Live data from Hacker News

Google Chrome's H.264 support not true "free" software

ianweller.org

31–40 of 85 posts

Re: Google Chrome's H.264 support not true "free" software

#31
post #22
post #21

I began to wonder how long it will take for someone to begin working on a fork of Firefox which would include support for h264. I have already seen patches for both DirectShow and GStreamer backends, so I'd guess it's not that hard technically. What we need is a catchy domain name, some marketing and a few devs and the spice shall flow again.

Who's going to pay the licensing fees? If you fork, you probably don't get access to the Google money spigot.

What licensing fees? h264? Well I'm not entirely sure how, but lots of FOSS projects support h264, most prominently x264 and ffmpeg, and software building upon them. Just do it like they do.

... or just use system codecs.

besides I'd guess Google would be happy to have h264-supporting Firefox, and end this farce, so getting some kind of sponsorship from big G wouldn't be that improbable.

Re: Google Chrome's H.264 support not true "free" software

#32
post #18

Earlier quoted context omitted.

I'm not sure you understand the problem here. I do, which is why I am so critical of Mozilla's "solution" to it, which has confused the issue of not being able to ship h.264 support for the reasons you mention with trying to promote a format that (whatever its merits) nobody wants. Ideally, that's the approach Mozilla would take, too. Indeed, it would be, but my understanding is that they are unwilling to do so or un…

Suggesting nobody wants Theora/Vorbis is absurd. Clearly someone wants it, as it was created. The important questions are whether it is useful enough to merit inclusion in a browser, and whether it's useful enough to outweigh the downsides. The same questions should be asked about H264.

But it wasn't created for this purpose. It was originally created by On2 to be sold commercially, but nobody bought it because it sucked. They dumped it as royalty-free abandonware, and some free-software partisans picked it up, changed some strings, and released it as Theora.

Nobody wanted it then, nobody wants it now. The dumping could even be seen as a brilliant strategic move to troll the hell out of the market for codecs by building a community of earnest activists yelling at everyone.

Re: Google Chrome's H.264 support not true "free" software

#33
post #24

What about x264? I know there is speculation that x264 wouldn't have much chance to withstand a software patent fight but OGG Theora isn't 100% safe either. If Mozilla is already going down that road they should just adopt x264 and see what happens. Challenge the patent holders to outrage the public and cripple the Internet. That's going to do more for software freedom than sticking their head in the sand.

x264 is a h264 encoder implementation. Mozilla couldn't "just adopt" it, mostly because they have little use of an encoder. Its the decoder what they need.

Re: Google Chrome's H.264 support not true "free" software

#34
post #23

Earlier quoted context omitted.

I'm not sure you understand the problem here. They do not have a license to use h264, and even if they did, they could not afford to share it with you. Distributing patent-encumbered GPL code violates the spirit of GPL-2 and the letter of GPL-3. If you don't like it, campaign against software patents. Chromium's approach to this is slightly better, but it's not all there yet. (Speculating here) Google can afford to d…

Patches are not welcome. Mozilla is not going to back down on this. Their implementation is designed to make it extremely difficult to swap in other codecs from a plugin or straightforward patchset. You'd have to fully fork at least Gecko and XULRunner, and extend Firefox. Due to their strong Trademark policies, your fork couldn't use their names or logos (see also: Iceweasel).

You're right on the former, I was under the impression that it was an economy-of-effort thing rather than an ideological thing. However, I'm not sure I agree that it would be as hard as you think. The mime types are registered here[1] and the impls are here[2]. Each media type is implemented as a state machine that could wrap ffmpeg. That's where I would start, I suspect.

EDIT: Looks like someone's already blazing that trail[3]. They're using gstreamer, but the research they do for that will show anyone how to add a new one.

--

1) http://hg.mozilla.org/mozilla-central/file/b27f05c8743b/cont...

2) http://hg.mozilla.org/mozilla-central/file/b27f05c8743b/cont...

3) https://bugzilla.mozilla.org/show_bug.cgi?id=422540

Re: Google Chrome's H.264 support not true "free" software

#35
post #28
post #22

Earlier quoted context omitted.

Who's going to pay the licensing fees? If you fork, you probably don't get access to the Google money spigot.

