Video, Freedom And Mozilla | Hacker News Reader

Video, Freedom And Mozilla

weblogs.mozillazine.org

55 pointsfelixmar65 comments

"- But I could just download gst-plugins-ugly and I'd be OK. - That's a selfish attitude. Everyone should be able to browse the Web with a free software stack without having to jump through arcane hoops to download and install software (whose use is legally questionable)."

I'll tell you what's selfish: going out of your way to cripple the experience of your users just so your extremist ideological views can prevail (and they won't anyway).

I don't think the deliberate weakening of the #1 open-source browser over such trivial matters is a win either, ideological or otherwise.

I love Firefox and I hope this sorry state of affairs will be quickly resolved sensibly before irrecoverable damage has been done.

</opiniated>

It's interesting to draw an analogy between this and Apple's lack of support for Flash, especially given the lack of support for the prior but the large amount of support for the latter on here.

I can only conclude you think it's equally selfish for Apple to cripple the experience of their users so their ideological views can prevail (and that they won't anyway).

I can also conclude that you don't think that the deliberate weakening of the #1 smartphone browser over such trivial matters is a win either, ideological or otherwise.

That's not quite a fair analogy. If Apple allowed Flash on the iPhone, it would have real consequences: lower battery life, higher RAM usage. I don't doubt for a minute that there's an ideological component as well - Apple want to see Flash gone. But at least that decision is based on technological problems with Flash.

On the other hand, if Firefox were to support OS H.264 decoders, it would harm nothing but Mozilla's ideologies.

The technical problems are trivially overcome by requiring the user to press a button to activate Flash and/or providing an option, so that's pretty much moot.

Pressing a button does not change the abysmal performance of Flash on a Mac, why would one expect it to change this problem on a iPhone? The positions being taken by Mozilla and Apple in these cases are similar, but Apple's technical justification has merit while Mozilla's political stance has no technical merit to hide behind.

I have flash lite on my shitty noname nokia I bought for 20 eur, I can play youtube videos on webpages, and it even played some flash games I made with HaXe. I am certain iPhone has much better cpu than this ordinary phone, so I don't see technical merits in apple's decision but exclusiveness and closing itself in walled garden.

same for java btw..

If in couple of years all web clients support only patented video codecs (which will happen if Mozilla doesn't succed here). Imagine doing a video startup or anything related to video that isn't irrelevant then.

Everybody's selfish, but Mozilla is also short-sighted ... they cannot and will not be able to have an influence on the defacto standards for codecs if they aren't popular enough.

This was the right time to really push for HTML5 and standards ... to push Flash/Silverlight/Quicktime out of the way. Safari, Chrome, Opera and Firefox combined have enough market-share to do that.

But no ... they take the same stance as Linux distros, all proprietary stuff is evil, ban it all. Unfortunately, while they have a point (H264 is not safe) ... they'll just lose users to Chrome or Safari.

This is exactly how we ended up with things like ActiveX and Flash in the first place. It's not enough to make things just work today. Mozilla is right to be thinking about what happens a few years down the road.

Mozilla has enough market share to make it relevant to web developers. Who will have to learn about the issue, and support Firefox. The users don't matter nearly as much.

Talking of ActiveX, apparently Mozilla had code to support that and made the ideological decision not to. That worked out well for everyone.

> This is exactly how we ended up with things like ActiveX and Flash in the first place.

No, and I really don't see the resemblance. We ended up with ActiveX, Flash and Java (you forgot about it) in the first place, because there was no other way to get around the browser's limitations.

In fact it's quite the opposite ... wasn't XmlHttpRequest implemented first as an ActiveX object in IExplorer 5? Or do you think some standards body magically came up with it and everyone was happy?

H264 is the current standard for video, with lots of tools and lots of momentum ... in order to replace it you'll have to put a whole lot more on the table than ideology.

Mozilla's decision to push for Ogg seems like a classic programmer's response. They're attempting to solve the problem with an engineering solution.

Unfortunately, this won't be enough. If Mozilla want to usurp H.264, they'll either need a huge market share or budget (which they don't have) or technological superiority over H.264 (I don't think this is the case either). They're also fighting against Apple, who want H.264 to succeed.

They do have a huge market share when it comes to HTML5 browsers, though. (Hm, the moment I’m writing this I’m starting to think about the iPhone and Android. So maybe not :)

Mozilla isn't making any choices. supporting theora is their only option due to patents surrounding H.264. It has nothing to do with engineering, and everything to do with copyright and patents.

