TV service providers typically dedicate a channel to advertize their programmes through a Mosaic displaying several channels at once.
From this specific channel, the end user is able to pick one of the programmes and switch to the corresponding channel.
The UC would be to allow a web application to provide mouse and keyboard interaction with the sub-sections of a video frame corresponding to the channels of a TV mosaic without requiring too much specific code.
The UC can be solved with existing technologies in at least two ways:
First, the web application can apply a set of well-organized HTML elements as an overlay to a video element, and associate anchors to these elements to trigger the appropriate action when the end user activates them.
Second, the application can catch all input events received by the video element, and apply a specific calculation algorithm to trigger the appropriate actions when the portion of the video corresponding to a channel is meant to be selected.
Drawbacks of these solutions:
Note: What about accessibility ?
Using a declarative syntax to describe the areas corresponding to the mosaic channels in the video frame and associating actions or links to them would be much simpler and would preserve the ability to use the video tag native controls.
This requires an evolution of the HTML5 specification
Since the HTML5 map element already provides the expected features to the img or object tags using only a declarative syntax, the proposal is simply to allow the HTML5 map element to be applied also on video elements.
How about simply tuning each channel into individual <video> elements and on clicking "zooming in" that channel while ramping up the volume? That seems to make better use of the capabilities of the platform as a whole...