[sdw] What would implementing WebVMT mean for user agents?

tidoust has just created a new issue for https://github.com/w3c/sdw:

== What would implementing WebVMT mean for user agents? ==
First examples in the [Proposed Solution](https://w3c.github.io/sdw/proposals/geotagging/webvmt/#proposedsolution) section show an HTML page composed of a `<video>` element and a `<div>` element that is to contain the map.

The link between the `<div>` element and the WebVMT track seems to be made within the WebVMT file itself through the `div-id` map property. I believe that this is wrong: the WebVMT file is embedded in the HTML page, and should not know anything about the page. The same WebVMT file could be used in different HTML pages for instance. The association with the `<div>` element should rather be made in the HTML page itself.

More broadly, this begs the question of what it would mean for a user agent to implement WebVMT. Does it:

1. Parse WebVMT files and trigger map cues at the right time, as `VMTCue` instances (interface which has yet to be defined)?
2. And render the map all by itself alongside the video?

User agents need to natively support map rendering to be able to do 2. They do not for the time being, and the Web platform does not have any notion of map for the time being. Something like the MapML proposal could perhaps change that situation in the future.

It would be useful to clarify the goal. If it includes 2., I believe this introduces a dependency on defining a notion of map on the Web, as linking to a generic `<div>` element and expecting rendering magic to happen is likely going to trigger interoperability issues, especially if you envision that the user can interact with the rendered map (pan, zoom, rotate, click, etc.). That is not necessarily a huge problem per se, but seems worth keeping in mind.

Alternatively, 2. could be something we envision as a longer term goal, but in that case we'll probably need to think about the transition between implementations that only support 1. and implementations that support both 1. and 2, e.g. how an application can detect that a user agent supports map rendering.

Please view or discuss this issue at https://github.com/w3c/sdw/issues/1019 using your GitHub account

Received on Friday, 16 March 2018 09:54:13 UTC