Re: [orca-list] Orca and D-Bus

[Piñeiro <apinheiro igalia com>]

From: Jason White <jason jasonjgw net>

> Willie Walker <William Walker Sun COM> wrote:
> > Jason pretty much has it right. :-)
>
> That's encouraging.
> > To expand upon this, however, I'll often refer to this as the
> > "Bonobo Deprecation" task for GNOME. As outlined in the following
> > WIKI page, this task covers AT-SPI, Magnification, and Speech:
> >
> > http://live.gnome.org/Accessibility/BonoboDeprecation
>
> One issue that I couldn't find discussed on that page concerns the
> performance of a DBus-based AT-SPI. As I remember, there were, initially,
> performance concerns, which I hope have been addressed. This should also help
> to prepare this infrastructure for the move to mobile devices, where resources
> can be highly constrained - but even for netbook-type devices, these issues
> can be expected to arise.

Although I'm not a expert in the internals of both, one of the reasons
in favor of the move (not major, but one) was also the constrains on
mobile devices, although in this case about memory.

As you pointed, on devices like mobile devices and netbook types, the
resources are highly constrained, but not only about performance it
also about memory.

DBUS will be used extensively, and bonobo, orbit and so on only for
a11y. Some people point that will be a waste of memory have bonobo,
orbit, etc on the device just for that, when DBUS could make the same
work.

About the performance, a study was made several years ago, when this
work started [1], but some years have passed so the numbers now would
be really different.

Just my two cents.

[1] http://live.gnome.org/GAP/AtSpiDbusInvestigation/

===
API (apinheiro igalia com)