They're not supporting H.264 because they don't want to promote Windows 7 (or Mac OS X) over Linux. They could easily implement H.264 support for all operating systems that license H.264, and then let websites dynamically choose what to send based on what the operating system supports. However, that means Firefox would work better on Windows 7 and Mac OS X than it would on 100% legit Linux distros.

I think this is an incredibly silly decision. Basically it boils down to "operating systems A and B support feature X while C doesn't, but we don't want operating system C to be at disadvantage so we will just disable access to feature X on A and B too."

That's already the case, though. JS performance and flash is far superior on windows than Linux already. So saying they don't want to make the windows experience better is not the case. Something like 80% of firefox users are on windows, and that's their primary target platform.

The difference is that Mozilla could make installing a H.264 decoder plug-in from some third party source on the first occasion the browser meets a H.264 video ridiculously easy if they chosen to. They did it with Flash, for example. But they've chosen not to.

Er, Flash plays H.264. They've already done that.

Yes, but this discussion was about H.264 support in the <video> tag.

Yes, and there exists javascript libraries and user agent hacks to fix that already. My point was that this isn't an engineering problem.

If the H.264 video tag wants to win over Flash, they'll either need a huge market share or budget (which they don't have) or technological superiority over Flash (I don't think this is the case either). They're also fighting against Mozilla, who don't want H.264 to succeed.

Chrome supports H.264 <video> and it has both huge budget and rapidly growing market share.

Chrome supports Theora <video> and it has both huge budget and rapidly growing market share.

Chromium supports only Theora <video> and it has both huge budget and growing market share.

Google is in the process of paying $100 million for On2, a company whose business model was building video codecs that didn't infringe on anyone elses patents, and the original source for VP3 which was the basis for Theora.

Almost nobody builds Chromium without trivially linking in ffmpeg.

In the real world, Chromium supports every damn codec and container format. All of them. Safari has identical support with Perian installed, and Opera does too with gst-ffmpeg.

The official Chrome supports Theora, Vorbis, H.264, AAC, MP3, and WAV in all reasonable container formats. Mozilla only supports Theora and Vorbis in ogm containers, plus WAV/RIFF for <audio>.

Good luck with redistributing that ffmmpeg-linked build of Chromium.

Or do you think that end-users are going to compile their browser?

Chromium always links with ffmpeg to support video, the question is whether you disable patented codecs in a fit of spite. Ubuntu distributes both those options: https://launchpad.net/~chromium-daily/+archive/ppa/+build/14...

There's some preemptive wankery in labeling the useful version of the ffmpeg plugin package as 'nonfree' despite the fact that Ubuntu's normal ffmpeg package in multiverse ships with all the Decoders enabled (though there's a separate 'nonfree' ffmpeg with all the Encoders enabled)

Mozilla makes $80 million a year. That's not huge enough?

And has over 50% market share in some locations and about a third globally. And both Chrome and Opera are supporting Theora too which adds to the total for this particular fight.

In this case the browser market share doesn't mean much. It is mindshare among the people providing the videos that matters. The way this particular battle is going to play out is that Mozilla will stick to its guns for a while and will continue to lose marketshare to Chrome, video providers will encode in HTML5/h.264 and Flash and eventually Firefox users on video sites will start seeing messages similar to the ones IE6 users see on some sites. The wildcard in this is Microsoft. If someone over there develops a clue they will jump on the h.264 bandwagon just to hose Mozilla -- IE8 will stop bleeding marketshare and this will provoke the video providers to move hard and fast to HTML5/h.264, forcing MozillaCorp to eat the costs of buying h.264 licenses and sticking a sharp stick in Adobe's eye as icing on the cake.

The first video provider to offer HTML5 video happened nearly a year ago and was dailymotion, who happen to have a euro focus, hence greater Firefox (+ Opera) adoption in their userbase. So that's a data point to Mozilla in the video provider mindshare column. Wikipedia's worth a mention too.

In revenue. About 50% of that go to operating costs. Developers need to eat too and they need to have a buffer for the near future. It's not wise to spend it all on giving unlimited H.264 licenses to the world.

