Google and the <video> tag

Though H.264 plays an important role in video, as our goal is to enable open innovation, support for the codec will be removed and our resources directed towards completely open codec technologies.

From HTML Video Codec Support in Chrome

Well that sucks.

Gruber asks a few “simple questions” here. Aside from the question of hypocrisy w.r.t. Flash bundling, I think his points are more than neutralized by these ten questions:

You are a proponent of Apple using its influence to diminish the importance of Flash for the web. Yet, when Google makes similar moves to rid the web of a similarly closed and patented, albeit different type of technology, you do not support them. Why is Apple promoting an open web a good thing, but Google promoting an open web a bad thing?

I think that if Google pulled Flash support from Chrome there would be no question that Google were on the side of the angels (although it would still be a dumb thing to do), but since there’s no hint of this it seems purely like a cynical move to hurt Apple’s anti-Flash campaign which will damage HTML5 <video> adoption. I think you can make the argument that HTML5 <video> adoption with H264 as the defacto standard codec is a Bad Thing.

Anyway it’s a bigger mess now than it was before Google decided to do this. Ultimately it will come down to “what will let the most people see the most porn using the most devices?”

Postscript: “standards”

One of the arguments made in favor of WebM/VP8 is that it can be part of the W3C standard, unlike H264, because it’s not encumbered by license fees. The problem here is that WebM/VP8 almost certainly is encumbered (as was GIF in earlier days), it just hasn’t been sued yet because no-one uses it. But this is beside the point — the CSS font-family property supports any font, and almost all the fonts that anyone cares about are encumbered (i.e. subject to royalties, copyright, and so on). Just as CSS font-family can specify a non-free non-open-source font, there’s no reason why a video tag can’t point to an arbitrarily encoded video.

To put it another way:

There’s no conflict between the HTML specification being open and royalty-free and H264 video playback being supported in HTML5 video tags as long as the codec doesn’t need to be implemented by the browser. Just as a slab of text with font-family “Verdana” won’t necessarily display on every browser correctly (if the font is not installed) it would follow that not every video will play back in every browser.

As a practical matter, it would be nice if serving a page with video were as simple an affair as possible. E.g. figuring out which video to serve didn’t involve sniffing the browser, operating system, and so forth; better yet, if one video format worked everywhere. As a practical matter right now H264 is the best candidate. VP8/WebM will never be the best candidate because by the time there’s a critical mass of hardware support out there it will be obsolete. This is a stupid, stupid fight.

And yet one more thing:

It’s interesting that the companies still in favor of h264 (Apple, Microsoft) are precisely those companies who do not implement the codec in the browser. Apple and Microsoft both implement h264 as a plugin architecture at OS level rather than a plugin at browser level (a much worse thing — see this excellent piece that daringfireball brought to my attention).

Is a it a font or an image file format?

The flipside of my argument that H264 should be considered analogous to a font is that, generally speaking, text is still legible when presented in the wrong font. By that argument H264 is more like an image format (JPG, PNG, etc.). If we accept this argument — which I’d say is the most h264-hostile stance (within reason) to take with respect to video codecs — then consider that most browsers simply let you display pretty much any image that’s convenient inside an <img> tag (sometimes badly, as per Internet Explorer’s notorious mishandling of PNG files over the years), generally by using the underlying OS’s APIs for handling images, which is exactly what I’d suggest the idealistic and pragmatic approach for video ought to be.

Would it be great if there were one codec out there that worked everywhere that web developers could target? Sure. But that doesn’t mean not supporting video codecs that happen to be around anyway, just as I can click on a PSD or TIFF in Safari and see it in the browser.

Ultimately, Google’s stance would have web browsers simply refuse to play back content with non-standards-based content (unless it’s Flash). What kind of “principled” or “non-evil” position is that? Again, if Google were to drop Flash support and make the argument that HTML5 is “the platform”, then it could make some kind of argument about consistency, but that’s not it. Google is making Flash part of “the platform” but not H264.

Climate Change Solved

I just finished reading Superfreakonomics. Bad title, and really not as good as the first book (essentially because it spends more time dealing with fewer, less novel ideas), but the final chapter (Why is Al Gore like Mount Pinotubo?) — which discusses Climate Change — basically underlines the fact that we know how to solve Climate Change and the question simply is how long before we feel compelled to employ a (real) solution?

The solution put forward in the book (which is the same as that described in the Atlantic article, linked above) is to increase the Earth’s albedo by injecting sulfur dioxide (using a hose suspended by balloons) into the stratosphere. This is what Mount Pinotubo did when it exploded, lowering the Earth’s temperature by about 1°F. The amount of sulfur dioxide involved is quite small, the negative effects are understood, and if we really find the idea of deliberately polluting the atmosphere repugnant, we can simply divert pollution we’re going to create anyway (e.g. from existing coal-fired power plants).

There’s a reason that most environmentalists don’t like this solution, and the core of it is that it doesn’t involve becoming more virtuous in other ways. In fact, it enables us to keep on doing the bad things we’re doing that got us in this problem in the first place — i.e. eating beef, burning fossil fuels, clearing rainforests, and covering huge areas in concrete. I imagine that if, say, the sun were mysteriously increasing its radiation output and we were all going to die unless we did something, injecting sulfur dioxide into the atmosphere would obviously OK. The big problem here is that we’re fixing the outcome of one kind of bad behavior with, in essence, a new form of bad behavior. When might it all end?

The “correct” way to fix the impact of human beings on the earth is depopulation. Anything less than this — recycling, driving a Prius, eating locally grown alfalfa, is just tinkering at the margin. Unfortunately, very few people would like to be part of this solution, so we need to figure out something that works. Yes, the sulfur dioxide solution is “bad”, but the question is do we do something bad now, while we still have most of our biodiversity, coastal real estate, and so forth, or wait until our “less bad” solutions, like trying to convince China and India that they don’t get to have air conditioning or nice cars, fail abjectly and then do it anyway?

The environmental movement is going to have to go back to trying to persuade people to do Good Things, like recycle, because they’re The Right Thing To Do, and not because it will solve climate change.

The big question for me now is what would we do if we needed to raise Earth’s temperature? An ice age would be far more catastrophic than global warming; is there a similar hack we could employ if the sun starts to cool or we go overboard with the sulfur dioxide?

Post Script

Ars Technica looks at five geoengineering options we have as alternatives to actually cutting carbon emissions, of which one is pumping Sulfur Dioxide into the stratosphere, which is dismissed thus:

They also evaluate a frequently referenced scenario, pumping sulfur dioxide into the stratosphere. Based on current estimates, a doubling of CO2 would add four Watts per square meter to the climate system. Canceling it, the authors calculate, would require the equivalent of a Pinatubo-sized volcanic eruption every other year. We could do it, but it “would lead to several undesirable consequences such as disruption in precipitation patterns and stratospheric ozone, and do nothing to avert the continued absorption of CO2 by the global ocean leading to rising acidity and ecosystem damage.”

The other options aren’t especially convincing either (collecting all our cars’ carbon emissions for example). I’d suggest that changing precipitation patterns is a given whether we do something or not, so it’s probably not a compelling argument on its own. Arguing that it doesn’t solve other problems (because it doesn’t actually remove CO2 from the atmosphere) is also almost beside the point (the same argument applies against the giant space mirrors, but I suppose that wasn’t worth mentioning since the idea is impractical to begin with).

The bottom line is, we’re not going to melt our icecaps and drown because before we get that far we’ll start pumping Sulfur Dioxide into the stratosphere (unless we find a better option). In the mean time, we still need to cut CO2 for its own sake.