W3C home > Mailing lists > Public > public-html-admin@w3.org > February 2013

Re: EME and proprietary plug-ins

From: Mark Watson <watsonm@netflix.com>
Date: Wed, 13 Feb 2013 23:48:34 +0000
To: Tab Atkins Jr. <jackalmage@gmail.com>
CC: Glenn Adams <glenn@skynav.com>, Henri Sivonen <hsivonen@iki.fi>, "Chris Wilson" <cwilso@google.com>, Robin Berjon <robin@w3.org>, Andreas Kuckartz <A.Kuckartz@ping.de>, Sam Ruby <rubys@intertwingly.net>, Fred Andrews <fredandw@live.com>, "public-html-admin@w3.org" <public-html-admin@w3.org>
Message-ID: <EFBF822F-B9E4-4E52-AD7F-4C2EB4B513FB@netflix.com>


Sent from my iPhone

On Feb 13, 2013, at 1:15 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:

> On Wed, Feb 13, 2013 at 11:26 AM, Glenn Adams <glenn@skynav.com> wrote:
>> On Wed, Feb 13, 2013 at 11:00 AM, Tab Atkins Jr. <jackalmage@gmail.com>
>> wrote:
>>> On Wed, Feb 13, 2013 at 9:13 AM, Glenn Adams <glenn@skynav.com> wrote:
>>>> On Wed, Feb 13, 2013 at 12:50 AM, Henri Sivonen <hsivonen@iki.fi> wrote:
>>>>> EME poses the threat of unleveling the playing
>>>>> field for browsers even within operating systems in addition to
>>>>> keeping the playing field unlevel among operating systems.
>>>> 
>>>> If unleveling means moving away from the status quo of using only
>>>> Flash/Silverlight for distribution of protected media content, then you
>>>> are
>>>> correct.
>>> 
>>> You know what Henri means, and it's not that.  He means, very
>>> obviously, that individual browsers may be locked out even on a given
>>> OS.
>> 
>> I understand Henri's point, but I believe it based on speculation and not
>> necessity. I believe EME can function perfectly well on all OS/UA
>> combinations, and the issue of which CDMs will be available on those
>> combinations cannot be predicted. For example, I don't believe one can claim
>> with certainty that any given CDM will not be available on some OS/UA
>> combination, e.g., on Ubuntu/FF.
> 
> I believe we have an actual statement from Mark (Netflix) that they
> don't expect widely-used DRM modules to be available on Linux.  

If I said that it was a mistake. What I may have said is that if you have a 100% GPLv3 system then it can't include a DRM module, because DRM requires non-user-modifiable components and GPLv3 explicitly forbids this.

> (I'd
> have to dig up the email to find his exact words, but I believe this
> is an accurate paraphrase.)  Aside from
> 
>> The choice of which CDMs can be used for deploying some given content will
>> be determined by the content owners, as is their prerogative. The existence
>> of the EME solution that supports a variety of CDMs will increase the number
>> of options for content owners.
> 
> Content distributors are neither the only nor the highest constituency
> we care about.  (They're roughly a subset of "authors".)  Something
> that's good for them (increasing choice) can still be bad for the
> platform as a whole if it's bad for users and the rest of the
> "authors" category, or very bad for browsers and the technical purity
> of the platform.  Henri, myself, and others have been arguing that the
> negative to the rest of the platform greatly outweighs the positive to
> content distributors.
> 
>>>> The issue of OS playing field is a non-issue.
>>> 
>>> Several people disagree.  I'm not sure why you think it's a non-issue,
>>> or why you think it's *such* a non-issue that it can be dismissed
>>> out-of-hand like that.  Producing technologies that will only be
>>> usable on particular OSes is a bad thing.
>> 
>> There is nothing about EME that prevents it from being implemented on any
>> OS. Whether a given CDM is supported on an OS/UA combination or natively in
>> a OS is a deployment decision outside of the scope of the EME specification.
>> There is nothing in principle that prevents any CDM from functioning on an
>> OS. That's why this is a non-issue.
> 
> As we discussed a few months ago when this started, ignoring reality
> and pretending that everything will turn up roses is simply wrong.  We
> *know* from existing DRM experience outside the web platform that most
> DRM modules work on only one or two platforms.  A lot are written only
> for Windows (or just specific versions!) or for Windows and Mac (or
> just specific versions of each!), because supporting other platforms
> is extra engineering effort that the DRM vendor doesn't find
> sufficiently profitable.
> 
> I don't think it's possible to honestly theorize that *this* time
> it'll be different, and DRM vendors will magically decide to expend
> the necessary effort to support the same set of OSes that browsers do
> today.  Browsers have users on a lot more things than just "the latest
> version of each major OS", and if it looks likely that many of those
> users will be locked out of modern content using DRM through EME,
> that's a problem.  (Flash, at least, still supports a lot of older
> platforms.)

The problem is the same - only worse - with Flash/SL. The vendors of those components will only support the embedded DRM capabilities on the platforms of their choosing. For them
to support the whole of Flash/SL on a platform is a lot more work than supporting a CDM with far more constrained functionality.

It's been said that Flash/SL can be made to work on a platform without explicit support from the vendor, but I think it unlikely that content providers would entrust their content to a client software stack which is not explicitly
supported by the vendor of the content protection being used.

> 
>>>> EME will enable new opportunities, while the status quo keeps content
>>>> locked
>>>> out from the web or locked into the Flash/Silverlight solutions.
>>> 
>>> The "new opportunities" are locking content into new plugins.  It's
>>> not materially different from the status quo, and you shouldn't try to
>>> pretend that it is.
>> 
>> It is substantially different from the perspective of those content owners
>> and content providers that are actually delivering content. I don't feel it
>> necessary to repeat again the reasons I and others have cited before.
> 
> I understand that the content distributors consider Flash's and
> Silverlight's DRM to be different from their preferred DRM module.  To
> everyone else on the platform, the difference is unimportant (after
> all, it's ideally invisible), except in how it affects them.  Henri
> and others have argued well for how it is probable that the EME spec
> will produce a "worse" DRM situation for the other parties in the
> platform.
> 
> ~TJ
> 
> 
Received on Wednesday, 13 February 2013 23:49:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 13 February 2013 23:49:03 GMT