On Mon, 2006-07-24 at 16:38, Chris Jones wrote: > Bill Haneman said: > >I don't recall ever "refusing" to fix any corepointer > >issues, for instance - at the time you asked, there were multiple > >research efforts under way to try and address them. > > Maybe I put that too strongly. You stated that the fix was to have a > second input device and were unwilling to consider, or at least gave > me the impression that there was not anything that could be done about > it. This may have been a flaw in my own interpretation.
My point was the using the corepointer, 'as is', was not acceptable from an accessibility point of view because of lockout scenarios and serious conflict with toolkit pointer behavior. I stand by that - the only reason I suggest it can be done with Xevie is because Xevie gives us a means of intercepting the pointer events before the toolkit or applications see them. > I also count the dependence on sticky keys a configuration issue. StickyKeys is available on all recent x servers. So in that respect, it's not a configuration issue, since the OSK can programmatically set this (and should, for various reasons). No modification of the default Xserver configuration is necessary. > As a general rule no application should be too dependent on the > configuration of the underlying operating system be it Xorg > configurations or a11y configuration. Completely aside from the > inconvenience to the user, conflicts between applications with > different configuration requirements could become quite likely. Of course. Simiilarly, applications must also not implement features that are too similar to existing platform features, because that can also result in conflicts from the end-user's point of view. This is especially important where accessibility is concerned, and in fact the Section 508 accessibility requirements explicitly require conforming applications not to conflict with the operation of existing platform accessibility features. This is one reason why applications that wish to use keyboard latching or locking must be very careful about how they implement it, and preferably they should use the existing StickyKeys services built in to the X server. Bill > It's very easy for an application developer to think that his or her > application is the most important thing running on a users machine, > though it is not a fair assumption to make. > > -- > Chris Jones _______________________________________________ gnome-accessibility-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnome-accessibility-list
