Re: [Usability] Close buttons on instant-apply {dialogs|windows}

[Reinout van Schouwen <reinout cs vu nl>]

Hello Daniel,

On 3 Jan 2002, Daniel F Moisset wrote:

> 1) The point for including it is to ensure there is an (usable) way to
> close the window even when the WM doesn't provides one, increasing
> accesability providing a larger target to hit, and having a Tab/spacebar
> navigable way to close these windows.
> 2) The point for not including it is to have a single way to close this
> kind of windows using a mouse click (there are another alternative ways
> like WM menus, taskbars, shortcuts, etc.). This is good for consistency
> and to avoiding ambiguity. This would (theoretically) increase usability
> by not forcing the user to make a choice where there's nothing to
> choose.
>
> [If I missed something in any side, please, add it].

I think to (2) could be added that the rationale against a Close button
includes that there shouldn't be one _because_ a instant apply window
isn't a dialog. There is no clear left-right top-down order one has to
work through the window contents. Imagine an instant apply Preferences
window with multiple tabs. The user could make some changes on the lower
part of tab A, move to tab C and enter something in a textfield on the top
of that page, then dismiss the window with whatever way the wm provides. A
'Close' button on the dialog introduces uncertainty whether the changes
will be permanent or not. (Agreed, 'Done' or 'Dismiss' is better for
that).

As for the Help button: how about doing providing advice with tooltips?

> "capabilities" of other desktop components (in this case, the WM). In
> this case, there should be some kind of utility-window widget that has a
> close button or not according to what the WM/theme answers when queried
> about "Do you have a close button in the frame?". If the WM/theme
> doesn't support the protocol, or there is no WM, a compromise choice
> should be taken (I would say 1, which is less bold but less risky).

A very nice idea, I think. I kind of doubt it will work but if you can do
it, it would be a way to satisfy both sides... (although then you could
argue that a user could be confronted with two different behaviours of the
same app on different machines...)

thanks for your input, anyway!

regards,

-- 

Reinout van Schouwen
e-mail    : reinout cs vu nl
mob/voice : +31-6-44360778 / 084-8750706

VU/W&I studiereis naar Ierland! - maart 2002 - www.cs.vu.nl/storm/ierland