CLosing apps

[Matthew Woehlke]

Maciej Pilichowski wrote:
> On Monday 16 March 2009 21:12:54 Matthew Woehlke wrote:
>>> * close direct parent container of this window (except desktop)
>> Is an empty container allowed? If not, your "close all siblings"
>> overlaps. My preference would be for just this one.
> 
> Good question. Initially I was against it, but more and more I think 
> about it and I opt for -- yes.
> 
> Pro:
> * empty desktop exists

Special case, but...

> * more natural feel for TAI -- after closing the last tab, you don't 
> have to create container from beginning

...this one feels more important. Especially then you can make 
container, put windows in it.

Ok, so add wm-close-parent-children? (Name isn't really important, 
anyway, a "close all siblings and me" action. Leave it unbound by default?)

And, as it relates to the other discussions, I would still say a 
container may not have input focus *unless* it has no children.

> Cons:
> * may looks odd (containers are frames really)

Well, yeah, I don't think it would look /that/ odd though :-).

> So I opt for:
> * TAI can be empty
> * GAI can be empty
> * single app. (SAI) -- cannot be empty (by definition)

"SAI" = window not in a container ;-).

>> Maciej Pilichowski wrote:
>>> On Monday 16 March 2009 19:59:15 Matthew Woehlke wrote:
>>>> I'll try the "poetry" again.
>>>>
>>>> If I have:
>>>>   container
>>>>     konq
>>>>     konq
>>>>     okular
>>>>     konq
>>>>
>>>> ...I'm not convinced we need a shortcut to close just the konq's
>>>> in the above. (Unless it is "close all siblings" :-), which is
>>>> probably a good action to have, though I'd leave it unbound by
>>>> default.)
>>> I also don't think we need shortcut for close _konquerors_, but I
>>> am 100% sure we need (I need) shortcut for close _this
>>> konqueror_.
>> What is "close this konqueror" in the above?
> 
> Let's say the container is TAI --> closing one tab.

== wm-close-window, no? I have increasing feeling we are agreeing but 
confusing each other by inconsistent terminology :-).

>>> If yes -- I also suggest "close all siblings except this one".
>> ...and /this/ is wm-close-siblings :-).
> 
> So implicitly you are saying there are no empty containers. Hmmm...

Beside "see above", part of what I am saying is I am not my own sibling 
;-), i.e. "sibling" implicitly excludes "self". That's how I would use 
it in anything user-visible, anyway.

> So we are both on wm-close-siblings-similar? Ok.

If we agree there is no reason to close only those windows in an app 
that are the same app, then yes.

>>> Could you please you accept / don't accept those.
>> Except for still not sure what you want from "close doc" vs "close
>> window", +1.
> 
> Let's say I start kpdf+doc. Closing doc, would be closing the opened 
> file. Closing window would be closing kpdf as application.

Didn't we have this discussion a while back? :-) Hmm... okay, so TWI 
brings some useful sanity to the concept of an "empty tab/window/etc". 
Maybe wm-close-window shouldn't be ctrl-w, and app-close-document should 
never close a window and restore window to "empty/nothing loaded" state? 
I think I would be okay with that.

>> Again I prefer that apps should strive to eliminate 
>> MDI entirely and leave it to the WM, in which case "close doc" and
>> "close window" are the same thing.
> 
> I am against this. IMHO we should have two-level structure for 
> application:
> * application (menu, toolbar, etc) 
> * document (the content)

If you mean it should be possible to get empty window, okay. If you mean 
we shouldn't discourage apps from having MDI stacked on top of WM 
features...

-- 
Matthew
Please do not quote my e-mail address unobfuscated in message bodies.
-- 
"I don't question your existence -- God" (seen on a church billboard)