Re: EME and proprietary plug-ins

[Moving to public-html-media and returning to the original point from Henri] …

On Feb 12, 2013, at 6:51 AM, Henri Sivonen wrote:

> On Tue, Feb 12, 2013 at 4:01 PM, Robin Berjon <robin@w3.org> wrote:
>> The question that lies open before us is: given that DRM exists, should it
>> be implemented through proprietary plugins or should it be possible to hook
>> it somehow into the open web platform?
> 
> That's not really the question EME poses. Rather, EME poses the question:
> 
> Should DRM be implemented as a large non-user-modifiable black box
> whose licensing characteristics are known[1] with a broad existing
> API[2] that doesn't integrate into <video> or should it be implemented
> as a smaller non-user-modifiable black box whose licensing
> characteristics are unknown with a narrower yet-to-be-defined API[3]
> that integrates into <video>?
> 
> It's easy to see making the black box smaller and the API narrower as
> an improvement over the status quo. Moving from known licensing and
> API to yet-unknown licensing and API (potentially multiple licensing
> models and APIs) may not be an improvement.

This is a good characterization of the issue and demonstrates the significant overreach of some of the more florid rhetoric we've seen (from others) on this topic.

I would add the following points though:

- the existing plugins do not support all platforms. Much less than half our video streaming is on platforms where SL is supported, for example. Something new is needed if we want a solution for all platforms (I am thinking mainly of TVs and like devices here, but also mobile platforms).
- the licensing and API situation with CDMs is indeed as-yet-unknown and so we can speculate that it may not be an improvement (for the subset of platforms where Flash/SL are available today). But we could equally speculate that it may be an improvement. The same pressures to get wide deployment as applied to Flash/SL will apply to CDMs. The EME approach, with multiple CDMs and common encryption, means that content need not be locked to a specific DRM. This makes it easier for one DRM to take market share from another - for example by doing things that promote wide distribution and support of their client component. This is to say that the incentives are present to address this aspect.
- a further difference is that we expect browsers to be in control of which CDMs they support and so able to offer greater protection to users (in terms of privacy and security) than they do with the plugins. For example a browser could place whatever security/privacy requirements they feel appropriate on CDMs as a condition of integration.

…Mark

> 
> [1] On Mac and Windows, the user gets the DRM box (Flash or
> Silverlight) for $0 and the browser vendor doesn't need a contractual
> relationship with the DRM vendor.
> [2] NPAPI
> [3] EME doesn't define the API between the CDM and the browser.
> -- 
> Henri Sivonen
> hsivonen@iki.fi
> http://hsivonen.iki.fi/
> 
>

Received on Friday, 15 February 2013 21:43:59 UTC