The H.264 Debacle: We're Complaining to the Wrong People
osnews.comThis is complete nonsense.
Read this instead: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-Jun...
Straight from the horses mouth, that is the state of HTML5 video and why we got there. The posted article is built on the idea that HTML5 requires H.264 when that is not the case at all. The spec does not specify a codec for use on the video tag, just as it doesn't for the img tag.
More importantly, the authors claim "we can cut to the chase and try to get the HTML5 spec fixed--in which case the commercial vendors would have to fix their implementations in order to be considered compliant." is bogus as well. The spec is a delicate balance. If something is spec'd out, but major vendors are not going to implement it, then you have accomplished nothing by putting it in the spec, and in fact, the spec actually harmful at that point because people will believe it is implemented correctly. The spec is only useful insofar as it is a set of things that all the players can agree to implement.
I do recommend reading through the email I linked to, in it Ian Hixie, by responding to emails he received, lays out a lot of the thinking that guided the making of the spec.
The link is good, but you repeat the highly misleading claim that video shouldn't specify a codec because img tag doesn't specify an image format.
If that's the case then why did Hixie spec an interoperable codec in the first place? I think the answer is obvious when you count the number and quality of competing, royalty-free image formats that uphold the basic principles of the web versus the number and quality of competing, royalty-free video codecs.
Claiming that not specifying a codec will lead to anything other than a de-facto standard of the non-royalty free H.264 is ridiculously, shockingly disingenuous. It's just shady lawyering to claim that pushing for a codec, that by its very nature can't be in the spec, and having its rivals removed from the spec, is anything other than the equivalent of having it in the spec.
I never said that the spec should not specify a video codec, I was just trying to set the record straight. In fact, I didn't claim anything about the spec at all, I just summarized some of Hixie's points. If you aren't satisfied with what I wrote, I recommend reading Hixie's words. Please let me know if I misrepresented him. And, I disagree that not having it in the spec is the equivalent of having it in the spec. The problem is that the spec is useless when, as Hixie puts it, it becomes fiction. Mandating either Theora or H.264 in the spec would have been useless because either Firefox/Opera would be violating the spec or Safari/IE would be. So the only option left to them was to leave it out of the spec, something that Hixie was not happy about but seems like the right thing to do. If a codec comes along that all the vendors can agree to implement, then it is likely that will go into the spec, but until then it will remain blank.
I agree with Hixie that putting it in the spec is pointless if Apple and Microsoft are going to go against the spec.
I disagree that just by forcing the spec to be changed from requiring Theora that Apple and Microsoft now get off the hook for their actions.
You said, repeating an often made claim, The spec does not specify a codec for use on the video tag, just as it doesn't for the img tag. Which is factually correct now, after Apple refused to implement the spec as it was and Hixie edited the spec to reflect reality. Hixie claims he will add a codec back, once one is found suitable, which means both royalty-free and implemented by the big players.
You appear to have done it by accident, but you'll find comments elsewhere in the thread that make the argument that he has no right to specify any codec in HTML5, whether Theora or a future royalty-free codec, and make the parallel to the img tag not specifying an image format. This is clearly not Hixie's position, and I don't think it generally holds any water.
>I agree with Hixie that putting it in the spec is pointless if Apple and Microsoft are going to go against the spec.
I disagree.
The spec should enable the web to be used unencumbered by patented software/codecs.
Ogg Theora or any other free+open container and codec (VP8?) would enable people to use FOSS software to create video and put that video on the web without being beholden to the H264 patent holders group ("MPEG LA").
Microsoft or Apple Computers or whoever would then be required to implement a viewer in their browser in order to comply with the spec. If they don't then they don't get to label themselves "standards compliant" without committing fraud and can't badge themselves with W3C compliance without trademark/copyright infringements.
If Opera, Firefox, Konqueror, etc., implement the full spec this applies pressure to obstinate commercial megoliths to also move to implement it.
What we have here is akin to knowing that the GIF patent situation (Unisys trying to get money from nearly everyone on the internet) is around the corner, already having PNG but not requiring that browsers implement PNG.
I think the situation is far worse than the GIF situation though, eg http://www.osnews.com/story/23236/Why_Our_Civilization_s_Vid... . Now http://www.engadget.com/2010/05/04/know-your-rights-h-264-pa... Engadget did a review of the situation in which the MPEG-LA say that they're not really going to hold their license terms (which mean that you can't shoot film on _any_ camera and use it commercially without purchasing a license) against people and basically they're really nice and we shouldn't be afraid of them. Which I agree, we have nothing to fear until the spec is widely implemented and the open license on H264 gets pulled out from under everyone in 2015 and MPEG-LA attempt to pull a Unisys.
