (for example)
a typical office worker.  And yet, a significant amount of GNOME users
*are* developers and otherwise-advanced computer users.  Perhaps we
should attempt to identify typical user types, and evaluate what it
would take to make GNOME 3.0 work for them.

But we don't have the money or time for this, so maybe there is
existing data that can help us out a bit.  Any thoughts?

Sandy