Re: gnome-mag and gnopernicus magnifier services TODOs

[Carlos Eduardo Rodrigues Diogenes <cerdiogenes yahoo com br>]

Hi Bill:

Bill Haneman wrote:

> Hi Carlos:
>
> Carlos Eduardo Rodrigues Diogenes wrote:
>
> > What are the plans in gnome-mag, mainly for the full-screen. There is 
> > any roadmap of what can be used to acquire a good full-screen service.
>
>
> Fullscreen magnification now works if you use a recent Xserver with 
> the DAMAGE and XFIXES extension, and with the 'dummy driver' virtual 
> frame buffer configured.  Please see a recent version of the Gnome 
> Accessibility Guide appendices for technical information on 
> configuring this.

I really like the results using this way of fullscreen magnification. 
This works much faster then the 'diopter' program developed by Keith 
Packard. Otherwise, I don't know if this is the best way to follow, 
because I don't know much about  the underlay involved, someone can 
clarify this?

> > I saw this video http://vizzzion.org/stuff/xgl_wanking.avi and is 
> > showed some operations in zoom in and out with very amazing results 
> > (anyone can identify what was used to acquire that?). The Composite 
> > Extension is a good solution?
>
>
> That demo uses GL to do lots of its rendering.  There is talk of 
> moving much of the X server to using GL for its rendering, so we'd get 
> a lot of this performance "for free" then.

If I'm uderstading how gnome-mag works, this is because gnome-mag uses 
Gtk to capture and to expose the images on screen and because Gtk always 
try to use the resources from Xserver to improve it's perform? I saw 
that the Gtk 2.8 supports Cairo, what permits anti-aliased render and 
other good thinks with an X server using GL in it's underlaying.

Anyone have ideia how the magnification is acquired in the video cited 
above? There is anything in the OpenGL API that permits scale 
transformations with that quality?

> Composite will help too, but gnome-mag doesn't include any code that 
> uses COMPOSITE yet; doing so will require a significant amount of new 
> gnome-mag development.

Do you think that this a way that we can follow? Or you think that we 
must clarify other things to take a decision?

If you think that this is a way that we can follow I'm want to help. 
Otherwise, what are the points that must be clarified?

> Either Olaf or Gunnar Schmidt (sorry, I can't remember which of the 
> two is working on this at the moment) is currently working on a 
> COMPOSITE-based magnifier for the free desktop.  There are a few 
> technical issues which need to be solved in the extensions/xserver 
> itself before such a composite-based magnifier can become a practical 
> reality.  The primary issues have to do with conversions between the 
> two resulting mouse coordinate systems, and ensuring that the mouse 
> pointer appears at the appropriate position on screen (and the no 
> "extra" mouse cursor appears).

How I yet said, I saw the 'diopter' application developed by Keith 
Packard that do a COMPOSITE-based magnifier in the xapps directory of 
the free desktop cvs tree. Is this software that you are saying about? 
Do you have any reference where to find informations about this and 
contact the developers?

> > Someone here saw Lunar or ZoomText magnifier for Windows? They have 
> > some resources very interesting in half magnifier screen mode. How 
> > the same effect can be acquired? I think that using the composite 
> > extension can be a way. What do you think?
> >
> > The documentation of gnopernicus (the magnifier part) and gnome-mag 
> > are very poor. What is the better thing to do, an API reference with 
> > doxygen or class diagrams? An API reference is simple to generate 
> > with doxygen, why this was not done yet? There is no need for it yet?
>
>
> We haven't introduced a dependency on doxygen yet, but I agree that it 
> would be a good idea.  If someone wants to produce a patch to run 
> doxygen if it is available, then I will update the inline source 
> documents so that the result is more clear.

produce a patch is modify the configure.in file to the generated 
configure script verify if the doxygen is avaliable and use it to 
generate the documentation? I think that it is what must be done, but I 
never play with the autotools, so I get in doubt. If is this I can do it...


Carlos.