RE: [Exposed] attribute on our APIs

> -----Original Message-----
> From: Dominique Hazael-Massieux [mailto:dom@w3.org]
> Sent: Wednesday, May 20, 2015 2:57 PM
> To: Mathieu Hofman; Martin Thomson; Harald Alvestrand
> Cc: Stefan Håkansson LK; public-media-capture@w3.org; Iņaki Baz Castillo;
> annevk@annevk.nl
> Subject: Re: [Exposed] attribute on our APIs
> 
> On 20/05/2015 21:57, Mathieu Hofman wrote:
> > While I don't want to delay LC, what's the best approach to make sure we
> can one day support these use cases?
> 
> If we think this isn't likely to be spec'd/implemented in the next few months
> (which I think is correct), the best approach is probably to delay this to a next
> version or an extension of the Media Capture and Streams specification.
> 
> If that's indeed the approach we take, I think it would be worth documenting
> the use cases you highlighted somewhere (a github issue? a "future use
> cases" document?). Alternatively, someone (you?) could maybe start
> drafting what such an extension would look like.
> 
> Dom

>From what I understand, the changes to the spec are fairly straightforward:
- Mention that MediaStream(Track) implements structured cloning (by implicitly calling the `clone()` method) in order to be used with HTML5 Web Messaging's 
postMessage API
- Make sure MediaStream(Track) are exposed to worker context so that these objects can be used from a Web Worker

We need to get the wording right but that's probably not too hard. Stefan gave it a stab back in October 2013 (search archive for message titled " Make MediaStream usable with postMessage") 

However I'm confused by the relation with the likeliness of this being quickly implemented by browsers.
There are plenty of other specs which have parts that are not implemented for years even after reaching Candidate Recommendation. A relevant example is the ability to use a WebSocket in a Web Worker, which has been part of the 2012 CR and has only landed in Firefox as of version 38.
If we can agree on the spec, and that implementors agree it's feasible (though possibly complex), would it be possible to include this in the current version?

I will try to put together a list of use cases, and update the previously proposed wording to the current spec.

Mathieu

Received on Wednesday, 20 May 2015 23:02:07 UTC