Closing of last tab

Leo Savernik l.savernik at aon.at
Sat Sep 20 03:03:57 BST 2003


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am Freitag, 19. September 2003 19:20 schrieb Thomas Zander:
[...]
>
> I see you failed to read the document and the mailing list I pointed you
> to. These issues were discussed and answers were given.
> In short; they are called KDE guidelines instead of editor guidelines for a
> reason.

I did look at the lists, but none of the postings deal with Konqueror issues 
particularly (If there actually are such postings, I'd be grateful if you 
point me to them).

Whatever, the user interface guidelines itself state that "[t]he guidelines in 
this style guide are only suggestions. [...] Following rules will not do 
that, but common sense will". This is exactly what konqueror does in many 
places.

Ctrl+N: Unconditionally opens a new "application" (using the guide's 
terminology). Concerning the guide, it should not do that if it is in its 
initial state. But for konqueror, this makes more sense.

Ctrl+O: Unconditionally opens the given document into the same window, 
regardless of it being out of initial state. Again, deviating makes more 
sense.

Ctrl+S: Konqueror provides Save As like specified by the guidelines and sticks 
to it to the point. This behaviour is perfectly sensible here.

Ctrl+P: Print, same as Ctrl+S

Ctrl+Q: Unconditionally closes the whole window, whichever state it is in, and 
only the window, not the process -> perfect compliance.

Ctrl+W: Closes a tab. However, it's not working up to the guide. As long as 
more than one tab is open, it will close documents. But on the last tab, this 
stops working.

According to the guide, one of the following actions should take place:
(1) Leave the application with an empty canvas.
(2) Open an empty document in the main-window.
(3) Open a dialog which allows the user to select a new or empty document in 
the main-window.

(3) can be dismissed immediately. (1) doesn't prove useful because konq is no 
MDI application (though with the advent of tabs, it's no longer a pristine 
SDI either), yet it'd be possible, as KDE 2.0's konqueror started with an 
empty canvas. Leaving (2) which essentially means returning to its initial 
state, be it the homepage, the introductory page about:konqueror, or some 
initial directory.

But konqueror does nothing of these three actions, leading to the perception 
that (a) the pre-tab konqueror behaviour was retained without considering the 
guide, or (b) the guide was considered, but deemed not suitable, and for lack 
of something better, the pre-tab behaviour was retained.

Concludingly, konqueror does not follow the guidelines here, but the current 
solution is not sensible enough as to be content with it. Otherwise no 
attempt would have been made to improve it.

When the guide ends, the application is on its own to determine consistency 
and usability. Being able to close the last tab the same way as all other tab 
is undoubtedly more consistent than the current behaviour. That closing the 
last tab will also close the whole window is no worse than the current 
behaviour.

Any user making use of tabs, and getting accustomed to the close shortcut 
Ctrl+W, will expect the document to be closed. It's more disrupting that the 
last tab cannot be closed the usual way, as if the whole window were closed. 
Or put it another way: The user simply expects the document to be closed on 
Ctrl+W, regardless of the count of tabs. The current behaviour is 
non-orthogonal.

So the only task left is to decide in which way the document is closed -- 
whether konqueror follows the style guide, and provides one of (1), (2), (3), 
or if konqueror deviates, and performs a window close. The current solution 
is the least consistent one.

>
[...]
> Global version that are hidden for the user and for new developers are also
> a nightmare to maintain; and thats ok for some reason?

You are right, you have made up my mind. It shouldn't be configurable, it 
should be the default.
>
[...]
> Your feeling of 'inproved usability' is setteled on just feelings and not
> on any rationale that holds for anyone but a very select set of people.

My propositions are no more or less objective than any propositions that have 
been made by other people throughout the mailing lists, and out of which the 
kde ui guidelines have finally been compiled.

Btw, if you have access to any scientifically conducted tab-related usability 
study, please share them. I'm really interested in its findings.

mfg
	Leo
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE/a7WOj5jssenUYTsRAl3tAJ0ajJbSskNrPsAgmYsM1j9Dm0ascQCgnGcT
3wpMIHt7yXgo1j9OI8uccbk=
=CnxO
-----END PGP SIGNATURE-----





More information about the kfm-devel mailing list