Very informative, Thanks.
A lively discussion, but most seem to miss the fact that the spec did require Theora (though not exclusively, just as a baseline to ensure interoperability) and Apple, with a little help from Nokia, had it removed from the spec because they refused to implement it.
So if Apple aren't the right people to complain to, I don't know who is. (Probably if Microsoft had implemented Theora then it would have been added back to the spec, as Apple would have been in a clear minority then, so complain to them too).
The working group tasked with creating HTML5 has a charter that defines what are they supposed to work on and deliver. The codec of the content served through the video tag is outside of the scope defined in said charter. Apple, among others, have argued only that. What Microsoft could have done might have had impact on the outcome of h.264 vs. Theora, but it would not have changed the HTML5 specification.
Can you actually point to Apple arguing that about the charter scope? I recall them arguing that Theora was not efficient enough, that it didn't have sufficient hardware support, that it had patent issues and probably a few other arguments. I don't remember them ever making that specific point about the HTML5 charter.
Unfortunately I cannot off the cuff cite a source (if I find one I'll add a link to it, though). I can tell you the context in which I came across this. A couple of months ago there was a story that Adobe were "sabotaging" HTML5 - in reality it was Larry Masinter, Adobe's representative in the HTML WG, arguing that the canvas API falls outside of the charter of the working group. In one of the blogposts in defense of Adobe's position the author quoted similar objections that Apple and Microsoft have made regarding <audio> and <video> codec support when HTML5 was still in transition from WHATWG to W3C.
Apple have most certainly expressed concerns about Theora's performance and patent situation and I'm sure they still have them. However, those concerns were pertinent when HTML5 was being developed by WHATWG, because WHATWG are not limiting the scope of any work they do. That's not the case with W3C's HTML WG.
I think you could throw Google's name on the list there as well. Ian Hixie says:
>Google ships both H.264 and Theora support in Chrome; YouTube only >supports H.264, and is unlikely to use Theora until the codec improves >substantially from its current quality-per-bit.[1]
If the most popular video destination on the web continues to use H.264, then Theora adoption is a little less meaningful. (but still meaningful)
[1]: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-Jun...
I don't really mind Youtube (or Apple Keynotes, Quicktime Movie Trailers etc.) not using royalty-free codecs that much. If every web decision was made through the lens of "is this image format suitable for Flickr", or "is this text encoding suitable for Wikipedia" then you'd end up with a very strange kind of web. Youtube seems to be regularly predicted to kill the internet as a whole anyway due to its size and popularity, with or without Theora.
If anyone is going to make extra work for themselves by using multiple codecs, or be subject to licence fees then it should be the giants of the web, not folk uploading videos of their cats to their blog (a demographic for whom Theora is a perfectly good solution if it was supported by Apple and Microsoft).
I agree with you. But even so, wouldn't a lot of people (everyone using firefox/Opera) then be unable to use youtube? When you have such an amazingly high profile site that is incompatible with major browsers, isn't that a bad thing? I guess you just fall back on flash.
W3C is in no way, shape or form "allowing H.264 to infiltrate its way". The HTML5 spec does not require or endorse h.264 or any other codec. Opera's, and Mozilla's I believe, attempt to sneak a codec requirement in the spec (Theora) was shot down, and rightly so, because it falls outside of the scope of HTML5, just as the encoding of image files displayed through the img tag is outside of its scope.
"""The fact is, the W3C is violating its own principles by allowing H.264 to infiltrate its way into the next HTML spec."""
I think this is a very valid point, w3c is suppose to promote open tools for the web of consumers and authors regardless their technology. A patented tool does not promote this.
Within W3C's mission statement ( http://www.w3.org/Consortium/mission ), this is not mentioned. Can you provide a reference to back up this claim? (that is, that W3C is supposed to promote open tools or would violate any of its principles by not requiring to support any particular video encoding)
What kind of bullshit is this? The spec does not forbid you to use any codec you want. Or do you want it to ban H.264 so inferior codecs could have a chance? Way to go.
Horse shit.
If the spec forbade "the" codec, the offending vendors will simply violate the spec and claim some kind of justification for it.
Yes, we're complaining to the wrong people. No, going to the W3C is not the answer.
I tend to think that if we can get people to stop using h.264 for anything, then freer codecs could gain traction - particularly On2's (now Google's) VP8.
Or, taking the bigger perspective, we should be complaining to the US Congress: patents are the problem, not Microsoft's choice in codecs. If there were no patents to license, this wouldn't be an issue at all.
The real fix would be to abolish software patents, and with it, the licences.