[Tfug] Speaking of desktops (little 'd')...

[Bexley Hall]

--- On Tue, 11/11/08, John Gruenenfelder <johng at as.arizona.edu> wrote:

> >> I always turn on focus follows cursor *and* auto-raise
> >> (500ms delay).  As for
> >
> ><frown>  I dislike delays in the user interface.  After about
> >300ms, people tend to get fidgetty -- "why isn't the thing
> >responding to me?".  Deliberately adding a delay in order to
> >implement a "feature" seems "forced" to me  :-/
> 
> Well, you could certainly alter the delay.  500ms is just
> my preference.
> Gnome gives you a slider to set it.  And you always have
> the option of just clicking and skipping the delay altogether.

Understood.  I just object to using "time" (delay) as a "control"
in a user interface.

> >Do you like/prefer having incoming mail "announced"?
> >Should "out of paper" be an audio annuciator (instead
> >of a distraction on the screen)?
> >
> >What if you had a stock ticker with "speech" output... should
> >it be babbling incessantly in the background?  Or, perhaps,
> >configured to just announce changes in particular stocks of
> >interest to you?  (I'm just making up imaginary apps that I
> >can imagine "wanting to 'talk' to you" even though you are
> >not actively "focused" on them)
> 
> This is one reason I'm generall a big fan of tray icons
> in whatever form the WM/desktop chooses to present them.  
> They allow new information and alerts to
> be displayed without forcing me to acknowledge them ASAP. 

Exactly.  However, I'm not sure there is a parallel
"mechanism" for the audio channel.  :<  In the desktop
environment, it (currently) isn't a real problem because
few things emit more than a beep, etc.

> Tray icons with bubble text are a grey area... To me that says "more
> critical - pay attention!" but is often used for not so important
> items.

Yes.  I don't see how the API can enforce rules on the developers.
People seem to rationalize doing whatever they want when it
comes to the interface -- ignoring the fact that there are
N other "things" that could be equally cavalier about stepping
on the interface...

[grabbing the server]

> >*Should* this be available to an application writer?  I.e.,
> >this comes back to the question I posed in previous message:
> >is there *ever* a time when the system should grab the server?
> >And, now I ask, should an *application* be able to do this
> >(knowing that developers are free to do whatever they want to
> >do with the API you make available to them -- regardless of
> >how stupid it may seem)?
> 
> I'd error on the side of yes.  How about the example of
> a UPS monitoring app.
> Something critical occurs and it wants to you know right
> away.  That's not so bad.  But, you're right that it can, 
> and often is, easily abused.

Should, instead, the application have to notify the "system"
(WM, etc.) of its intent and let the "system" decide how this
is conveyed?  I.e., this "feature" lets an application act like
root and stomp on your display at will...