Re: No policy? Re: Is EME usable regardless of the software/hardware I use ?

On Thu, Jun 6, 2013 at 7:58 PM, piranna@gmail.com <piranna@gmail.com> wrote:

> EME itself can be compatible with GPL-2, but EME alone are useless and
> require of CDMs, that are their main complement and that by definition are
> not compatible, so the tandem of EME-CDM combined is not compatible with
> GPL-2 at all.
>
Thank you for pointing out this distinction. It's important that we have
explicitly excluded from the EME proposal anything which could be
incompatible with W3C policies. I believe this includes anything which
might be incompatible with FOSS licenses. It would make no sense to propose
a specification to W3C which didn't have such material excluded.

Now, you, Duncan and others argue that *in practice* the only purpose of
EME is DRM and DRM is incompatible with FOSS (and also with W3C policies,
specifically on RF licensing). Further, that this should be a reason not to
specify EME in W3C, despite the paragraph above.

This amounts to a request that the entire implementation below any given
web API must be implementable in FOSS, however far down the stack you go.
As discussed in other threads, the request extends below the OS/Application
boundary and further to firmware running on hardware components of the
system. This may be a noble goal, but it's a very strong requirement.

Clearly there can exist FOSS implementations of the whole stack below EME,
but as with other web APIs these may not be suitable for all applications.
However, it's argued that there are no applications where such an
implementation would be suitable. I'm not sure how we could know this when
there remain no proposals for FOSS-compatible content protection. I think
the only way we could know this would be through experimentation - meaning
that such solutions are made available and we see whether or not they are
popular. The oft-cited AES+HTTP proposal - consisting as it did of a
handful of cursory emails on technical mailing lists - doesn't provide any
kind of interesting evidence IMO. I also find it strange that so many of
the people arguing strongly for non-DRM solutions to this problem do not
seem interested in developing or experimenting with them.

To drop EME on this basis would imply first that we give up immediately on
the possibility of a useful FOSS-compatible CDM *and* that we accept the
requirement that the whole implementation be FOSS-compatible.

...Mark

El 07/06/2013 04:52, "Jeff Jaffe" <jeff@w3.org> escribió:
>
> On 6/6/2013 10:07 PM, Duncan Bayne wrote:
>>
>>> I'm not an attorney, but I agree that the EME draft document may be
>>>> incompatible with GPLv3.
>>>>
>>> Definitely, I agree.
>>>
>>> Re. the premises you stated I held:
>>>
>>>     * A premise that W3C has a Recommendation in this space.  At the
>>>>      moment there is a draft proposal.
>>>>
>>> That's correct.  Sloppy language on my part; I was envisaging the state
>>> of affairs should the draft proposal proceed to a recommendation.
>>>
>>>     * A premise that EME = DRM.
>>>>
>>> The reason EME is being proposed is to enable DRM.  Netflix, Microsoft
>>> and Google are interested in it for no other purpose.  No-one (to my
>>> knowledge) has proposed that EME might be used for any *other* purpose
>>> than interop with DRM systems.  Therefore, EME is a component of DRM
>>> systems, nothing more, nothing less.
>>>
>>> However, note that I didn't mention EME in my premises.  I was quite
>>> specifically talking about CDMs, as they are the reason for the
>>> existence of EME.  I addressed CDMs because they're central to your hope
>>> that movie companies will abandon closed-source, proprietary DRM
>>> systems.
>>>
>>>     * A premise that GPLv2 (which may be consistent with EME) is not a
>>>>      FOSS license.
>>>>
>>> That's not my opinion.  GPLv2 is definitely a FOSS licence, and an
>>> implementation of EME could be compatible with GPLv2.  I think we're
>>> agreed on that, too.
>>>
>>> So, restated, & elaborated:
>>>
>>> Consider what will happen if the EME proposal is accepted, and becomes a
>>> recommendation.  Vendors will use this to interop with DRM CDMs (the
>>> sole purpose of EME).
>>>
>>>   - major content providers will not implement and release CDMs that can
>>>   be trivially bypassed
>>>
>>>   - a CDM released under *any* FOSS license is, by nature, trivial to
>>>   bypass
>>>
>>>   - therefore, no major content providers will release CDMs under FOSS
>>>   licenses
>>>
>>> This is equivalent to EME being incompatible with any FOSS license.  EME
>>> exists for one purpose, and that purpose is incompatible with FOSS
>>> licenses.
>>>
>>
>> I don't understand.  You said GPLv2 is FOSS.  You said that EME could be
>> compatible with GPLv2.   So how is EME incompatible with any FOSS license?
>>
>>
>>>  I hope my response above addresses your question.
>>>>
>>> Not exactly, but it's facilitated some clarification, which I greatly
>>> appreciate.
>>>
>>>
>>