There was a recent HN post on this topic that garnered a lot of discussion:

http://news.ycombinator.com/item?id=1070780

This new blog-post from Robert O'Callahan (one of the Mozilla developers who worked on the <video> element) explains Mozilla's position much better than anyone in the HN thread did, touching on things like patent licensing, using GStreamer, and idealism.

Idealism is an excellent way to loose users to Chrome.

Or a way to make users even more die-hard Firefox users. I don't understand how this is extremely idealistic. This is an excellent way to show how software patents are bad.

Sure, I'm agreed software patents are bad. But average users just want stuff to work.

Remember those music players that didn't play mp3 files so they didn't have to pay a few cent royalty? No me either, but they existed for a time. Everyone wanted to play mp3 files, so the players that could play them won out.

>Sure, I'm agreed software patents are bad. But average users just want stuff to work.

Right. Yet you've said before that you'll get rid of your iPhone if "they put Flash on it". How do you reconcile your position that "average users just want stuff to work" against being so vehemontly against Flash on the iPhone?

? Flash is amongst the worst examples of 'stuff that works'. Flash doesn't work. It pegs your CPU at 100%, crashes, and generally ends up in mess.

IF they can get flash to work, and not crash, and not eat battery, and not peg CPU, then I'd probably welcome it. But they haven't managed to do that in the last 10 years with the standard versions of flash, so I think it's extremely unlikely they'd manage it on slimmed down hardware like the iPhone.

And how is that working in other areas? How many Linux desktop-users do you know? And out of those, how many use "pure" distributions, like gNewSense?

Other issues aside, I don’t understand why supporting a proprietary Flash plugin from a single vendor is better than opening support for a standardized (albeit similarly patent-encumbered) video format with open-source implementations.

I (as an end user) don’t care so much about freaking patents. I only see that my MBP sounds like a fighter jet when playing flash videos, not when playing h.264 videos.

Since 2004, Mozilla has had baked-in support for automatically installing Flash on the first encounter if the plugin is not found. A nice little yellow infobar pops down (a brilliant UI innovation), prompting you to install it with a few clicks, even without root access on both Windows and Linux.

They've also recently implemented automatic update checking for Flash. Since it's their biggest security hole, they throw up a big nasty "update now" warning on launch if you're using a known-vulnerable version.

Mozilla initially distributed the Flash binaries under license themselves via addons.mozilla.org -- I'm not sure if they still do so.

Flash is shitty, nonredistributable, closed-source, restricted-platform, proprietary, and patented but they're willing to go to great lengths to help their users use it. Why not do the same thing for ffmpeg, which is merely patented?

From the article: It pushes the software freedom issues from the browser (where we have leverage to possibly change the codec situation)

This is what it's all really about: trying to use Mozilla's leverage to foist an unpopular and inferior technology on an unwilling web. Remind you of the antics of any other company with a popular browser?

Power corrupts. Thank god Firefox doesn't have the same market share as IE did. Webkit ftw, here.

The discussion I've seen surrounding the issue is mostly centered around the user (browser) and video quality. The thing that most stuck out to me from the post, buried in the end, was the (potential) restrictions that could be placed on creating h.264 content and publishing it on the web. I think that this is the biggest relevant issue I've seen yet with h.264. Apparently these issues are still potential and still under wraps, but I would not be in favor of effective restrictions on including the <video> tag in a website.

Sorry for being very off-topic, but that three-column layout (or more in higher res?) of the text makes it hard for me to read. Resizing also makes me lose where I was reading.

Sorry, again.

It looks great in IE--its all one column.

Sometimes lacking a "feature" can be a feature in itself.

Once google completes its merger with On2 (February), we'll get free and open video codec for free on youtube. Firefox and chrome will include it. This is why I think we'll never get Theora on youtube, but all hope is not lost thanks to vp7 or whatever they end up calling it.

With all the discussion going on, I was looking for a comparison of the codecs and found this website: http://people.xiph.org/~greg/video/ytcompare/comparison.html

If you ask me, the h264 vs ogv versions were not that different at the same bit-rates (sure, maybe a little). My point is, I'd really like to see more actual data before we enter endless discussions and most of us don't know what's actually talked about.

I'm not accusing greg of anything, however, I personally would feel a little uneasy trusting someone that works for one of the entities involved, to be making unbiased comparisons.

I'd distrust a comparison page hosted by an Apple employee for the same reason.

That was just the first comparison that I came around at it sounded reasonable. I would actually prefer if someone else would show more of these. But not a lot of folks do that. Everyone is just bitching around.

