4 ms·

The reasoning goes that EME is a standard that specifically allows for attaching a closed-source black-box to the browser in a standardized way. This code is no

The reasoning goes that EME is a standard that specifically allows for attaching a closed-source black-box to the browser in a standardized way. This code is not auditable by outside sources, and therefore could be a breeding ground for malware and exploits (standard open-source philosophy that "sunlight is the best disinfectant"---many eyes on open source software minimizes the opportunities for exploitable bugs to hide). We've seen codec software---even open-source codec implementations---used as an attack vector in the past. This reasoning, IMHO, ignores the fact that the user-agent still asks for consent, so the user may choose not to allow anything they can't audit. It's just the "Do you want to allow Flash to play this cool video?" story all over again.

danudey 9y ago
Presumably they feel the same way about all closed-source browsers? Also, the comparison to Flash is not quite right. For example, Flash was an entire black-box, top to bottom, with a runtime, codecs, UI, a JS engine, etc., which downloaded code from the internet to run it. Web DRM, on the other hand, is a standard that allows for an encrypted video stream to be decrypted. The only differences that I can see between this and the current state of web video are that now the stream can be encrypted. Everything else is still using Javascript, HTML, CSS, HTTP, h.264/265, etc. The thing that gets me is that open-source browsers are free not to implement this extension if they're not comfortable with (or opposed to) it. The theory is that if companies start encrypting their content then open-source browsers will have to follow suit, but we already have that kind of junk; the difference is that now, with EME, we can keep 99% of the stack open standards, open-source, and auditable, rather than having to rely on Flash or Silverlight to do our decoding. EME isn't the death of the open web, it's the death of the closed web. It's the death of awful, insecure, and obsolete technologies like Flash and Silverlight by taking the one single thing they're still good for and moving it to the browser. If EME failed and no one implemented it, Netflix wouldn't just stream their content unencrypted; they'd keep using Flash or Silverlight, or maybe implement something themselves so they didn't have to.
- gsnedders 9y ago
  > For example, Flash was an entire black-box, top to bottom, with a runtime, codecs, UI, a JS engine, etc., which downloaded code from the internet to run it. And, notably, the DRM modules can be sandboxed far more than Flash can.
davexunit 9y ago
The user often does not have a choice because websites that they must use for one reason or another will require these extensions, just like some required Flash. And that's assuming that the extension even works on their system. The extensions are native code IIRC. Ever get a "sorry, your OS isn't supported!" message on a website that you really need to use to get something done? It sucks, a lot. I don't know why the W3C would want to steer themselves in such a direction, it's the antithesis of what the web was supposed to be. The web is supposed to have a well defined set of standards that anyone can implement and use to view the web. This was never true in practice because of things like Flash, but the W3C never endorsed such a thing. Now they are, and that is sad.
pmoriarty 9y ago
"It's just the "Do you want to allow Flash to play this cool video?" story all over again." Another comparison to existing tech is Javascript. A lot of sites break if you disable Javascript in your browser, but if you enable it you open yourself to Javascript vulnerabilities.
- khedoros1 9y ago
  Open browsers have open, auditable Javascript engines. There can be vulnerabilities, but you aren't waiting for some company to put out a closed patch containing a fix. EME blobs are like Flash plugin blobs. So, it's similar in the sense that enabling JS increases the potential attack surface, just like installing EME/Flash/whatever does. But at least the JS engine can be examined by whoever wants to.
problems 9y ago
> This reasoning, IMHO, ignores the fact that the user-agent still asks for consent, so the user may choose not to allow anything they can't audit Aaaand we're back to the days of good ol' ActiveX.

danudey 9y ago

Presumably they feel the same way about all closed-source browsers? Also, the comparison to Flash is not quite right. For example, Flash was an entire black-box, top to bottom, with a runtime, codecs, UI, a JS engine, etc., which downloaded code from the internet to run it. Web DRM, on the other hand, is a standard that allows for an encrypted video stream to be decrypted. The only differences that I can see between this and the current state of web video are that now the stream can be encrypted. Everything else is still using Javascript, HTML, CSS, HTTP, h.264/265, etc. The thing that gets me is that open-source browsers are free not to implement this extension if they're not comfortable with (or opposed to) it. The theory is that if companies start encrypting their content then open-source browsers will have to follow suit, but we already have that kind of junk; the difference is that now, with EME, we can keep 99% of the stack open standards, open-source, and auditable, rather than having to rely on Flash or Silverlight to do our decoding. EME isn't the death of the open web, it's the death of the closed web. It's the death of awful, insecure, and obsolete technologies like Flash and Silverlight by taking the one single thing they're still good for and moving it to the browser. If EME failed and no one implemented it, Netflix wouldn't just stream their content unencrypted; they'd keep using Flash or Silverlight, or maybe implement something themselves so they didn't have to.

gsnedders 9y ago
> For example, Flash was an entire black-box, top to bottom, with a runtime, codecs, UI, a JS engine, etc., which downloaded code from the internet to run it. And, notably, the DRM modules can be sandboxed far more than Flash can.

davexunit 9y ago

The user often does not have a choice because websites that they must use for one reason or another will require these extensions, just like some required Flash. And that's assuming that the extension even works on their system. The extensions are native code IIRC. Ever get a "sorry, your OS isn't supported!" message on a website that you really need to use to get something done? It sucks, a lot. I don't know why the W3C would want to steer themselves in such a direction, it's the antithesis of what the web was supposed to be. The web is supposed to have a well defined set of standards that anyone can implement and use to view the web. This was never true in practice because of things like Flash, but the W3C never endorsed such a thing. Now they are, and that is sad.

pmoriarty 9y ago

"It's just the "Do you want to allow Flash to play this cool video?" story all over again." Another comparison to existing tech is Javascript. A lot of sites break if you disable Javascript in your browser, but if you enable it you open yourself to Javascript vulnerabilities.

khedoros1 9y ago
Open browsers have open, auditable Javascript engines. There can be vulnerabilities, but you aren't waiting for some company to put out a closed patch containing a fix. EME blobs are like Flash plugin blobs. So, it's similar in the sense that enabling JS increases the potential attack surface, just like installing EME/Flash/whatever does. But at least the JS engine can be examined by whoever wants to.

problems 9y ago

> This reasoning, IMHO, ignores the fact that the user-agent still asks for consent, so the user may choose not to allow anything they can't audit Aaaand we're back to the days of good ol' ActiveX.