Answering why gnome-shell might not work for me (was Re: What does gnome-shell give us?)

[Sandy Armstrong <sanfordarmstrong gmail com>]

I really don't want to have a thread that devolves into gnome-shell
bashing, but then again this is not an inappropriate venue to answer
your question.  I've changed the subject because I think it's a
different discussion from the rest of the thread.

On Fri, Oct 2, 2009 at 11:37 AM, Colin Walters <walters verbum org> wrote:
> On Fri, Oct 2, 2009 at 3:04 PM, Sandy Armstrong <sanfordarmstrong gmail com>
> wrote:
>>
>> Hey folks,
>>
>> Taking a step back from the excitement over shiny new things, I was
>> just wondering if somebody could give a few bullet points that explain
>> what gnome-shell gives us over our current desktop experience.  I feel
>> like there's a lot of stuff that's different or missing but I'm not
>> sure I'm understanding where it's "better".
>
> Also, I would subdivide "better" into "better right now",  "better once bug
> 12345, bug 12346... are fixed", and "i don't see how this could ever work
> for me".

Sure, bugs/regressions are to be expected, so what I was asking about
(and what you answered) was more along the lines of "how will
gnome-shell 2.30 be better", or maybe "why should convince the release
team to include gnome-shell in GNOME".

> We certainly have some serious regressions right now, the application browse
> in particular is really unloved, but it's also among the things we'll be
> working on quite soon.  I could absolutely understand if you were saying
> "not better right now" (for me personally it's in the middle state, a few
> bugs are fairly bad).  But if you're saying something closer to the latter
> that's more of a concern.

Well, since you're asking, I am certainly concerned that gnome-shell
might fall in the latter category for me.  Again, I haven't used it
since before GUADEC due to dependency issues, but from what I can tell
it is still the same basic workflow, so I don't feel completely
unjustified in my concern.

I agree with the idea of trying to achieve an activity- or
application-centric UI.  I also agree with the stated goals of
providing a visual refresh, and exposing recent documents in a useful
way.

The real killer for me is that I get the feeling that gnome-shell is
not activity-centric, but shell-centric.  Shell is set set up like a
dashboard, which is not an unpopular type of UI, but it feels
unnatural to me in a desktop situation.  I have "getting stuff done"
mode, where I'm doing the actual work I have a computer for in the
first place, and then I have "shell mode", where I manage
activities/applications/workspaces/documents.  These two modes are
completely separate from one another, have very different behavior
from one another, and "shell mode" is completely different from every
other popular desktop UI.

It's kind of hard for me to explain why this separation is so jarring
for me.  It could be the foreign nature of it, as I am a little slow
to accept change.  But the best way I can think to describe it is that
it removes the magic from the desktop experience.  Arthur C. Clarke
said that any technology sufficiently advanced is indistinguishable
from magic.  I think it makes sense to strive for UIs that seem
magical to users.  When I use gnome-shell, I feel like it's a very
hands-on approach to activity management.  But I don't want to learn a
new application that requires my full attention to manage activities.

With current default GNOME panel, I can switch workspaces and
applications through a minimalistic UI without changing modes.
Finding and launching applications is tedious, but it too doesn't
require a mode switch.  These basic desktop features blend into the
overall GNOME UI and fit familiar (if imperfect) interaction models.

For myself, I've solved the problems with the current task switcher
and application launcher by using gnome-do's Docky UI.  I get a fairly
application-centric approach to task switching (which works across
virtual workspaces), and a quick method of application launching that
does not require me to open a full screen application like
gnome-shell.

I'm not saying that GNOME should adopt Docky instead of gnome-shell.
Instead, I'm trying to demonstrate that there seem to be ways to
achieve the stated goals of gnome-shell without introducing a new
full-screen mode that, for me, interrupts my work.

> Also a few other goals to the list:
> * Actual design for how workspaces behave (We didn't really have this in any
> consistent way, but now is an opportunity to fix it right, e.g. make moving
> an application to a workspace a persistent operation, etc.  Actually a lot
> of cool things are enabled when we have an application based system)

I'm very interested to see this progress.  I think this type of
improvement to workspaces could be very beneficial, and with zeitgeist
we may even be able to automatically populate workspaces and bring
some of the magic back to the desktop.

> * Help push forward the technology stack by being a consumer of Clutter,
> introspection etc., including hopefully providing a motivation for graphics
> drivers

This is a good long-term goal, too.

Sandy