nanogui: Thread: use windows id's for buttons?

Dear <your name here>!

We are developing a slide show application for a STB under nano-X. The slide
show needs some buttons (it is a rather advanced slide show). When a button
is active (mouse over) it changes appearance as well as putting descriptive
text in a rectangle on the slide (plus more).

The question is whether we are best off treating the areas on the screen (an
activated button might change the appearance of several rectangular ares on
the screen) as windows with window id's or whether we should just let the
button handle the areas in some other way?

We believe that there is a point in making the areas as windows, thus be
able to receive events "directly", but we have had many discussions about it
and we wanted to get a the gut feel from more experienced people.

K. Nikolaj Berntsen  Technical Consultant
--------------------------------------------------------------
AGENCY.COM Interactive TV

Islands Brygge 57, DK-2300 Copenhagen S
T: +45 3693 6000  F: +45 3693 6001  
Direct: +45 3693 6120  
E: ####@####.####  www.agency.com 

On Thu, 28 Jun 2001, Nikolaj Berntsen wrote:
> The question is whether we are best off treating the areas on the screen (an
> activated button might change the appearance of several rectangular ares on
> the screen) as windows with window id's or whether we should just let the
> button handle the areas in some other way?

It would probably be easiest to make each area be a seperate window. Don't
worry about creating too many windows; they have a pretty low overhead.
When each logical item (button, message window, tooltip, etc.) in the
display is a window, moving it or unmapping it will automatically move or
unmap any child windows. Also, if you move a window to a different
location on the screen you don't need to worry about what the new absolute
location of the items displayed within the area, are as they are still in
the same position relative to their parent window. Just my humble opinion.

-- 
------- Alex Holden -------
http://www.linuxhacker.org/
 http://www.robogeeks.org/

I agree 100% - In fact, there are many distinct problems with *NOT* using 
seprate windows.  FLTK/FLNX is one example of a toolkit that draws everything 
on 1 window, and everyone who has ever worked with the FLTK/FLNX internals 
will tell you that it is a most confusing mess.

Jordan

On Thursday 28 June 2001 01:40, Alex Holden mentioned:
> On Thu, 28 Jun 2001, Nikolaj Berntsen wrote:
> > The question is whether we are best off treating the areas on the screen
> > (an activated button might change the appearance of several rectangular
> > ares on the screen) as windows with window id's or whether we should just
> > let the button handle the areas in some other way?
>
> It would probably be easiest to make each area be a seperate window. Don't
> worry about creating too many windows; they have a pretty low overhead.
> When each logical item (button, message window, tooltip, etc.) in the
> display is a window, moving it or unmapping it will automatically move or
> unmap any child windows. Also, if you move a window to a different
> location on the screen you don't need to worry about what the new absolute
> location of the items displayed within the area, are as they are still in
> the same position relative to their parent window. Just my humble opinion.