A few weeks back, I wrote about the troubling prospect of DRM being baked into HTML5. At the centre of a related piece was a post by Brendan Eich, CTO and SVP of Engineering for Mozilla. As I noted then, it was somewhat opaque, in that I found it hard to understand how exactly Mozilla intended to react to the W3C's pernicious proposal to discuss DRM – specifically, the idea of adding Encrypted Media Extensions (EME) to HTML5. By a happy chance, Eich was passing through London recently, and so I was able to find out more about Mozilla's attitude and plans in this area.
He began by talking about codecs – another hot subject in the light of Cisco's so-called open-sourcing of H.264, and something else that I was sceptical about. Eich noted that in the mobile world, encoders for video recording are simply indispensable, otherwise no one will buy your smartphone:
We were in a position of not being competitive on mobile unless we supported H.264 as Chrome did all along, and as Android did all along. We're trying to move the world piecewise towards zero-price, free as in beer for these patent-encrusted formats.
However, that does not mean that Mozilla is happy with "patent-encrusted formats". As I discussed a couple of weeks back, the organisation is now working on Daala, a fully open next generation codec. That seems to be the broad approach for DRM, too: to support the bad stuff as little as possible, while working towards better solutions. Here's Eich explanation of what's happening here:
We're not in a position, having not rejected plugins [like the one for Microsoft's Silverlight] to say "no DRM" - we accommodated DRM through plugins. If we had been serious about open source we would have banned plugins. By serious, I mean exclusive or absolutist. Instead we have to move the world incrementally and get enough users to have scale with developers and therefore influence with standards bodies. If we say no to H.264, if we use only free and unencumbered codecs, if we say no to DRM whether plugins or the new CDMs [Content Decryption Modules] in EME – we'll probably lose share.
And as we take more of this risk, our purity point will be high with certain users, which we like, they're very valuable and important users and we respect them. But the mass market of users across Geoffrey Moore's chasm - where we went in 2005 after Firefox 1 – most of those users sadly don't know what's going on and they don't really care. I hope that explains why we're not all or nothing on this topic.
The key thing that emerged for me was that Eich felt that Mozilla had to be pragmatic here so as not to endanger its installed base and thus lose its power to move things in directions that are helpful rather than harmful to the open Web and free software. That reflects the organisation's position in the open source world, where it has tended to prefer this approach rather than refusing to compromise. Whether you agree with that probably correlates strongly with your views on Richard Stallman.
Eich went on to explain what he thought the EME moves within the W3C were really about:
EME is really a fig leaf over what is already shipping in Windows, now in Chrome OS. As a standard it's woefully underspecified, so you can't have anything but a toy clear-key CDM. That's not a robust system according to Hollywood, so no major studio will trust it.
This means that something particularly troubling is taking place:
If browser share depends in part on these streaming services being available, and if the EME fig leaf is just concealing a vertical integration from hardware up through OS and browser vendor and DRM vendor – where Microsoft has PlayReady, and Google has acquired Widevine and Apple has Fairplay – that could be a serious threat to Mozilla.
The question then becomes: what are Mozilla's options in this vertically-integrated world?
We can say: we reject it outright. We can say: oh, this is a new kind of plug in – can we have a new plugin like Silverlight? And then the problem becomes robustness. A lot of these systems are tied into not only the browser but the operating system. The worst case, which I do not believe will happen, we see the Web perverted into walled gardens - a sort of underverse of content where DRM is everywhere, and you have to opt into Time-Warner, Sony, Fox or Disney.
In that case, Mozilla might be forced "to look for side deals with DRM/OS/browser vendors," as Eich says, "that's not a good situation to be in." As with H.264, what Eich hopes will happen is that a bad solution – DRM – can be replaced with a less bad one:
Watermarking is better than DRM. We are still expecting Hollywood, one of the big six, to announce a streaming service like that. I have no news there. We are working behind the scenes, and this is another example of Mozilla not being so pure that we just abhor all these players. We're doing what we can.
The big question is: why is the W3C going down this route? Eich's views are not comforting:
W3C has gone off the rails again just like it did in the past, by just looking for paying members. To hear both Jeff [Jaffe, CEO of the W3C] and Tim [Berners-Lee] talk about it, they [say]: well, shouldn't we have everybody in the world join, so we can all make the best Web standards we like? Maybe not, because a lot of them are commercial interests that are operating against the open Web principles. Inherent conflicts among these members are going to be make bad specs or no specs, or cause litigation anti-patterns or anti-specifications to be encoded, to leave everything open to a particular OS, or to let the court sort it out later.
Eich points out why this problem is not likely to go away soon:
So long as people want Hollywood movies, and Hollywood is used to getting its way, and DRM vendors are pushing to perpetuate this and codify it, and put it under a fig leaf from the W3C, we have a real problem.
Although Mozilla could object to the W3C, Eich believes that might turn out to be a terrible idea:
We can make a formal objection, and that gets resolved by the [W3C] director, and that's Tim. Tim has said some things that seem to go even farther. He 's talked about why not make egalitarian DRM, DRM for all content. And it's actually a really bad idea. If we appeal to Tim, he's perhaps going to do something which would make DRM something of a virtue that we should spread among all of the Web standards. That would be a disaster.
And that's putting it mildly.
If I didn't before, after talking to Eich I had a strong sense that Mozilla is constrained by conflicting desires – to do the right thing, for example, while retaining enough browser share that it remains able to do the right thing. As you might expect, there are no easy solutions. Fortunately, things seemed more hopeful regarding another topic we discussed: I'll be exploring that in my next column.