Re: [g-a-devel] proposing ATK/AT-SPI API additions

[Mike Gorse <mgorse novell com>]

Hi Peter,

On Tue, 18 Jan 2011, Peter Korn wrote:

> From my own experience I haven't see any AT actually need these setter APIs. 
> Not all apps want to let users move windows around (e.g. MessageBox style 
> windows/dialogs).

If an app doesn't want to let a user move a window, then this can be 
handled by the ATK implementor.  GAIL tries to handle this by only 
moving/resizing a widget if gtk_widget_is_toplevel returns TRUE for the 
widget corresponding to the ATK object.  If GAIL allows a widget to be 
resized through ATK where it cannot be resized through normal user 
interaction, then this could be considered a bug in GAIL that should be 
fixed there, so I don't see this as a reason not to expose these ATK 
functions over AT-SPI.

> Exposing API for doing things programmatically that 
> otherwise cannot be done through the UI is treacherous ground - something we 
> need a very compelling accessibility reason to do.
>
> Supporting a macro package doesn't pass that bar for me.

I am going to withdraw that part of the proposal at least for now, since 
several people have expressed concerns about it and so it seems 
questionable/unclear that it belongs in the APIs.

Thanks,
-Mike