Instead of a complete followup post, I therefore decided to bring back the original text basically unchanged, and added a separate update section that addresses the individual point of views brought up here.

Thanks to everyone for chiming in!

Anyway, I look forward to your follow up post.

Wikipedia has 14 definitions. 12 say “no patent fees”, 1 is a historical accident, much like someone saying that they had “Open Source” or “Free Software” years before those gained defined meanings.

That leaves one definition, created by the patent lawyers of the telecoms standards group that developed H.264 (somewhat of a coincidence no?). If you read the statement that is quoted in full you find they are actually reacting against the attempts by various governments and standards organizations to establish “Open Standards”.

Note that in Steve Jobs’ Thoughts on Flash he refers to HTML, CSS and Javascript as “Open Standards” repeatedly, but calls H.264 an “industry standard” instead. He knows there is a difference and doesn’t generally mess it up.

He will do clever and misleading things e.g. saying that FaceTime could be an open standard is true just like saying HTML5 video tag is an open standard. This doesn’t mean they can’t both use proprietary video codecs (or image files etc.) and therefore make the combination non-open.

I have been programming voip software for lots of years, and I can guarantee you that these is an OPEN software architecture. As usual, people are confusing OPEN with FREE (as in beer). The architecture as outlined here, is nothing special, and is basically similar what most voip systems (the one that are used by telecommunication companies, not Skype) use these days. The whole point is that you can easily (*) build a voip system (a desktop app or a mobile phone or whatever) that can interoperate with this. It’s standards compliant. And that’s the whole point – you are NOT forced to use a mobile phone from 1 manufacturer, you can choose your own (aside from restrictions that your mobile provider might impose, but that’s something else).

Note that the use of a separate connection (over the internet) to let video stream bypass the normal phone company network, isn’t particularly new. But Apple might have problems at first, since many phone companies might not like it (I have programmed servers before that explicitly blocked this). Maybe if they hide it in a propriety SIP header instead of in the RTP body ? And the QoS might be a bit difficult to guarantee if it needs to be passed through more than 1 ISP. We”ll see.

The only thing a bit surprising is the limited number of codecs built-in (H.264, MPEG-4, M-JPEG). It seems that Steve seems to force other manufacturers to use H.264 as a common choice – not a bad choice from a quality point of view (H.264 has a pretty good compression profile for this type of video). The point that are are other codecs -free or not- free is moot, if there are no voip clients that currently implement them. Google might be a big influence – they’re capable of pushing VP8 in their android phones. But even they might be forced to support H.264 too, to be able to inter-operate. H.264 is also found in many IMS voip systems, which is the new standard for phone companies, so I wouldn’t be surprised if they will. And maybe force Apple to support VP8 too (I hope), and the other common codecs (H.261, H.262, H.263, …).

Note that the use of codecs for video calls has nothing to do with their use in HTML5. But smartphones will obviously share their codecs between their browsers and their phone software.

(*) Yes, I have implemented (parts) of such a system. Really.

It’s really that simple.

To be fair, no handheld device will ever be capable of encode/decode without a hardware encoder/decoder. Regardless of motivations for or against webM, apple didn’t have the time to build webm support into their chips while they already included h.264 for video recording and playback.

Yes, definitely a muddy pit I stepped straight into, sigh.

Well, you’re certainly entitled to your opinion about what “open” means. But it might be informative for you to read about RAND licensing and varying definitions of terms like “open standard” and “open specification”. In many contexts, it is considered fair to charge a bucket of money for a patent license on an “open” spec/standard.

http://en.wikipedia.org/wiki/Open_standard#Patents
http://en.wikipedia.org/wiki/Open_specifications

Of course, I wish circumstances were different, but I think it’s clear that “open” is still a very muddy term.