Re: A few comments regarding Gnome-shell

[Robert Park <rbpark exolucere ca>]

On Mon, Mar 7, 2011 at 5:29 AM, Juergen Mangler
<juergen mangler univie ac at> wrote:
> On 03/07/2011 11:43 AM, David Prieto wrote:
>> Hi Jürgen,
>>
>>    I do not comprehend the advantages of the menu on top approach
>>    (apple approach). With high resolution, big screens and several (not
>>    maximised) windows, the distance i have to move the mouse grows much
>>    higher.
>>
>> My experience with top-panel menus has been that yes, they're farther
>> away but no, they're not more difficult to reach because I could "throw"
>> the cursor to the top of the screen and be sure it would land on the menu.
>
> Okay, i buy this one.

David is describing Fitt's Law:

http://en.wikipedia.org/wiki/Fitts%27s_law#Success_and_implications

The disadvantage of the MacOS approach here is that it renders
focus-follows-mouse worse than useless (as it becomes impossible to
access the application menu if there are any other windows in between
that window and the top of the screen, which would be common for any
non-maximized window).

On a side note, I was recently annoyed when I tried out v0.0.6 of the
GNOME ISO (from gnome3.org) and I was not able to activate the clock
menu by clicking on the topmost row of pixels above the clock applet,
I had to move the mouse down in order to click. That's a failure to
implement Fitts Law.

>> As for the advantages, I think it would go well with Gnome3's efforts to
>> remove clutter and distractions, by hiding every menubar except for the
>> one from the active window. Seriously, I can't recall ever having wanted
>> to open the menubar of an inactive window, partly because in most cases
>> they're half-hidden anyway and you need to raise the window first, so
>> why not go all the way and hide them entirely?
>
> Gnome-shell has some tiling features built in *, and when the windows are
> side by side its two clicks vs. one. Of course it can be weighted against
> additional clutter on the screen. I like having everything in one place (the
> window). But maybe thats old-fashioned.

I'm a HUGE fan of how GNOME3 has implemented tiling. It's so smooth,
and the UI for it is so deliciously consistent with how maximization
now works. I've been pondering the implications of widescreen monitors
for a while now, and it frequently occurs to me that vertical space is
incredibly precious, while horizontal space is so ample that I can't
even think of ways to use it all up. It's easier to read text when
it's in narrow columns, so it makes sense for UIs to allow windows to
be arranged into columns (and indeed, I already have configured my
mail client to have columns for the inbox and message view, rather
than the more traditional inbox above message view below), and similar
for my RSS client, etc.

> Regarding unity: i confess i like that in unity the titlebar is integrated
> with top panel when windows are maximized. So I think removing the close
> button from windows altogether (as others have done), and maybe moving it to
> the top panel (not only context menu of activity, but explicit) is at least
> worth a try.

Yeah. I'm absolutely in favor of anything that can reduce the usage of
vertical space. On my netbook right now (screen size 1024x600), I
actually have a very wide gnome panel along the left edge with various
launchers, and then the panel across the top (with a clock and
notification area) is actually set to autohide because I just can't
spare a single pixel of vertical space.

What if titlebars occupied the left edge of the window instead of the
top edge? Rotate the text 90 degrees. It's only slightly less
readable, but it would save a LOT of space for people with wide
screens.

-- 
http://exolucere.ca