|
Hello,
Well I am sorry perhaps I'll ask too much but please try to at least consider this. We all are kind of jealous that on Windows while using Firefox and NVDA they have verry good accessibility support and switching into / out of focus mode is kind of automatic and it is usually right so why not to try borrowing some ideas from them. I know it may not be that easy but I am thinking that although we originally wanted to have something completelly different on linux avoiding forms mode vs browse mode paradigm now we are about to introduce something verry similar. On windows when using so called focus mode the screen reader does nothing to the content and nothing to the browser caret it just interprets stuff. This is mostly equivalent to what we do have on linux while gecko caret browsing is active. When forms mode is off we are using screen reader specific feature named so called browse mode on Windows where content is already partially preprocessed and indexed for us so we can use quick navigation keys etc. On linux when the orca caret browsing is active the content is not preprocessed nor indexed instead orca controls the browser's caret trying to create similar if not the same experience. So to say it verry simply the difference is that on Windows the content is preprocessed by the screen reader and on Linux it is not however what the end user receives is or should be the same feeling. My suggestion is to combine the two settings we are discussing here into a single feature. If gecko is controlling the caret you can't and usually don't want to use quick navigation commands and most of the related screen reader features. However when orca caret browsing is active this is the case and what we are doing is we are using quick navigation commands and expecting orca to handle arrow key presses. Then let's remove all the logic that may affect performance where orca tries to determine where and when to control the caret and where not. If we'll do this we will get ability to manually toggle whether we want orca to control browser's caret and at the same time whether we want to use structural navigation keys and other screen reader related features which happen to be part of the gecko toolkit support script. This is more or less equal to the paradigm there is on Windows we do have two modes of operation. One where we strictly rely on browser capabilities and the other where screen reader is tryyng to control what it can. Then let's examine what NVDA is doing on Windows in order to see whether it would suit our context and whether we'll like it. NVDA has two settings related to this. In the Browse mode settings window of NVDA preferences menu there are: - Automatic focus mode for focus changes: If this is enabled it examines curent control whenewer system focus changes inspecting the control's role, state and ancestry. If we'll change focus into for example text field, text area, combobox (also equivalent aria widgets) it will toggle the focus mode on and disable browse mode. Translating this to our scenario would be that it should switch into gecko caret browsing. The important part to note here is that no mather what we do in order to change the focus e.g. click a mouse, tab around, or a web app changes the focus the gecko caret mode will kick in. - Automatic focus mode for caret movement: When this is on NVDA also does the same kind of checking when it's browse mode cursor enters such a focusable control. Translating this to our scenario means when orca controls the caret and it moves into such focusable control / aria widget the gecko mode should get enabled for us. I am not sure if this might be possible to distinguish since orca now controls browser's caret directly thus moving the caret would trigger focus change and the first option would apply regardless of the second option. And switching back to browse mode in NVDA happens automatically when new page loads. Other than this switching from focus mode to the browse mode i.e. switching from gecko caret browsing to the orca caret browsing is done after pressing the designated shortcut key. This can also be pressed anytime when orca controls the caret it should change to gecko controlling the caret and vice versa. Not to clash with these names and still being able to distinguish between these two modes on linux desktop with orca and firefox I think we might call them browser control and screen reader control. The names don't sound too verbose and are still providing clear hints on what's going on. On the key bindings we currently have orca+F12 and orca+z occupied by this new feature merging gecko vs orca caret browsing setting as well as structural navigation on / off toggle. If that's free in this context I guess we might also add orca+y into this to avoid qwertz vs qwerty keyboard layout differences and we would be all done with the changes. I am not sure if it can really be fully adopted this way but for sure it's a good proposal. Again what's most important to me that this is already verified on Windows with NVDA. So question for people knowing underlying API's and technical background: is it doable? Doesn't it introduce additional performance issues and other possible conflicts? Question for all ordinary users and technically inclined: Do you like it and can you add something into it to make it sound even better? Thanks and greetings Peter On 08.08.2014 at 06:15 Storm Dragon wrote: Howdy, |