Also, maybe there are problems with theora, but then again, it's better to push it to more attention and fix the problems now than settle with a halfassed, non-sustainable solution and having the same discussion same time next year again. Seems to be very ineffective.

Problem is not that Theora isn't capable enough now that MozCo has poured money on it. The problem is that Google and every other video service has already their videos in h264, which is playable in Flash. So if YouTube were to support Theora they'd need to re-encode all their videos, and store them as duplicates.

Another problem with Theora is hardware. There is a lot of hardware h264 encoders and decoders, allowing even cellphones to play HD h264. AFAIK there is absolute no Theora hardware currently.

Sounds like a chicken and egg problem to me. If we collectively ignore ogv because there are no current hardware en/decoders, then why should anyone build it? These things are built when enough demand is present. For that, the push from the Mozilla foundation seems to be reasonable. Once there is more demand, the tech comes automatically. If we clinge to H264, nobody will ever build the right hardware. It's basically a demand based resource allocation problem, and Firefox is doing the demand part. Seems reasonable to me.

Mozilla has not poured any money into Theora that I know of. The primary developers are all working on it in their spare time or as part of their jobs with other companies. For example, Monty is employed at RedHat.

The fact that Theora is competitive to the best proprietary codecs without any real funding is a huge success. Imagine what would happen if Google or Mozilla put a little effort behind that?

I'll state up front I think that patent-royalty free codecs are the future, and that Theora is the best of that class at the moment.

But codec comparison is a complicated subject. That link isn't H.264 vs. Theora, it's the particular H.264 encoder and particular settings used at that time by Youtube versus some similar bitrate Theora. It was prompted by an ill-informed comments by Google's Open Source dude who basically said that Youtube using Theora would melt the internet. The test only shows that Theora isn't a ridiculous choice for most web video, not that it's better or even as good as H.264 in general.

The idea is that open trumps closed over time. That's why we're using linux on our servers, and web technologies on top. These were (and are in some cases) toys compared with the competition, but got better by being open. It's only a matter of time till video codecs go the same way, supporting Theora will speed this up and we'll all benefit.

He touched a very important point. Many people do not realize that GPL and LPGL are fundamentally incompatible with patents (intentionally) and that in many cases it is a ticking bomb.

When he writes "The software license permits you to redistribute and use the code, but the MPEG-LA can still stop you", he means section 7 of the GPL (2.0). I think many people who use the GPL for their software do not realize that later on a single company (patent holder) can stop the redistribution of that software.

Any patent holder with relevant patents can stop redistribution of a software, regardless of license ... GPL or not.

If one were to create an open-source "proxy extension" for Firefox to shim H264 support via proprietary binary blobs dynamically loaded into the proxy, how do you think that would go?

why proprietary binary blobs and not FOSS ffmpeg/libavcodec?

edit: but I like that idea... Would it be possible to create an Fx extension/plugin to enable h264 playback? I mean, QT and Flash plugins do already play h264 video with Firefox in <object>/<embed> -tags.

That's exactly what I'm trying to ask. Not if it were technically possible, but if it would be worth the while.

Why bother using a FOSS implementation of a patent-encumbered technology? FOSS for EMCA standards (.NET, Java) is great, but FOSS for h264 is kinda pointless (which is part of the argument in TFA)

Because most of the world is not under US jurisdiction.

No, it's not possible to do it via XUL and NPAPI.

You would have to fork both Gecko and XULRunner.

This is idotic. By not supporting h.264 and not getting in trouble if and when MPEG-LA decides to cease and desist, they are making it much easier for h.264 to become completely ingrained because h.264 is what works NOW and it is what is widely supported NOW. Why? Who among the vast majority of blithering idiots in the world knows or cares about the patent issues with h.264? Outside of Hacker News and the Mozilla corporation, I'd put that number at just about nil.

Yo, has anyone started work on a Mozilla fork that does support h.264 either through native support or plugins?

edit: Also, if no one does sue, then the whole thing will have been simply an exercise in idiocy.

Well, if Bilski goes our way, then this whole issue is mute, right? Or are there some other restrictions on the codec?

I agree that OGV is pretty much Mozilla's only option for now.

Not mute. No.

It's called a moo point. You know, like a cow's opinion.

Heh, sorry, it's early.

*moot

Reply on news.ycombinator.com