Comment by d--b | Hacker News Reader

Comment by d--b | Hacker News Reader

What people don't get is that EME sets a strange precedent in the history of HTML.

Web browsers have always been very hackable. HTTP meant you could always look at the traffic being exchanged. And because there was little point in obscuring anything, web browsers allowed you to look into and modify everything:

- view/modify document source

- view/modify DOM

- debug script

- and so on

This is how CSS was defined too. It was supposed to be a compromise between how the user liked things styled, and how the vendor suggested styling the content.

EME brings something new to the table: locked LOGIC. This not a bad thing per se but takes html in a completely different direction than what it used to be.

The main concern is not DRM. The main concern is that this is a step in a direction where web browsers become unscrutable virtual machines running code that cannot be looked into. It's basically a step turning HTML into silverlight. This may happen for instance if the gaming industry decides that they need EME for in-browser games.

And most importantly this is done for all the wrong reasons: EME cannot stop anyone from copying the rendered content. And it certainly doesn't prevent anyone from downloading copied content.

So EME is just a stupid thing that technology-dumb media dudes are imposing on web developers for no reason and that may have far reaching consequences on the future of html... That's what's worth talking about.

I do not understand DRM/silverlight properly. Does this means that I won't be able to see what code is running in my browser for eg. JS collecting data in the background or use things like adblocker, greasemonkey etc.

No. The changes that are happening only affect DRM. So it means that when you watch a video/audio in the browser, there will be some encryption applied deep within the browser.

But it is a paradigm shift in the sense that it is the first time that a change is introduced in HTML that restricts the user from accessing the actual content that is being displayed.

Effectively, similar changes to the standard could be introduced to prevent users from tinkering with ads, or to prevent users from opting out of data collection.

In that sense, it opens the way to your browser being more like your phone.

Again, it's not _bad_ in itself, but it is a paradigm shift.

> I won't be able to see what code is running in my browser

Yes, but it's not JS. It's a plug-in the EME spec calls "CDM", and it is allowed to be anything, including kernel modules and hardware.

Websites that sign contracts with CDM plug-in vendors can use these unrestricted plug-ins, but you aren't allowed to check what they're doing because of DMCA.

HTML have always had plugins too... this ends that.

Think of EME as a slightly better sandboxed plugin with a single very small purposes... As oppose to the security horrors from Flash

If there aren't already enterprise software firms investigating how to offer their SuperWidgetron 10000 HR system through EME in the browser, I will eat my hat.

The single very small purpose will soon enough be worked around, messed with and extended to do more than it was ever "supposed" to - just look at JavaScript.

EME offers an interface to decode audio and video stream. How is what you're talking about even possible in that context?

Some kind of virtual webpage that's delivered over video?

Where is the source code of the DRM module that shows that the only thing the code do to the stream is decode audio and video?

Oh, that part is not part of the open part of the standard. The super cookies, the administrative hooks into the operative system, the access to everything from the webcam to the hard drive, all that is what the EME is talking to through an interface labeled audio and video stream. A perfectly sand boxed interface. The buss transporting data from the browser to the module is not very useful as an attack surface when you full control of the module itself.

Those are good questions – and you might find Firefox's sandboxing implementation worth study – but that's not what we were talking about. Currently there's no talk of implementing restrictions on code, etc. within browsers and it hurts the cause to spread easily disproven claims that it's happening now.

Reply on news.ycombinator.com