unity-design team mailing list archive
-
unity-design team
-
Mailing list archive
-
Message #03168
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
2010/6/29 Matthew Paul Thomas <mpt@xxxxxxxxxxxxx>
> > If I have chosen to actively use another application/window, the new
> > application should never take the focus even if it takes it 24 hours
> > to start.
> >...
>
> Of course, but that's *still* assuming the question. How do you define
> "chosen to actively use"?
>
> Define it by clicks, key presses, and milliseconds. Any other definition
> is useless to a window manager.
>
>
Click on window, type in window, switch to window (by whatever means,
ALT-TAB or any switcher). Actually, for cases where it's possible to clearly
identify this, even intent to switch should be enough. The case I think of
here is ALT-TAB, if I've initiated ALT-TAB but not chosen a window yet, it
still should count as if I had and start any other app unfocused.
Any other thing that anyone can come up with that unambiguously means that
I've actively given this window my attention. "Unambiguously" and "actively"
are keywords, and I can well imagine that they are hard for many cases, but
click, type, switch to - are those unclear? If so, can they be made clear?
I don't see time entering into it at all. I have interacted with a window,
or I have not. The interaction happens instantly, on mouse down, key down
(or mouse over if using sloppy focus I guess), no milliseconds or anything.
Just as the interaction never times out, if it happened after a launch, it
stays forever for that launch, so that application can never come to front.
(Well, actually, for a complication, I guess it would if all interacted-with
windows has gone before it launches, as in: "start OOo, type in gedit, save,
exit, gedit is gone, OOo pops up in foreground").
Actually, it's probably trying to be too smart and guessing by timeouts and
other things that make the current window managers not work, to the point
that most of us think that there is no algorithm at all? I still hardly
can believe that it does, because there are no signs whatsoever that the
desktop respects my focus choices. :)
I'm not sure though, how the window manager knows when an application is
started? Is it that focus is not on any window, since a launcher, menu or
similar has been used, so that it can notice interactions with windows after
that point in time - or does it actively need to know that something has
been started? If I assume the former, the basic suggestion becomes something
like:
* WM keeps a list of windows that has been interacted with, defined as:
- clicked on (mouse down)
- switched to (by any means given the focus)
- typed in (if that is possible without any of the others first?)
* Addition to list is instant, according to the above rules no waiting, no
timeouts
* WM empties this list when no window has focus, as would be the case when:
- launcher, panel menu, desktop icon is used
- or, if that's a bad way to go about it, list empties whenever a new app
is launched
- or, however this point in time is detected :)
* Window is removed from list when it is closed
* If list is empty upon window creation, it gets to open on top, with focus
* If list is non-empty upon window creation, it opens unfocused, behind
- It could be discussed if it should open behind all others, or directly
behind the last focused window
* Window list is NOT per-user, but per-WM (if that's the problem with the
gksu windows popping up all over, otherwise maybe superfluous requirement)
* Special cases may include counting initiated ALT-TAB and similar "intents
to switch" as having a non-empty list, though I'm not sure it's necessary or
even a good idea - testing perhaps?
Probably not perfect in any way, but are there cases that are unclear in
that? Are there cases that are unwanted in that? I've probably overlooked a
lot of obvious cases, or don't have an understanding of what is possible,
but I do think something like that is at least the base of how I want my
system to respect my choice of focus.
/ K
References
-
Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Sense Hofstede, 2010-05-24
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Conscious User, 2010-05-24
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Mark Shuttleworth, 2010-05-24
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Alex Schoof, 2010-05-24
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Tyler Brainerd, 2010-05-24
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Mark Shuttleworth, 2010-05-25
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Greg K Nicholson, 2010-05-26
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Frederik Nnaji, 2010-05-27
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Conscious User, 2010-05-27
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Kristoffer Lundén, 2010-05-27
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Matthew Paul Thomas, 2010-05-28
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Greg K Nicholson, 2010-05-29
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Matthew Paul Thomas, 2010-06-21
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Frederik Nnaji, 2010-06-21
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Matthew Paul Thomas, 2010-06-21
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Kristoffer Lundén, 2010-06-21
-
Re: Papercut or not? Bug #495403 in One Hundred Paper Cuts: “Do not raise windows or dialogs without user input”
From: Matthew Paul Thomas, 2010-06-29