Re: [Usability]Keeping the Quit menu item

[Calum Benson <calum benson sun com>]

On Thu, 2003-03-20 at 17:28, Calum Benson wrote:

> FWIW, some anecdotal evidence from one of the other usability guys
here:

And some more:

> To counter Dave (Curbow)'s comment...   Dave is correct that there are certain 
> user expectations today, but those expectations are not universal. IE on 
> Windows does not have Quit/Exit. Neither do File Manager windows. 
> Historically, the Apple Lisa had no Quit anywhere and that didn't cause 
> significant problems, but then GUIs were very new and people didn't know 
> quite what to expect. Smalltalk has no Quit command.
> 
> That said, removing Quit reduces the user's sense of control. How is a 
> user confident that the application is off. But do you think your TV is 
> really 'Off'? Does it need to be?
> 
> If an application is (thought to be) misbehaving, what is a user to do? 
> But then the Quit command may not be accessible either, so this in 
> itself isn't a good reason to retain Quit.
> 
> Jeffrey Baker makes three points:
> 
> 1. If all of an application's windows are closed there is no way to 
> interact with the application. This is false. Just do the same thing 
> that is normally done to start the app or open a related desktop icon. 
> This requires an OS enhancement.
> 
> 2. Resource utilization. Partly true, but this should have minimal 
> impact. The OS should be free to kill the process if needed to reclaim 
> limited resources.
> 
> 3. How do we deal with 'application' state as opposed to window-related 
> state, e.g.. cookies and connections? This is a little tougher since 
> there are so many different situations. Quit deals with all of these as 
> a side effect, e.g. it breaks existing logins. In general, closing the 
> last window could do this as a side effect, though that might not be the 
> correct thing for some apps. The latter would need to have an explicit 
> command.
> 
> More to the point. The bottom line to all of this is that to remove the 
> Quit command, there must be a very clear and well supported 
> non-application user model, e.g. the document models of Star and Lisa. 
> (This goes to Dave's comment about user expectations too). If people are 
> aware that they are 'starting' processes, if that is part of the user 
> model, then it is likely that there must also be a way of 'stopping' them. 

Cheeri,
Calum.


-- 
CALUM BENSON, Usability Engineer       Sun Microsystems Ireland
mailto:calum benson sun com            GNOME Desktop Group
http://ie.sun.com                      +353 1 819 9771

Any opinions are personal and not necessarily those of Sun Microsystems