On Aug 19, 2013, at 22:46 , [email protected] wrote: > Well, with this constraints I believe I will not be affected, so then I > should not be worry about EME since I don't think I will never need to use > media protection the same way it was not offered to me until now (I use > Linux), but what about the ethical questions? Also, as a standarization > organism, should W3C promote only standards that don't require third party > elements to work? In other words, shouldn't it promote only standards that > can run anywhere and deployed from scratch from a technical point of view?
The virtual interface of the web layer is something that needs underlying resources. I am actually arguing the same as you in a different context -- device-specific APIs. Given, for example, a Vibration API, one can construct a web site that is not usable by devices without a vibration ability, or by people unable to detect vibration. The web has (for the most part) relied on an abstract layer of screen+audio+keyboard+pointer, which can be instantiated in various physical ways. I agree that piercing this veil of abstraction is dangerous, be it for software capabilities or hardware. Plug-ins of any kind suffer from this problem. EME may reduce the 'surface' of the plug-in (from handling the media and the interaction, as well as the protection, to handling only protection) but I agree, it doesn't eliminate it. I can't see how, right now. > EME would allow to play CDMs in a binary blob, but if it's not available for > my OS, what's suposed I should do? Use a diferent OS? Wouldn't it be better > that W3C mandates that CDMs specifications are available so they could be > implemented by third parties (me)? This would reduce the amount of available > CDM mechanism, but also would allow to have a common base available from > everywhere and also they will be the best, flawless ones… It's been discussed before, but there is a trust chain involved in content delivery. The distributor has to trust that the client software is abiding by their desires to make it difficult enough to make an unprotected copy. If the client interface (the network transactions) is open to anyone to implement, it's hard to see how to do that. The SUN OMC took the approach of using code signatures, I think. David Singer Multimedia and Software Standards, Apple Inc.
