Expose special HTML content as voice-activable commands in UWP

[jdm]

XAML buttons in UWP automatically get support for voice commands on HoloLens devices. It could be a great extension of this to recognize certain parts of web content (effectively piggybacking off of accessibility roles) and expose them to the same mechanism.

[jdm]

I imagine this would look something like:

• when certain roles are recognized by HTML element attributes being set, notify the document
• the document is responsible for creating a mapping (ie. uuid) to the element and notifying the embedder about the element and its role
• the embedder exposes a new voice command based on the text provided with the event
• if the voice command is recognized, the embedder activates a servo API with the associated identifier
• this API sends a message to the document which retrieves the original element and invokes a trusted click event on it

[jdm]

https://github.com/Microsoft/Windows-universal-samples/blob/master/Samples/SpeechRecognitionAndSynthesis/cpp/Scenario_ContinuousDictation.xaml.cpp
https://docs.microsoft.com/en-us/windows/mixed-reality/voice-input-in-directx