Re: [Ayatana] Global menu in Oneiric Ocelot (11.10)

[GonzO <gonzo@xxxxxxxxxxx>]

On Wed, May 25, 2011 at 8:49 PM, Ed Lin <edlin280@xxxxxxxxx> wrote:

>> With all due respect: I do not think your cases are all that common.
>
> I never claimed the contrary. BTW I don't use Unity and probably never
> ever will (I'll spare you the reasons, it wouldn't be pretty).

I do not know whether or not anyone Official(TM) reads this list.
However, if the reasons for never ever wanting to use Unity aren't
aired, it'll never get better.

Better out than in, I say, but that is something best left to your
discretion.  :-)


> you have mine: I want a great FOSS OS for everyone) and I gave plenty
> of reasons why for those the global menu is worse than what they had
> before.

I'm not sure anyone disagrees.  Well, wait, I think one person argued
for them in the past week, but for my part, I think its one of Apple's
two big UI mistakes (apple-q to actually _quit_ an application is the
other).


> The only counterargument I heard is Fitts's Law

Which, frankly, people bring up far too often and understand far too
little about.

It's a good goal to aim for, but there's a difference between "in
theory" and "in practice".  FL is good in theory, and helpful in most
cases, but not all.  And when we run up against a situation where its
better to ignore it, we should do that.


> I know stuffing panel indicators into the launcher is an afterthought
> and therefore very, very difficult to solve from a design and
> usability perspective. But just because it's difficult doesn't mean we
> couldn't try.

The problem I have with this is not its implementation, but the idea
itself.  I don't think putting indicators into the launcher is a good
idea, what-used-to-be-a-panel notwithstanding.  My reasons for this:

#1 - The launcher is vertical.  This, in general, is a good idea, but
in order for it to _remain_ a good idea, some things need to be
considered.  Like the fact that, on MOST monitors (the ones not
rotated 90 degrees), there are fewer pixels in the vertical direction.

Which means, in simplest terms, that the launcher needs to be really,
really choosy about what is and isn't allowed to use its valuable tile
spaces.  In fact, it should only be limited to things that are
launchers.

And this is something Unity is _already_ somewhat bad at.  Trash isn't
a launcher - or, it launches Nautilus, which already has a launcher.
Virtual Desktops, not a launcher.  Two different lenses taking up two
different tiles, when the functionality of each is merely a spin-off
of the Dash.  They _are_ launchers, but that's two more launchers than
is really necessary or desirable.  And none of these are removable
except the lenses, where you actually have to uninstall Unity
functionality to remove them.

With 1050 pixels to deal with, four extra squares isn't all that bad.
But then again, that's just the starting line.  Actual USE fills the
thing up, even on a big screen, pretty darned fast.  Every app that
isn't already a launcher?  New square.  Every device you hotplug?
Yeah, new square.  In about twenty minutes of actually _doing stuff_
on my computer, I already have a few collapsed slides, even on one of
my old 1600x1200 displays.  And the overflow behavior, as its
currently implemented, makes this more aggravating.

Because of this, I am forced to be really choosy about which actual
launchers I put on the bar, even though putting program launchers
there is supposed to be the whole point.  I don't think it can afford
to lose any more space to any more non-launcher things; in fact, it
needs to go on a diet already.

#2: Indicators that contain text, like the date/time app, are almost
impossible to display sanely in a launcher square.  "September" all by
itself won't fit unless the text is so small its squint-worthy.  The
username indicator would be a problem in my office, where usernames
are typically full first+last name.

#3: Top right (and horizontal) is, IMO, a perfect place for them.  I
don't think they'd work as a notification and interaction device if
they were located anywhere else on the screen.  But that would appeal
to me, as an English reader (right -> left, top -> bottom).  For
example, I doubt I'd even see, much less pay attention to, any
indicators that are near or on the bottom of the screen.  I always had
this issue in Windows (until I moved the taskbar to the top, like I
usually do).  And Ubuntu has been making this area the place for these
things for a while; Apple already does; and Windows can be made to by
moving the taskbar (they're already on the right if your bar is
horizontal).

#4: Interaction.  I frequently adjust the volume by placing the cursor
over the speaker and mouse-wheeling.  I change my status with that
user indicator; I change networks with the network indicator.  Most of
this wouldn't work if several indicators were grouped into a single
tile (the only sane way this can be done, IMO), and in that
implementation, I'd have to click twice to get anywhere with any
indicator at all.  I'm not even sure the volume trick would even work,
even if it had its own tile to itself (which, again, probably just
wouldn't work).

#5: Indication.  I'm not sure we even _can_ display a speaker icon
with varying numbers of "waves" to indicate volume, or a "link
down"/"link up" pair even if each indicator would be its own launcher
tile, which wouldn't work.

I _could_ see a single, expandable tile functioning as a "systray"
area on the launcher; Cairo-Dock has this already, and this is how
Windows behaves if you move the taskbar to either side of the screen.
This would solve lots of those problems, but still be on the dock, and
still incredibly hard on text-based indicators (someone with a long
username, date/time, etc.).  Not to mention, if you're an indicator
freak, that tray area could grow to two or three squares, and how do
you collapse it?  Do you allow it to be collapsed?


> 3 solutions: keep the full panel, keep a partial panel, completely
> remove the panel.
>
> Would you not agree the cleanest solution from a high level is
> completely removing it?

Too clean, IMO.  I think this solution is the same kind of high-level
perspective that prompts so many to yell out "Fitt's Law!" at every
available opportunity.  It is so clean that it makes indicators unable
to be effective or interactive indicators, and shoehorns them into a
an already overburdened device.

Its not that I have any love for the Panel-which-is-no-longer-a-panel.
 I agree that if global menus are removed, a good deal of its
real-estate would become mostly worthless.  But if removing it means
moving indicators from the top-right and confining them to 48px
squares, then I think that keeping the panel or giving it the
WingPanel treatment are both vastly better solutions.


> Any ideas on the organisational side of things here? Maybe launchpad
> is suited better because it forces us to split up the issues into
> little manageable chunks but then the "grand vision" is lost.

None whatsoever.  I was pointed here from the forums, by someone who
seemed to be in the Ubuntu organization, apparently because no one
Official(TM) reads the forums.  o.O  Yeah.  I don't get it either.
Maybe it was a forum mod, and not Official(TM) Ubuntu brass.

In closing, I apologize if the last few messages come off as a "Hey,
everyone, let's get Ed Lin!" pile-on.  By and large, I agree with a
lot of the ideas you've had; I just think the indicator-in-launcher
thing is a non-starter.

--G