Apple (QT) and Microsoft (DirectShow) pay the license fees when they distribute their Operating Systems. Nobody gives a shit about distributing the x264 source code, and only the fosstards even have a hiccup about distributing binaries, especially from hosts outside the US. Ubuntu may not ship with the gstreamer plugin, but it does prompt you to download it the first time it's asked for by an application.

Sorry, my US bias is showing. I totally forgot about the "screw patents and just ship it from Europe" solution.

Re: Google Chrome's H.264 support not true "free" software

#36
post #22
post #21

I began to wonder how long it will take for someone to begin working on a fork of Firefox which would include support for h264. I have already seen patches for both DirectShow and GStreamer backends, so I'd guess it's not that hard technically. What we need is a catchy domain name, some marketing and a few devs and the spice shall flow again.

Who's going to pay the licensing fees? If you fork, you probably don't get access to the Google money spigot.

VLC is one of the most popular open source programs in the world. Over 100 million copies have been downloaded since the 1.0 release, with an all time download count of nearly 400 million.

They support not only H.264--which has a relatively reasonable, low-cost licensing system--but many formats whose licensing models are absurdly overpriced and unreasonable (e.g. MP3). And furthermore, VLC not only supports open standards like H.264, but also plenty of proprietary formats, like VP6, MSMPEG4, and Real Video.

They don't pay a cent to anyone.

Re: Google Chrome's H.264 support not true "free" software

#37
post #16
post #15

Earlier quoted context omitted.

This is what I was wondering. If this is true that what we're talking about is linking against proprietary libraries that already exist on the user's machine, then I don't understand the argument against it. By the article's logic they should pull the plug on firefox for windows & mac since they are more "non-free" than h.264.

http://weblogs.mozillazine.org/roc/archives/2009/06/directsh... Firefox won't use system codecs because they might be insecure, they might be buggy, and Mozilla wants to encourage Web authors to use Theora.

Firefox won't use system codecs because they might be insecure, they might be buggy

This does not address the original objection. Calls to the Mac and Windows are not known to be any less buggy or insecure than calls to the h264 APIs. So why accept one and not the other? The only sane answer of course is that you can't build an app at all without calling the OS basic APIs of your host machine, but you can try to have an ugly work around h264.

Re: Google Chrome's H.264 support not true "free" software

#38
post #20

Earlier quoted context omitted.

There's a relatively simple solution to this that another heavy hitter in the FOSS world has been using for years: Prompt for a "restricted package" install when the user first encounters an h.264 video, similar to how Ubuntu deals with non-FOSS drivers, plugins, and codecs. It's the best of both worlds; you stay free and you get to be compatible.

Mozilla links directly with liboggplay and actively avoids abstracting it in any way. It can't handle any non-ogm container formats, and ogm was intentionally designed not to handle any other codecs. They're just assholes. If software patents are repealed in the US they'll mourn the loss of the mainstay excuse with which they force earnest activist wankery on their users. If patents were the real obstacle they'd dist…

The "ogm" file format is an old, unsupported, deprecated variant of the Ogg container. Do you mean "ogv"?

Sorry, this may seem like nit-picking, but such a fundamental mistake makes it harder to take your strong claims about Mozilla's design and intentions at face value.

Re: Google Chrome's H.264 support not true "free" software

#40
post #9

Earlier quoted context omitted.

I value watching non-crappy video enough to pay $0.20 for an H.264 license. I value the runtime on my battery operated devices enough to pay $0.20 for an efficiently decoded format. (Maybe that is more than $0.20. Depending on volume the H.264 is $0.00, $0.20, or $0.10, but I think there is a base license also required that I can't find. In any event, adding H.264 to a computer probably costs less than half a can of…

So it looks like your value your freedom to $0.20. "And remember, I am still allowed to do anything I want with the code. I just have to give MPEG-LA half a can of beer do it." Then free software would'nt be ablo to be free software. You cannot preserve freedom this way.

You mean software freedom, I guess. Personal freedom is greater with BSD than GPL, and with public-domain software than with any of them. Personal freedom includes the right to sell software at $0.20 or $200, and the customer's right to buy it or not. For the record, I'm against software patents, but also against being forced to publish one's code, which incidentally maximizes personal freedom.
Post reply on HN