Panel Accessibility (long, but please read) [was Re: Panel Status - GNOME 2.0]
- From: Bill Haneman <bill haneman sun com>
- To: Sander Vesik <Sander Vesik sun com>
- Cc: George <jirka 5z com>, Glynn Foster <glynn foster sun com>, gnome-2-0-list gnome org, gnome-hackers gnome org
- Subject: Panel Accessibility (long, but please read) [was Re: Panel Status - GNOME 2.0]
- Date: Sat, 17 Nov 2001 18:39:01 +0000
Sander Vesik wrote:
>
> On Sat, 17 Nov 2001, Bill Haneman wrote:
>
> > Sander Vesik wrote:
> >
> >
> > > The prority of having a p[anel is far more higher than having an
> > > accessible panel, so all accessability work on panel should be postponed
> > > unril after such time as we have a working panel and people can work on
> > > aplets (panel is blocking work on them).
> >
> > I disagree with Sander here, at least the "far more" part. If we don't
> > address accessibility in the panel *now*, it won't get addressed, and
> > without an accessible panel then Gnome-2 accessibility is basically a
> > lie from the end-user perspective. Though we can add accessibility to
> > most apps and non-core toolkits later, we need for that accessibility
> > which will be coming on line in the Gnome2 timeframe to work *with the
> > gnome2 platform*.
> >
>
> I'm not sure I follow - how is not having a real panel in gnome2 (and
> using a pretty much ad-hoc launcher) better than having a panel that
> while maybe not completely accessible can be made to be so?
You have already raised the spectre of a Gnome 2 without a panel (or one
that "doesn't work") - or was that just a rhetorical device? What ever
happenned to the concept of "show-stoppers" ? If we are considering a
library-only Gnome 2 release (as opposed to "platform only", which I
think we've all conceded now), then that message needs to be transmitted
now and alarms raised. Personally I think that a libraries-only Gnome 2
should not be called Gnome-2.0, it will just make us an easy target for
the slashdotters, etc. Call it Gnome-2 Developer Edition or Developer
Early Access, and stick to the existing schedule.
I am not sure we need to be that draconian but I think that there is
some subset of the 'desktop' that needs to be working in the Gnome 2
release, and at least some of that needs to be using Gnome 2 libraries.
I believe that the original intent was that *all* of the Gnome 2
"platform" should use the new libraries but there may be issues with
meeting schedule for some things, like librepgtk/sawfish, so I assume
exceptions could be made with appropriate review.
Since accessibility has been touted as a key part of Gnome 2 we are
under pressure to show "something". In fact we've come a long way, we
have patches outstanding to provide full keyboard navigation for GTK+2,
and to enable more accessible themes. With some edited stock icons we
will have the accessible themes themselves. But if GTK+2 is keyboard
accessible and themeable but the Panel is not, I think that any claim to
accessibility support (even in a limited way, like keyboard navigation
and themes) will ring hollow.
We can retrofit enhanced versions of libgail, at-spi, etc. for users who
need assistive technologies (and those technologies will only be coming
on line for Gnome 2 in the immediate post-G2 timeframe anyway). But if
the panel does not implement accessibility or keyboard navigation or
respect themes, we have a bigger problem since a user installing the
base Gnome-2 package from, say, a CD will expect it to work properly
with the latest accessibility libraries.
> > We don't want to require, say, a "panel patch" in order to use assistive
> > technologies with Gnome 2, and we can't afford to wait for 2.2 in order
> > to have at least the panel and control center accessible. As Telsa
> > pointed out, expectations for accessibility in Gnome 2 are very high.
> >
> > I think that it is already a problem to ship any inaccessible bits with
> > the so-called "core" gnome2, though we have conceded that such will be
> > the case. To ship an inaccessible panel seems to me totally at odds
> > with the Board's assertion that accessibility is a key part of Gnome 2.
> >
>
> From this it sounds like you have a plan - that you haven't shared with us
> - that details how the gnome core will be made accessible. Becuase
> otherwise this does not make any sense, I'm afraid.
Sorry Sander, there is no conspiracy here.
And I don't think there is any secret about how gnome-core is being made
accessible.
The gnome core is being made accessible via accessibility enhancements
to GTK+ and ATK, and enhancements currently being made/recently made to
gnomeui, gnome-canvas, libzvt, and bonoboui. Although not bug-free,
these changes are virtually completely API frozen now, implementation is
already substantially functional, and the key mechanisms for providing
support (libgail, atk) have been carefully designed to be extensible and
to cause minimum impact on existing libraries. We are almost there, but
if the first pieces of Gnome UI that a user encounters (Panel and
terminal) are inaccessible then the users will never get to appreciate
the "theoretical" accessibility of the platform, and even if apps are
later made 100% accessibility-friendly, an inaccessible panel will mean
that from the user perspective Gnome 2 is as inaccessible as Gnome 1.4.
Within the trimmed-down constraints we have targeted (complete support
for GTK+ widgets, basic support for gnomeui and zvt widgets, basic HTML
viewing [another email], and connection via bonobo-control for
bonobo-embedded GTK+ widgets; accessibility for the panel,
gnome-terminal, and control center), we are not in such bad shape. The
GTK+ work is API-complete and mostly in bugfix mode, the gnomeui and zvt
work is substantially complete, the bonoboui work has been sketched out
and researched. We are committed (Gnome Accessibility Team) to making
the gnome-terminal accessible via the zvt work, and think that the
additional work for Control Center should be tractable. The panel is a
worry, but it's the only area in this scaled-down "minimum desktop"
where we think we can't retrofit support if it's not included now,
during the port.
-Bill
p.s. - I have bcc'd marc.mulcahy, marney.beard and peter.korn at
sun.com, since they are part of Sun's Accessibility Program Office and
Marc has been doing the libzvt and libgnomeui work - to help prevent
their being included unnecessarily on all replies. If you wish to CC
them of course go ahead.
> > regards,
> >
> > Bill
> >
>
> Sander
>
> "I don't think there is intelligent life within our solar system"
> -- Brian Behlendorf
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]