W3C home > Mailing lists > Public > public-html@w3.org > February 2012

Re: Encrypted Media proposal (was RE: ISSUE-179: av_param - Chairs Solicit Alternate Proposals or Counter-Proposals)

From: David Singer <singer@apple.com>
Date: Mon, 27 Feb 2012 12:30:45 -0800
Cc: Adrian Bateman <adrianba@microsoft.com>, Maciej Stachowiak <mjs@apple.com>, "HTML WG (public-html@w3.org)" <public-html@w3.org>, David Dorwin <ddorwin@google.com>, Mark Watson <watsonm@netflix.com>
Message-id: <24435032-6809-446B-B627-1A045D98296F@apple.com>
To: "Vickers, Mark" <Mark_Vickers@cable.comcast.com>

On Feb 27, 2012, at 12:18 , Vickers, Mark wrote:

> David brings up the interesting issue of support for content protection in the existing spec. I have related questions for David and the proposal authors:
> 
> David Singer:
> 1. How would the application and user agent communicate the particular content-protection scheme? Would it be in the source element type and/or codecs attributes? Can you give a concrete example?

This is kinda assuming that I need to stop beating my wife :-).

If the content is marked, inside the content file "this is protected by ExampleDRM and the content-ID is XXXX", then someone who has ExampleDRM installed and has bought keys to content XXXX will be able to play it.  ExampleDRM may also be able to prompt "to watch this, you need 30 bezants in silver" if you have not yet bought keys…


True, it helps if the HTML source tag warns that a specific DRM is needed.  We could (and probably should have) defined the 'codecs' parameter for MP4 files to have, when the codec is 'encv' or 'enca' (protected content), a subtype of the scheme type; but that helps with selection only, playability is unaffected.  We should have done this in the recent 6381 revision of that parameter, dang.

It's only if you assume that content protection is managed by a combination of scripts, user-agent, secure code, and so on, do you need extra signalling.  That may be the case sometimes, but it's misleading to say it always is.

I'm not saying that more general problems don't need solving; I just want to start from an agreed set of underlying assumptions.

> 
> 2. How would the application key server connect to the user-agent key client? 

again, same assumption.  why is it always necessary that this be so?


David Singer
Multimedia and Software Standards, Apple Inc.
Received on Monday, 27 February 2012 20:31:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:17:46 GMT