On Sat, Feb 7, 2009 at 6:52 PM, Kevin Kofler <kevin.kofler@xxxxxxxxx> wrote: > Martin Sourada wrote: >> No, they won't be. They're using whatever framework there is on mac and >> they'll be using whatever framework is default on windows. It's no >> different from video intended from download - you rely on the customer >> to install the needed codecs himself. > > And this in turn will draw webmasters to Flash (or rather fail to draw them > away from it, the web is already infected with it all over the place), > which has a known set of codecs. :-/ > > The whole point of having native video support in HTML is that you don't > have to install stuff, so having some known-supported codec is essential. > Therefore, IMHO it was a big mistake that the clause which made Ogg > Theora/Vorbis support mandatory got removed from HTML 5, and the WebKit > approach is not going to be reliable. > > Using the native multimedia framework does make sense, but at that point the > browser should be responsible for installing the codecs Xiph.org provides, > not the user. Xiph.Org's statement on matter from the time of the argument was pretty good: http://xiph.org/press/2007/w3c/ It makes obvious sense to support as many codecs as possible via the HTML audio and video tags: Users have differing needs and formats improve over time. But without a *baseline* set of reasonably performing codecs web developers have nothing to target when they simply want to produce something that JustWorks(tm), and nothing reasonable to fall back to when superultramegacodec is not available. You've been able to <embed> video for years, yet it's not widely used precisely because of its unpredictability. Pretty much the only argument against including Ogg/Theora and Ogg/Vorbis as *baseline codecs* was the patent scaremongering by the likes of Nokia and Apple. As the conspicuousness of the existing widescale Theora deployments increases these arguments will vanish, and perhaps a new standard will correct the omission. Firefox is pretty much a static build on most of its platforms— it's the only way out of the DLL hell. It doesn't use a system SVG library on Windows, so why would we expect it to use system A/V codecs? The Fedora shipped Firefox does use system libraries, I will expect it to use the system copies of libVorbis and libTheora (and Speex and Dirac too). Using the native infrastructure would be ideal, but there is a different one for each platform, and using it would result in a lot of additional failure modes. Getting one single portable solution working well has taken enough effort, but perhaps they'll do that in a subsequent version. This has all gone pretty far off topic: Ogg/Theora and Vorbis and Dirac just work in Fedora today. It's the rest of the world that is catching up. Additional container formats should be supported in Fedora, so long as the software doesn't violate the licensing guidelines. I don't think Fedora needs to 'take sides' in one format vs another beyond the requirement that Fedora can only support and advocate formats that meet the licensing requirements. Fix packages, submit patches, nag upstreams, etc. No controversy there. On topic for this list would be the question of Firefox 3.1 in Fedora. It's going to be rather silly when the Windows world has rapidly switched to Firefox 3.1 and has <video> support, while Fedora is still shipping 3.0 without it for many months. (At Wikipedia I saw on the order of 1/4 adoption of 3.0 *the day after* the release) -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
